Savio.no - Blog

Tracking Content Scrollers, Scanners & Readers in Google Analytics

Eivind Savio — 2012-03-04

A couple of weeks ago Justin Cutroni wrote 2 awesome blog posts about Advanced Content Tracking using Google Analytics.

In Part 1 he shared the script he has written and is using, and in Part 2 he shared his findings and experience with this method.

I have been using his method with some small tweaks since before he published these blog posts, and since someone asked me if I could publish the script with my tweaks - here it is.

Differences between this script & Justin's script

Reader classification: 30 seconds. If you use more than 30 seconds to scroll the Content Area you will be classified as a Reader.
Check if content area has to be scrolled: Tracked as Scrolling Allowed.
Tracking of Content Length in pixels: Gives you more opportunity to analyze Readers vs. Scanners.
All tracking is done using Non-Bounce Events: You can change that, but I wanted to be able to compare Bounce Rate on a article level with Readers and Scanners.
Naming are different: The naming I'm using in my script is different from Justin's naming. I used a different Content Tracking script earlier, and kept the namings I used in that script.

Bonus: Analyzing your content in Excel

As Justin wrote, this data begs for Excel.

You can therefor download this Excel spreadsheet to analyze your data further. The spreadsheet requires the Next Analytics Excel plugin.

Using the Excel spreadsheet

In the _actions tab, change the following values:

Site Name
Date Start & Date End
Google Analytics Profile ID

The Google Analytics script for tracking Readers & Scrollers

Values in Blue: You may change these values
Values in Red: You have to change these values
The script requires jQuery.

Remove formatting from Code

// This script was originally written by Justin Cutroni, see http://cutroni.com/blog/2012/02/21/advanced-content-tracking-with-google-analytics-part-1/
var readerTime = 30; // Seconds after scroll to bottom of content before visitor is classified as "Reader"
var readerLocation = 150; // # px before tracking a reader
var callBackTime = 100; // Default time delay before checking location
// Set some flags for tracking & execution
var timer = 0;
var contentLength = 0; // Content Length -> Length of content area
var scroller = false;
var endContent = false;
var didComplete = false;
// Set some time variables to calculate reading time etc.
var pageTimeLoad = 0;
var scrollTimeStart = 0;
var timeToScroll = 0;
var contentTime = 0;
var endTime = 0;
jQuery(function($) {
// Check if content has to be scrolled
if ($(window).height() < $('#contentArea').height()) { // Replace contentArea with the name (class or ID) of your content wrappers name
pageTimeLoad = new Date().getTime();
contentLength = $('#contentArea').height();
_gaq.push(['_trackEvent','Page Scroll','Page Scroll: Allowed',window.location.pathname,contentLength,true]);
}
// Check the location and track user
function trackLocation() {
bottom = $(window).height() + $(window).scrollTop();
height = $(document).height();
// If user has scrolled beyond threshold send an event
if (bottom > readerLocation && !scroller) {
scroller = true;
scrollTimeStart = new Date().getTime();
if (pageTimeLoad > 0) {
timeToScroll = Math.round((scrollTimeStart-pageTimeLoad)/1000);
} else {
timeToScroll = ""
}
// Article scroll started
_gaq.push(['_trackEvent','Page Scroll','Page Scroll: Started',window.location.pathname,timeToScroll,true]);
}
// If user has hit the bottom of the content send an event
if (bottom >= $('#contentArea').scrollTop() + $('#contentArea').innerHeight() && !endContent) {
timeToScroll = new Date().getTime();
contentTime = Math.round((timeToScroll-scrollTimeStart)/1000);
if (contentTime < readerTime) {
_gaq.push(['_setCustomVar',1,'ReaderType','Scanner',3]);
_gaq.push(['_trackEvent','Page Scroll','Page Scroll: Content Scanner',window.location.pathname,contentTime,true]);
} else {
_gaq.push(['_setCustomVar',1,'ReaderType','Reader',3]);
_gaq.push(['_trackEvent','Page Scroll','Page Scroll: Content Reader',window.location.pathname,contentTime,true]);
}
endContent = true;
}
// If user has hit the bottom send an event
if (bottom == height && !didComplete) {
endTime = new Date().getTime();
totalTime = Math.round((endTime - scrollTimeStart)/1000);
_gaq.push(['_trackEvent','Page Scroll','Page Scroll: Page Bottom',window.location.pathname,totalTime,true]);
didComplete = true;
}
}
// Track the scrolling and track location
$(window).scroll(function() {
if (timer) {
clearTimeout(timer);
}
// Use a buffer so we don't call trackLocation too often.
timer = setTimeout(trackLocation, callBackTime);
});
});

Some final words

Before starting to use this script, you should read both blog posts from Justin Cutroni to get a grip of the differences between this script and his script.

Other people I should thank for this content tracking method is Thomas Baekdal (and I really recommend his Plus Content).

If you think the tweaks I have done to the code could be improved, feel free to comment (I'm not a programmer).

Behavioral Targeting using Facebook Connect & Google Analytics Custom Variables

Eivind Savio — 2012-01-31

In this blog post I explain and demonstrates how to use Facebook Connect as a Social Login function, how to track Facebook demographic data like Gender and Age Group, and how to do Behavioral Targeting based on this data.

The Behavioral Targeting will be examplified using 2 different methods:

Google Analytics Custom Variables
BTBuckets

I will not explain all my code in this blog post, it's simply just too much going on. However, you can download this demo at the bottom of this blog post.

This blog post covers:

Creating a Facebook App
Facebook Connect
Tracking Age Group & Gender as Custom Variables in Google Analytics
Behavioral Targeting using Custom Variables in Google Analytics
Reactivating Facebook Connect Tracking
BTBuckets Behavioral Targeting
Google Analytics Events
Facebook Connect Demo
Download Example Code

Create a Facebook App

The first thing you have to do is to create a Facebook App that the user can connect to (login to).

Facebook Connect

The next thing you have to do is implement the Facebook Javascript SDK that makes it possible for the user to connect/login to your Facebook App. My code is based on a Facebook blog post, and you find it in fb-connect.js.

What kind of Facebook User Data you can ask for is described in the Facebook Graph Api. In my demo I'm asking for your birthday, which isn't part of the standard data you can get from a user. You have to ask specifically for that.

Tracking Age Groups & Gender as Custom Variables in Google Analytics

As mentioned earlier I track Facebook Gender & Age Groups in my demo using Google Analytics Custom Variables. I track this in different slots and on a visitor level (you have to do that if you want to do Behavioral Targeting). I don't track the age itself, but as a age group (ex. 18-24). And since I'm a nice guy I don't group those below 18 year.

Since the number of Custom Variable Slots are limited, you can also combine the data if you have few slots available. Ex. like this: male_18-24.

Remove formatting from Code

window.fbAsyncInit = function(){
FB.init({appId: 'YourAppID', status: true, cookie: true, xfbml: true, oauth: true}); //Replace YourAppID with your real AppID number
....more code
//Set some Custom Variables
_gaq.push(['_setCustomVar', 2, 'Facebook Gender',response.gender, 1]); //Save Facebook Gender as Custom Variable in GA
_gaq.push(['_setCustomVar', 3, 'Facebook Age Segment',ageSegment(aSegment), 1]); //Save Facebook Age Segment as Custom Variable in GA
....more code

Behavioral Targeting using Google Analytics Custom Variables

To do the Behavioral Targeting based on our Custom Variables, we are using _getVisitorCustomVar. You will find this code in index.html.

Remove formatting from Code

<script type="text/javascript">
_gaq.push(function() {
var pageTracker = _gat._getTrackerByName(); // Gets the default GA tracker.
var visitorGender = pageTracker._getVisitorCustomVar(2);
if (visitorGender == 'male') {
document.getElementById('gaCV').innerHTML=('<p>If you are a male you will see this text.</p>');
} else if (visitorGender == 'female') {
document.getElementById('gaCV').innerHTML=('<p>If you are a female you will see this text.</p>');
}
});
</script>

Reactivating Facebook Connect Tracking

If a user deletes cookies from our domain, we will no longer be able to track Gender & Age Group if the user visits our website again. However, if the user still has authorized access to our Facebook App, we can reactivate this tracking.

You will find this code in fb-ga-tracking.js.

Remove formatting from Code

....more code
window.fbAsyncInit = function() {
FB.init({appId: 'YOUR_APP_ID', status: true, cookie: true, xfbml: true, oauth: true}); //Replace YourAppID with your real AppID number
function checkStatus(response) {
if (response.status == "connected") { // Check if user is connected to your Facebook App
_gaq.push(function() {
var pageTracker = _gat._getTrackerByName(); // Gets the default Google Analytics tracker.
var visitorGender = pageTracker._getVisitorCustomVar(2); // Slot 2 - Gender
var visitorAgeGroup = pageTracker._getVisitorCustomVar(3); // Slot 3 - Age Group
if (typeof visitorGender === 'undefined' || typeof visitorAgeGroup === 'undefined') // Check if Custom Variables for Gender & Age Group in GA are undefined. We only want to run FB.api if it's strictly necessary since it slows down our website.
//user is logged in and connected
FB.api('/me', function(response) {
....more code. Do the tracking.

BTBuckets Behavioral Targeting

BTBuckets is a free Behavioral Targeting solution that can be integrated with Google Analytics, and they also have an integration with Facebook Connect.

You will find BTBuckets scripts in my example code, but I will not explain this further except that the code is commented so you should be able to sort it out.

Unfortunately the BTBuckets Age Group tracking and Behavioral Targeting isn't working properly. I don't know why, but I think this is because of a bug in BTBuckets integration with Facebook Connect.

Google Analytics Events

User actions like Login & Logout will be tracked as Events in Google Analytics.

My script also tracks if the user is connected to our App using Event Tracking.

Facebook Connect Demo

Download Example Code

For the demonstration purpose I considered hiding the the download behind the Facebook Connect Login, but since I'm not that type of person, here you go:

Download Example Code (ZIP-file containing 3 files)

Some final words

If a user is logged in to Facebook and has authorized your Facebook app, you can in theory track the user across different devices. For example, I'm logged in to Facebook both on my PC (with 3 different browsers) and on my mobile. That means I could track myself across all those devices and browsers.

Don't abuse the possibilites and do that! You are not allowed to store Personally Identifiable Information (PII) in Google Analytics.

If you think the piece of code could be improved, feel fra to comment (I'm not a programmer).

Tracking Internal Search using POST request method in Google Analytics

Eivind Savio — 2012-01-02

Since Google Analytics is relying on query parameters in the URL to track Internal Site Search, Internal Site Search solutions that are using the HTTP POST method can't normally be tracked without some help from IT and server side implementation/programming. This because HTTP POST don't send any query parameters. And based on my experience, it can often take some time to get something like this implemented.

I needed a quicker solution for some friends that have slow opened their web shop Karma.no, and wrote a simple javascript/jQuery solution for them that also tracks Zero Result Searches.

Some example HTML code

Imagine that the search result page is found at the location http://www.yourdomain.com/search.aspx, and that the HTML-code below is generated on this page when a user has searched for something that resulted in Zero Results.

Remove formatting from Code

<form action="/search.aspx" method="post" name="intSearch">
<input name="keywordFieldName" id="keywordFieldID" value="some search phrase" type="text" />
<input name="searchButton" type="submit" value="Search" />
</form>
<div id="divName">
<p>No posts found. Try a different search?</p>

</div>

In the example HTML-code above, some part of the code have been highlighted with red text. We are going to use those values in the javascript/jQuery script.

The Javascript/jQuery solution for tracking Internal Site Search

To grab the search phrase, I'm using the jQuery .val() method. Replace values in red with your values.

Remove formatting from Code

$(document).ready(function() {
var strURL = document.location.href;
if (strURL.indexOf("/search.aspx") != -1) { // Only run the script on the /search.aspx page
var intSearchVal = $("#keywordFieldID").val(); // Keyword field ID. Replace with your text ID
var noResults = document.getElementById('divName').innerHTML; // ID of div where the text for "No posts found" is shown.
if (noResults.indexOf("No posts found.") != -1) { // Replace the text "No posts found" with (part) of the text you are using
_gaq.push(['_trackPageview','/search/?cat=No results&q='+intSearchVal]);
} else {
_gaq.push(['_trackPageview','/search/?q='+intSearchVal]);
}
}
});

Since we are using a virtual search result page name in the script above, we have to exclude the real search result page name (search.aspx) to avoid double counting of page views. The filter below will exclude search.aspx from the reports in Google Analytics.

Filter Name - Exclude Page: search.aspx

Filter Type	Custom filter
Filter	Exclude
Filter Field	Request URI
Filter Pattern	\/search\.aspx
Case Sensitive	No

Site Search Setting

The next thing you have to do is set your Site Search Settings.

Done! Happy tracking internal search phrases and zero result searches in Google Analytics from internal search using POST request method.

If you think the piece of code could be improved, feel fra to comment (I'm not a programmer).

Javascript Redirect & Google Analytics Cross Domain Tracking

Eivind Savio — 2011-09-24

Recently I had to write a Javascript redirect function that supports Google Analytics cross domain tracking. You find this function below.

The function below has a 5 second delay before it redirects. Change values in red according to your needs.

Remove formatting from Code

<script type="text/javascript">
_gaq.push(function() {
var myRedirectURL = 'http://www.domain2.com/';
var pageTracker = _gat._getTrackerByName();
myRedirect = pageTracker._getLinkerUrl(myRedirectURL);
setTimeout("document.location.replace(\myRedirect\);", 5000);
});
</script>

Also remember that you should normally add a regular link on that redirect page for those who have javascript turned off.

If you think the piece of code could be improved, feel fra to comment (I'm not a programmer). Happy Google Analytics cross domain tracking using Javascript redirect.

Tracking Facebook Likes, Share, Send & Comments as Social Engagement in Google Analytics

Eivind Savio — 2011-08-07

There are already posted "tons of articles out there" about how to track Facebook Likes, Send & Comments as Social Engagement in Google Analytics, but since I also uses my own blog as a reminder of how to do things and a "script bank", I decided to write this blog post nevertheless.

To be able to track Facebook Likes, Share, Send & Comments on your web site, you have to use the XFBML version (Javascript SDK).

Below I have listed what that this article covers, and what you should do (or consider doing):

Change your <html> opening tag
Implement Facebook Meta Tags (Facebook AppID & AdminID)
Register your web site as a Facebook App
Implement Open Graph Protocol (Meta Tags)
Use Facebook XFBML version (Javascript SDK) and load the script asynchronous
1. Use Facebook channelURL to avoid problems with older versions of Internet Explorer
Implement the Google Analytics Social Media Tracking Script
Implement my script for tracking Traffic from Facebook Likes, Send & Comments in Google Analytics (and read that article as well)

Facebook and Open Graph Meta Tags

I recommend that you implement both Facebook and Open Graph Meta Tags. Open Graph Tags will give you better control of how your shared content will appear on Facebook and Google+ (Google+ is also using the Open Graph Meta Tags).

It's recommended that you change your <html> opening tag when implementing these meta-tags. Your new <html> tag should look something like the code below.

Remove formatting from Code

<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" xmlns:fb="http://www.facebook.com/2008/fbml" xmlns:og="http://ogp.me/ns#" >

Facebook AppID and AdminID Meta Tag

To administer your page, you need to associate it with either your Facebook account or your Facebook Platform application. It is valid to associate your page with both user accounts and a Facebook Platform Application.

To associate the page with your Facebook account, add the additional property fb:admins to your page with a comma-separated list of the user IDs or usernames of the Facebook accounts who own the page.

Remove formatting from Code

<meta property="fb:app_id" content="130796876963854" />
<meta property="fb:admins" content="12345678" />

Replace values in red with your values.

Open Graph Protocol (Meta Tags)

The Open Graph Protocol gives you better control of how your content are presented on Facebook and Google+.

Remove formatting from Code

<meta property="og:title" content="Your Awesome Title for Your Page"/>
<meta property="og:type" content="article"/>
<meta property="og:url" content="http://www.domain.com/canonical-url-to-article"/>
<meta property="og:site_name" content="YourSiteName"/>
<meta property="og:description" content="Your Awesome Description up to 300 characters"/>
<meta property="og:image" content="http://www.domain.com/facebook-picture-150x150px.png"/>

Since Google+ also uses this data, I recommend that the size of your og:image (Object Type) has a width of minimum 150 pixels and not 50 pixels (as recommended by Facebook) .

What kind of og:type (Object Types) you should use can always be discussed. In the example code above I'm using "article" as Object Type, which is recommended for a blog post like this. However, if you use a different Object Type, Facebook Insight will show you deeper information about "Likes" for that particular page, and you can post follow up messages on Facebook to those who have liked the page.

Facebook XFBML version (Javascript SDK)

Remove formatting from Code

<div id="fb-root"></div>
<script type="text/javascript">
window.fbAsyncInit = function() {
FB.init({appId: 'YourAppID',
status: true,
cookie: true,
xfbml: true,
channelUrl : 'http://www.yourdomain.com/facebook/include/channel.html' // add channelURL to avoid IE redirect problems
}); };
(function() {
var e = document.createElement('script'); e.async = true;
e.src = document.location.protocol +
'//connect.facebook.net/en_US/all.js';
document.getElementById('fb-root').appendChild(e);
}());
</script>

Replace values in red with your values. The best place to put this code is right before the closing </body> tag according to Facebook.

Facebook channelURL

The contents of the channel.html file should be this single line.

Remove formatting from Code

<script src="http://connect.facebook.net/en_US/all.js" ></script>

Google Analytics Social Media Tracking Script for Facebook

Now that all the basic Facebook stuff is implemented, it is time to implement the script that will track Likes, Share, Send & Comments as Social Media in Google Analytics.

Replace values in red with your values. I also recommend that you don't implement this script as inline javascript, but instead save the script into a javascript file and point to that file to improve page speed.

Place this code below all your other Facebook scripts (before </body>).

Remove formatting from Code

The script above is inspired from Google Analytics and from Yoast. In addition I have also added tracking of Facebook Comments.

Yoast's version tracks interaction with Facebook Like & Send buttons (this will be tracked as "Share" or "Send"), and with the Facebook Like Box (this will be tracked as "Like").

Now that all the basic code and tracking scripts are in place, it's time to implement buttons, boxes or comment fields on your site.

Facebook Share & Send Button

Remove formatting from Code

Facebook Like Box

Remove formatting from Code

<fb:like-box href="http://www.facebook.com/Savio.no" width="265" show_faces="true" stream="true" border_color="#ccc" header="false" ref="facebook_fb-page"></fb:like-box>

Facebook Comment Box

Remove formatting from Code

<fb:comments width="500" num_posts="2" href="http://domain.com/canonical-url-to-article"></fb:comments>

Just so I have mentioned it, if you comment on your own site, your comments will not be tracked in Google Analytics.

See also information on Facebook about the Comment Box.

If you see anything that could be improved or errors (I'm not a programmer), or if you just found this article valuable, feel free to comment, Like, Tweet or G+.

Happy tracking Facebook Likes, Share, Send & Comments as Social Engagement in Google Analytics.

How to Track Traffic from Facebook Likes, Send or Comments in Google Analytics

Eivind Savio — 2011-07-12

OK, so you have implemented Facebook Like, Send or Comments on your website, and you see from your referral data in Google Analytics that you are getting traffic from Facebook. But are that traffic caused by Like Buttons or Comments? Referral data will not show you this.

Image by Ksayer1

In this blog post I explain how you can get better insight into your Facebook traffic by tracking traffic from Facebook Likes, Send or Comments as campaign traffic - and of course, how to do it.

The road to better insight

Unfortunately there isn't a easy road to what I consider to be the best solution for tracking Facebook Like, Send or Comments (I have tried several approaches). Here are what you will have to do:

(Probably) change your Facebook Like & Send script
Change your Google Analytics script
Add filters to Google Analytics
Exclude Facebook Parameters from your Google Analytics reports
Beware of SEO implications that Facebook Parameters may cause

Facebook Like & Send

To the Facebook Like & Send script we will add something called the "ref attribute". The purpose of the ref attribute is originally to be able to help you test different placements and types of Like buttons. Append the ref="" attribute to the Like button, and make sure that the value you choose is less than 50 characters (a-z, A-Z, 0-9, + / = - . : _).

The ref attribute is going to be our Google Analytics Campaign Name. Since we also are going to use the ref attribute in our Google Analytics Filters, part of the ref attribute must contain something that makes us able to identify the value as "Facebook". Se example below:

Remove formatting from Code

<fb:like href ="http://domain.com/canonical-url-to-article" send="true" layout="button_count" show_faces="false" font="arial" ref="facebook_your-campaign-name"></fb:like>

When somebody "Like" or "Send" your article, the parameter fb_ref is automatically added to the URL. This makes us able to identify this as either Facebook Like or Send traffic as the example URL below shows.

http://www.domain.com/url-to-article?fb_ref=facebook_your-campaign-name

Where on Facebook did the Link Get Clicked?

When you use the ref parameter, Facebook will also (but not always) add a fb_source parameter to the referrer URL, which includes the stream type ('home', 'profile', 'search', 'other') where the click occurred and the story type ('oneline', 'multiline') concatenated with an underscore.

The fb_source parameter will be tracked as Campaign Ad Content in Google Analytics.

Facebook Comments

You don't have to change anything in your Facebook Comments code. Example code below:

Remove formatting from Code

<fb:comments width="500" num_posts="2" href="http://domain.com/canonical-url-to-article"></fb:comments>

When a user comments and posts to Facebook a fb_comment_id parameter will automatically be added to the URL like this:

http://domain.com/url-to-article?fb_comment_id=commentidparameter123

The fb_comment_id parameter will be tracked as Campaign Ad Content in Google Analytics.

The Google Analytics script

To be able to track traffic from Facebook Like, Send or Comments as campaign traffic you will have to add Google Analytics Campaign Tracking Code to your Google Analytics script. In the example script below the Campaign Tracking Code is bolded.

Since the script below is handling both Like/Send and Comments, I have added a simple javascript checking for either the fb_ref or the fb_comment_id parameter in the URL. You can also do this server-side as I do on this site. Just change the script part in orange.

Remove formatting from Code

<script type="text/javascript">
// This script is provided AS IS without warranty of any kind.
// See http://www.savio.no/blogg/a/107/ for more information.
// Smacked together by Eivind Savio July 2011
var _gaq = _gaq || [];
_gaq.push(['_setAccount', 'UA-XXXXXX-X']);
if (document.location.href.search("fb_ref=")!=-1){
// Track traffic from Facebook Like & Send as Campaign traffic
_gaq.push(
['_setCampMediumKey', 'fb_ref'],
['_setCampSourceKey', 'fb_ref'],
['_setCampNameKey', 'fb_ref'],
['_setCampContentKey', 'fb_source']
);
} else if (document.location.href.search("fb_comment_id=")!=-1){
// Track traffic from Facebook Comments as Campaign traffic
_gaq.push(
['_setCampMediumKey', 'fb_comment_id'],
['_setCampSourceKey', 'fb_comment_id'],
['_setCampNameKey', 'fb_comment_id'],
['_setCampContentKey', 'fb_comment_id']
);
}
_gaq.push(['_trackPageview']);
_gaq.push(['_trackPageLoadTime']);
}
(function() {
var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true;
ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s);
})();
</script>

Google Analytics Filters for Facebook Likes, Send & Comments

Without filtering the data in Google Analytics, our Campaign Name, Source and Medium reports will look like rubbish.

Facebook Like, Send & Comments Campaign Medium Filter

Field A -> Extract A	Campaign Medium	^facebook_\|^fbc_
Field B -> Extract B	-
Output To -> Constructor	Campaign Medium	social-media

Field A Required	Yes
Field B Required	No
Override Output Field	Yes
Case Sensitive	No

Facebook Like, Send & Comments Campaign Source Filter

Field A -> Extract A	Campaign Source	^facebook_\|^fbc_
Field B -> Extract B	-
Output To -> Constructor	Campaign Source	facebook.com

Field A Required	Yes
Field B Required	No
Override Output Field	Yes
Case Sensitive	No

Facebook Comments Campaign Name Filter

Field A -> Extract A	Campaign Name	^fbc_
Field B -> Extract B	-
Output To -> Constructor	Campaign Name	facebook-comment

Field A Required	Yes
Field B Required	No
Override Output Field	Yes
Case Sensitive	No

Done! Now traffic from Facebook Like, Send or Comments will be tracked as campaigns in Google Analytics and will give you better understanding of your Facebook traffic.

Google Analytics Report Explanation

As mentioned in the beginning of this blog post the parameter fb_ref will automatically be added to the URL on Facebook.

Unfortunately not all traffic caused by a "Like" will have that parameter. The exception is when somebody "Like" an article, and then somebody on Facebook reshare that article. From my experience the posted URL on Facebook will then loose the fb_ref parameter, and traffic from the "Liked" URL will from now on be tracked as referral traffic in Google Analytics.

With other words, this solution isn't perfect, but I found it to be the best method.

Facebook Campaign Name Report in Google Analytics

Below you find an example of how the Campaign Name Report may look like.

In the example above you will see that I have differentiated between "Likes" in left column or below article (I have 2 sets of Like buttons on my sites for the moment). Perhaps not the best campaign names, but I know this traffic is caused by either Facebook Likes, Send or Comments.

Facebook Campaign Ad Content Report in Google Analytics

Below you find an example of how the Campaign Ad Content Report may look like.

In the beginning of this blog post I shortly explained that the fb_source parameter can be added to your URLs. If your traffic comes from a Facebook Like button, the fb_source parameter will be added.

If your traffic comes from a Facebook Send button, the fb_source parameter is message, but I have also experienced that it can be missing. The report above shows some traffic reported as (not set). That is in this case traffic from the Send button.

The Ad Content Report containing the fbc_xxxx value is traffic from Facebook Comments, and the value is the Comment ID.

Cleaning up your data

The parameters from Facebook will create duplicate content in your page content reports in Google Analytics. I recommend you exclude those parameters.

Some final recommendations

Implementing Facebook scripts on your website aren't necessary so easy as it may look like. Facebook scripts may slow down your site, cause problems for Internet Explorer, and finally, cause problems for search engines like Google (and with this, cause SEO problems).

These are my recommendations when you implement Facebook scripts on your website (the 5 first points are covered in my blog post Tracking Facebook Likes, Share, Send & Comments as Social Engagement in Google Analytics):

Use the HTML5 solution Facebook XFBML and not the iFrame solution.
Load your Facebook XFBML scripts Asynchronously.
Implement Open Graph protocol
Use Facebook Channel URL to reduce problems with Internet Explorer.
Implement Google Analytics Social Interaction Analytics
Implement Canonical tag or exclude Facebook parameters in Webmaster Central to avoid SEO problems.

If you see anything that could be improved or errors (I'm not a programmer), or if you just found this article valuable, feel free to comment, Like, Tweet or G+.

Happy tracking Traffic from Facebook Likes, Send or Comments.

How to Track Facebook iFrames with Google Analytics

Eivind Savio — 2011-05-16

When Facebook launched Facebook iFrame Tabs for Pages, tracking visitor interactions, goals or whatever inside the iFrame in for example Google Analytics suddenly seemed very easy. A common answer in forums when people asked about this was "just add your regular Google Analytics code to your iFrame page".

Unfortunately things aren't that easy. If you just add your regular Google Analytics script to your iFrame page, you are missing 2 important things:

Tracking traffic sources to your Facebook iFrame Tab is not possible
- Facebook will be the referral for all the traffic to your Facebook iFrame Tab
Visitors using Internet Explorer will not be tracked

In this blog post I show you how these problems can be overcomed.

Handling the Traffic Source Problem

In an earlier blog post LunaMetrics explained how tracking Traffic Sources to the Facebook iFrame Tab Page could be done (I recommend you read the article from LunaMetrics since they also explains some other important things). The method is to send the visitor to a page outside Facebook that will track the traffic source, and then redirect the visitor to the Facebook iFrame Tab Page.

However, if you just redirect the visitor, Google Analytics will probably not track the visit since the Google Analytics script will not be able to load before the redirect occurs. In other words, you must use a delayed redirect.

The delay can either be "wait 1 second, hope Google Analytics has loaded, and then redirect". Or as I do in this script, check if Google Analytics probably has loaded, and then redirect (or frame the page as I do).

Since I'm framing the page, I don't have to create an additional redirect page. Everything is handled by the script in the Facebook iFrame Page.

As an example, I have created a Facebook iFrame Tab Page at this adress:

http://www.facebook.com/Savio.no?sk=app_180487048663627

If I want to track the traffic source to that page, I will link to my iFrame Page and not to the Facebook iFrame Tab Page.

My Facebook iFrame Page can be found at this adress:

http://www.savio.no/facebook/

If you would track campaign traffic to the Facebook iFrame Page (ex. promote your Facebook Page in a newsletter), then you should add campaign tracking to the URL like this:

http://www.savio.no/facebook/?utm_source=newsletter&utm_medium=email&utm_campaign=yourcampaign

In regards of campaign tagging, I really recommend my own Google Analytics Campaign URL Tool. :)

Google Analytics tracking code for Facebook iFrame Tab Pages

Below is the Google Analytics code I'm using to track Facebook iFrame Pages. The code handles:

Campaign tracking
Delayed redirect if the visitor is landed on the iFrame Page
- The code is checking if _gaq.push is loaded before it redirect the user to the Facebook iFrame Tab

Change values in red, and add the code between <head> and </head> to your iframe page.

Remove formatting from Code

<script type="text/javascript">
// This script is provided AS IS without warranty of any kind.
// See http://www.savio.no/blogg/a/104/ for more information.
// Smacked together by Eivind Savio May 2011
var FacebookURL = 'http://www.facebook.com/YourFacebookPage?sk=app_123456789';
var _gaq = _gaq || [];
_gaq.push(['_setAccount', 'UA-XXXXXX-X']);
_gaq.push(['_addIgnoredRef', 'static.ak.facebook.com']);
// If Google Analytics is loaded and this page isn't framed inside our Facebook URL, frame the page.
if (_gaq.push && self.location == top.location) {
// Create a fake URL that we can filter in Google Analytics. This pageview also sends the traffic source data to Google Analytics.
_gaq.push(['_trackPageview','/facebook/campaign-tracking/']);
// Tracking done, let's frame the page
setTimeout(top.location.href = FacebookURL, 200);
} else if (self.location == top.location) {
// If Google Analytics doesn't load within 2 seconds, refresh and frame the page anyway. People dislike waiting. Change the time if you like.
setTimeout(top.location.href = FacebookURL, 2000);
} else {
// The Page is Framed. Let's track it.
_gaq.push(['_trackPageview']);
_gaq.push(['_trackPageLoadTime']);
}
(function() {
var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true;
ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s);
})();
</script>

Why Internet Explorer Privacy Settings can make your life miserable

To be able to make things work in Internet Explorer can be a nightmare, whether it is CSS, javascript or in this case - tracking Facebook iFrame Tab Pages in Google Analytics.

The problem isn't related to Google Analytics, you will get the same problem with other web analytic tools. The problem is related to Internet Explorer and handling of cookies from a third party website inside an iFrame.

If the users Privacy Settings in Internet Explorer is set to "Medium" or higher, Internet Explorer will block cookies from your iFrame Page and Google Analytics, and the visit will not be recorded.

The solution: P3P (Platform for Privacy Preferences)

To get around this problem you will have to send a P3P (Platform for Privacy Preferences) Header in your iFrame Page.

A P3P Header may look something like this:

P3P: CP="NOI DSP COR CURa ADMa DEVa TAIa OUR BUS IND UNI COM NAV INT"

What you add as a CP value doesn't really matter. CP="Eivind is Cool" will also work. So where do the quoted CP values come from?

I had a really hard time before I was able to get P3P work in Internet Explorer. On my way to a working solution I came across the P3P Policy Editor from IBM. This editor has generated the values for my P3P Header based on the information I added to my P3P XML file.

I have generated 2 files:

P3P reference file: http://www.savio.no/w3c/p3p.xml
- The reference file should always be located at the adress /w3c/p3p.xml
Full P3P policy: http://www.savio.no/w3c/full-P3P-policy.xml

Just to be specific, creating and uploading P3P XML files isn't necessary, but if you want to show your Internet Explorer visitors your Privacy Preferences upload these files.

Image: P3P XML File missing.

Image: P3P XML File Reference in place.

Below you find code examples of how to send a P3P Header in ASP.NET, ASP and PHP.

P3P Header for ASP.NET using global.asax

ASP.NET implementation may look like the following:

Remove formatting from Code

protected void Application_BeginRequest(Object sender, EventArgs e)
{
//
HttpContext.Current.Response.AddHeader("p3p", "CP=\"IE is often a NIGHTMARE\"");
}

P3P Header for "old" ASP

ASP implementation may look like the following (implement this as the first code in your .asp iFrame page):

Remove formatting from Code

<% Response.AddHeader "P3P","CP=""IE is often a NIGHTMARE"""%>

P3P Header for PHP

PHP implementation may look like the following (implement this as the first code in your .php iFrame page):

Remove formatting from Code

<?php
header('P3P: CP="IE is often a NIGHTMARE"');
?>

P3P Header for PHP using .htaccess

PHP implementation may look like the following (source):

Remove formatting from Code

<IfModule mod_headers.c>
Header set P3P "policyref=\"/w3c/p3p.xml\", CP=\"IE is often a NIGHTMARE\""
# OR THIS, SIMPLER
Header set P3P "policyref=\"/w3c/p3p.xml\""
</IfModule>

Source: Manipulating HTTP Headers with htaccess

The easiest method to identify if your P3P Header is sent, and if Google Analytics is tracking the visit in Internet Explorer, is to install ieHTTPHeaders.

If you see anything that could be improved or errors (I'm not a programmer), or if you just found this article valuable, feel free to comment.

Happy tracking your Facebook iFrame Page.

How to Track Multiple Links to the same URL automatically in Google Analytics

Eivind Savio — 2011-02-05

If you have multiple links on a page pointing to the same URL, you are not able to identify in Google Analytics which link were clicked. Neither the In-Page Analytics report or the Navigation Summary report (or what somebody calls a Next Page report) will show you this.

The usual method to overcome this problem is to manually tag links, which is a method that often requires programmers (and may cause SEO problems and duplicate content if done wrongly).

Because of this I have created a script that is easy to implement and configure. The script will automatically track which link the user clicked on, the Anchor Text of the link and the Link ID (number). The information is stored as Events in Google Analytics.

And as a "bonus" I have included a Excel Spreadsheet using the Google Analytics API that makes reports more readable.

How to Automatically track Multiple Links to the same URL

First, download the script (ZIP-file). Within the ZIP-file you find 2 javascript files:

adddomloadevent.min.js
ga-track-link-clicks.js

Open up ga-track-link-clicks.js and configure the settings (shown below).

What	Explanation
pageName	The name of the page you are going to analyze. 3 different options to identify the page: "" (empty) = Page URL will be used "title" = Title Tag of the page will be used "Your custom text" = Use this option if URL and Title Tag sucks, and if you are going to analyze just a single page. Title Tag or Custom Text should be your preferred option because of the Excel spreadsheet report functionality.
trackAnchorText	Use Anchor Text of the link to identify which link the user clicked on. If the link is a image and not a text link, the image URL will be used instead. If Anchor Text is missing, the text "Anchor Text Missing" will be used. true = Track Anchor Text false = Do not track Anchor Text
trackLinkID	Sometimes there can be several links on the same page pointing to the same URL with the same Anchor Text. In these occations the Link ID can help you to differentiate between the links. true = Track Link ID false = Do not track Link ID The script automatically counts links on the page. The first link will get number 1, second link will get number 2 etc.
trackExternalLinks	Should links to external sites also be tracked? true = Track External Liks false = Do not track External Links
trimPageName	Should pageName be trimmed for special characters? true = Trim (remove) special characters false = Do not trim special characters

Below are the settings in the javascript file shown. In the example the

Title Tag will be used to identify the page (pageName).
Link Anchor Text and Link ID will be used to identify which link the visitor clicked on.
We are not tracking External Links.
Special charachters are removed from the pageName.

/************************************
SETTINGS
************************************/
var pageName = "title"; // Name of the page you want to track/analyze. title/custom/url
var trackAnchorText = true; // Track link Anchor Text? true/false
var trackLinkID = true; // Track Link ID? true/false
var trackExternalLinks = false; // Track External links? true/false
var trimPageName = true; // Should pageName be trimmed for special characters? true/false

Implementing the script

The script is loaded as soon as the DOM has finished loading using addDOMLoadEvent.

The adddomloadevent.min.js and ga-track-link-clicks.js file must be implemented just after your Google Analytics Asynchronous script (between <head> and </head>):

** Google Analytics Asynchronous tracking script **
<script src="/path/to/js/adddomloadevent.min.js" type="text/javascript"></script>
<script src="/path/to/js/ga-track-link-clicks.js" type="text/javascript"></script>
</head>
<body>

The Link Click Report

Event	Value/Description
Event Category	Link Clicks
Event Action	URL/Title Tag or Custom Name for the tracked page
Event Label	http://www.domain.com/url-to-page\|Anchor Text\|LinkID like this: http://www.savio.no/article-name/\|Read More\|64

Below is a screenshot showing the Event Report in Google Analytics with data for Clicked Links, Anchor Text, Link ID and Time spent before Click.

Because the Event Label report in Google Analytics can be a little bit confusing to analyze, I have created an Excel Spreadsheet that splits the Event data into different columns and makes the data more readable.

The Excel Spreadsheet is using the Google Analytics API.

Download the Excel Spreadsheet (ZIP-file) and unzip the file
Open the Excel file and activate macros.
Go to the Settings sheet. Change "Sites" from YourWebsiteName to your site name.
On the same line as your site name, insert your Google Analytics ID number. This is not your UA-XXXX-X ID, it's the ID in the Google Analytics URL (ex. https://www.google.com/analytics/reporting/dashboard?id=123456)
Insert your Google Analytics username (email) and password in the Settings sheet.
Select date period on the Dashboard. If you click on a date cell, a date picker will pop up to make it easier to select dates.
Click Update Report.
Done.

The report works best in Excel 2007 or newer because of Conditional Formatting.

Visualising the data

Since the data isn't very visual, I like to create a kind of "click map", or more correct, a link color map in Excel. The first thing I do is to grab all the links from the page I want to analyze, and insert them into Excel (URL, ID and Anchor Text) as a reference.

Then I use different background color for links belonging to the menu, content, footer etc. using conditional formatting in my report. I have scripted Excel to do this automatically based on the Link ID. This makes it easier to visualise which part of the page the different links belongs to. Filtering by colors can also be very useful since you now can isolate different parts of the page.

The (censored) screenshot above, shows an example where I'm using Link ID to automatically identify the behaviour for different page areas for visitors that have converted. I can now attach outcome (money!) to page areas, links and message (Anchor Text). I've also found it easier to identify different buying modalities using this method.

This method will only work well on pages were the number of links don't change.
NB! The download doesn't include this functionality, but I show the screenshot as an example of how you can improve the spreadsheet.

Why not just use a Link Analysis/Heat Map tool?

The reason I made this solution in the first place was because I had a case where I wanted to know the outcome of the click, not just which links were clicked and how much they were clicked. In addition, I wanted a solution that was easy to implement for Google Analytics so I could avoid implementation of more tools in this specific case.

The Event Tracking report in Google Analytics will show you clicks tracked in connection to Ecommerce values like transactions, revenue and conversion rate.

If your outcome is not tied to Ecommerce, but regular conversion goals, you can find number of conversion by using Advanced Segments. There is already a predefined Advanced Segment called "Visits with Conversions" in Google Analytics.

So even if tools with Click Maps/Heat Maps reports will show you better click information (ex. if visitors clicks in a area they thought were a link), this solution makes it easy to track and connect different link clicks to outcome in Google Analytics.

What to be aware of

Although the script has built in some functions for improving reports like stripping away HTML in the Anchor Text, this is a "quick and dirty" method for identifying link clicks. Below are some points you should be aware of.

If the Page Name (Event Action) you are going to analyze contains some special characters, or you are using URL to identify the page, the Excel spreadsheet may not work properly. I have experienced problems with Norwegian characters, & and / to mention some.

The fix for this is to set trimPageName to true (var trimPageName = true) in the javascript file.

The Javascript

I'm not a programmer, so this solution is maid mainly by banging my head against the wall.

If you study the javascript in ga-track-link-clicks.js, you will see that I track clicks using onmousedown and not with onclick. The reason for this is because if I use onclick in the script, other onclick events may be neutralized. Using onmousedown will not affect scripts using onclick. Since fewer things are tracked using onmousedown, the chance for affecting other scripts are lower.

With this in mind, test after you have implemented the script to see if other scripts are affected.

Link ID

If your content changes or the number of links on the page changes, some of the links will get a different ID.

Anchor Text

"Bad" HTML may affect the report. If a image is the "link anchor", src (source) to the image file has to be in single or double quotes to be tracked properly. Ex.

src="/folder/image.jpg" and src='/folder/image.jpg" will be tracked like IMG-/folder/image.jpg

Sometimes the image anchor text will not be tracked by the URL. Instead the content of the Alternate Tag are used. Although it's not perfect, I can live with that. Feel free to improve the regular expression that grabs the image info.

Visitors using Google Translate or similar will also affect the tracked Anchor Text.

Bounce Rate and Time Spent

Time spent on page/site can be affected by this script, especially if you set trackExternalLinks to true.
Bounce Rate will be affected especially if you set trackExternalLinks to true. Since the script will track all type of link clicks, like file downloads and anchor links, this will affect (reduce) your bounce rate as well.

Download the Google Analytics Script and Excel Report

Download the Google Analytics script (ZIP)
Download the Excel Spreadsheet Report (ZIP)

Happy tracking, analyzing and optimization of link clicks and outcome.

If you find some bugs I haven't seen, improves something, or just find the solution useful, feel free to drop a comment below.

Website Optimizer Asynchronous Integration with Google Analytics

Eivind Savio — 2010-11-07

Best practices for running a testing experiment is always to integrate the test data into your web analytics tool for further analysis.

Below you find my code examples for a multivariate test where the different Website Optimizer variations are tracked as Custom Variables in Google Analytics.

This method is a rewritten version of An Update to the Erik Vasilik's "Poor Man's GWO/Analytics Integration": Using Custom Variables. Codes for A/B testing is different, and that is also explained in the article.

Although I have used this code successfully in testing, I really recommend that you Test your test first to see if there are something wrong with the codes or the implementation.

Test Page Code

Below you find the scripts for the test page. The Google Website Optimizer Control script isn't changed from the original script, but the Google Website Optimizer Tracking script is different. Because of this the validation within Website Optimizer will fail. Just ignore the errors shown there.

Replace XXX with your GWO and GA ID codes/numbers.

Remove formatting from Code


<script>
function utmx_section(){}function utmx(){}
(function(){var k='XXXXXXXX',d=document,l=d.location,c=d.cookie;function f(n){
if(c){var i=c.indexOf(n+'=');if(i>-1){var j=c.indexOf(';',i);return escape(c.substring(i+n.
length+1,j<0?c.length:j))}}}var x=f('__utmx'),xx=f('__utmxx'),h=l.hash;
d.write('<sc'+'ript src="'+
'http'+(l.protocol=='https:'?'s://ssl':'://www')+'.google-analytics.com'
+'/siteopt.js?v=1&utmxkey='+k+'&utmx='+(x?x:'')+'&utmxx='+(xx?xx:'')+'&utmxtime='
+new Date().valueOf()+(h?'&utmxhash='+escape(h.substr(1)):'')+
'" type="text/javascript" charset="utf-8"></sc'+'ript>')})();
</script>

<script type="text/javascript">
var _gaq = _gaq || [];
_gaq.push(['_setAccount', 'UA-XXXXX-1']);
// GWO code
_gaq.push(['gwo._setAccount', 'UA-XXXXX-2']);
_gaq.push(['gwo._trackPageview', '/XXXXXXXX/test']);
// End GWO code
var gwoVar = utmx("combination_string");
if (!gwoVar) {
gwoVar = "inactive";
}
_gaq.push(['_setCustomVar',1,'GWO_Test-Name', gwoVar, 1]); // Track GWO as Custom Variables in GA
_gaq.push(['_trackPageview']);
</script>
<script type="text/javascript">
(function() {
var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true;
ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s);
})();
</script>

Conversion Page Code

Remove formatting from Code


<script type="text/javascript">
var _gaq = _gaq || [];
_gaq.push(['_setAccount', 'UA-XXXXX-1']);
_gaq.push(['_trackPageview']);
// GWO Conversion code
_gaq.push(['gwo._setAccount', 'UA-XXXXX-2']);
_gaq.push(['gwo._trackPageview', '/XXXXXXXX/goal']);
// End GWO conversion code
</script>
<script type="text/javascript">
(function() {
var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true;
ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s);
})();
</script>

Page Sections

The Page Sections script in Website Optimizer is not changed. Below you see an example of this script.

<script>utmx_section("Headline")</script>
<h1>Buy This Product!</h1>
</noscript>

Successfull Web Site Testing Practices

The technical implementation and running the test isn't difficult. But that is also a reason for failing in testing, processes and hypothesis are missing. If you haven't done so before, I highly recommend to establish testing processes. Eric T. Peterson's 10 tips for successful testing shown below is a good start.

Free Excel Goal Funnel Report for Google Analytics

Eivind Savio — 2010-10-11

Have you ever came across Google Analytics accounts where Goal Funnels are missing or wrongly setup, and been wondering how you could analyze and track these missing goals easy?

Or you have been staring at the Goal Funnel Visialization report in Google Analytics and thought "Why can't I use Advanced Segmentation for this report?".

Or perhaps you just need to create and analyze conversion funnels on the fly?

My answer to all these problems was to create a Excel Spreadsheet using the Google Analytics API, which means that you can log into Google Analytics inside Excel and do Goal Funnel Tracking on the fly.

My Excel Goal Funnel Report for Google Analytics have the following functionality:

Funnels can be up to 10 steps.
URLs are matched using Regular Expressions.
Segmentation is supported.
The graph is dynamic, and shows visits per step and how many visits that came from the previous step.
Important data about your funnel are listed below the graph:
- Number of Conversions
- Funnel Conversion Rate
- Funnel Abandonment Rate
- Website Conversion Rate
- Time spent per step, average pageviews per step, entrances per step, exits per step

How to use the Google Analytics Goal Funnel Report for Excel.

Date Picker

Download the Excel Spreadsheet (ZIP-file) and unzip it
Open the Excel file and activate macros. (You may get an error the first time you open the file since the file doesn't contain any data. Just ignore the error).
Select date period on the Dashboard. If you click on a date cell, a date picker will pop up to make it easier to select dates.
Go to the Settings sheet. Change "Sites" from YourWebsiteName to your site name, and name your Goal.
On the same line as your site name, insert your Google Analytics ID number. This is not your UA-XXXX-X ID, it's the ID in the Google Analytics URL (ex. https://www.google.com/analytics/reporting/dashboard?id=123456)
Insert your Google Analytics username (email) and password in the Settings sheet.
Activate Regular Expressions for Excel (see explanation below).
Go to the Goal Path sheet and set up your conversion funnel.
- Name your funnel steps and add your URLs.
- Your URLs must be matched using Regular Expressions.
Go to the Dashboard sheet, click Update Report.
Done.

The Google Analytics Goal Funnel Report works best in Excel 2007 or newer because of Conditional Formatting.

Image showing Goal Path setup. URLs are matched using Regular Expressions. Up to 10 steps can be added.

Selecting "Microsoft VBScript Regular Expressions"
from References.

Excel and Regular Expressions

Excel's default setup doesn't normally include the Regular Expression "engine", and has to be activated.

To activate support for Regular Expressions do the following:

Hit Alt+F11.
From the menu, choose Tools -> References.
Select Microsoft VBScript Regular Expressions 5.5.
Click OK.
Done.

Fetching data from Google Analytics with Excel

Excel is fetching data from the Google Analytics API using VBA coding. These codes are written by Mikael Thuneberg, and he has explained on his blog how to use the codes. You will also find other Excel examples there.

Some closing comments about the Goal Funnel Report

This Excel Goal Funnel Report for Google Analytics is what I would call a Beta version, and I'm not a programmer. Therefor, if you find any errors, or think that something could be improved, please let me know.

Download the Excel Goal Funnel Report for Google Analytics

Excel Feedback Form Reporting using Google Analytics API

Eivind Savio — 2010-09-27

Recently I released my Free Feedback Form for Google Analytics. Although it is pretty easy to drill down into the Feedback Form data inside Google Analytics, exporting the data and putting together a useful report can take some time.

Because of this I have made a Excel Spreadsheet using the Google Analytics API, which means that you can log into Google Analytics inside Excel, and your Feedback Form Report will be automatically generated.

Excel Feedback Form Reports

The Excel sheet contains 3 reports.

Feedback Form Report Overview (Dashboard)
Qualitative Feedback Report
Search Phrases and Feedback Score

Feedback Form Report Overview (Dashboard)

The Feedback Form Dashboard gives you an overview of which pages visitors have given feedback on, how many votes for each page, total average score and the average score for each page.

On the Dashboard sheet you also select Date Period and which Question you want reports for.

Qualitative Feedback Report

The Qualitative Feedback Report sheet shows visitor comments (and hopefully Why visitors has rated your page good or bad), which page the comment was given on and page rating.

Search Phrases and Feedback Score

Not all visitors will tell you why they rated a page/your site good or bad. Although my Feedback Form supports settings for Mandatory Commenting, I would normally not force visitors to comment.

Analyzing Search Phrases in combination with Ratings can give you a clue about why visitors rated your pages like they did. For example, ranking and getting traffic for the wrong search phrases can result in bad ratings.

How to use the Excel Feedback Form Report

Download the Excel Spreadsheet (ZIP-file) and unzip it
Open the Excel file and activate macros.
Select date period on the Dashboard.
Go to the Settings sheet, and change "Sites" from YourWebsiteName to your site name.
On the same line as your site name, insert your Google Analytics ID number. This is not your UA-XXXX-X ID, it's the ID in the Google Analytics URL (ex. https://www.google.com/analytics/reporting/dashboard?id=123456)
Insert your Google Analytics username (email) and password in the Settings sheet.
Done. Your report is now generated.

To the right of each report you will see which Google Analytics API metrics and dimensions that are being used. Example:

Sort results by Votes:	TRUE
Metrics:	eventValue&totalEvents
Dimensions:	pagePath

As standard reports are sorted from highest to lowest value. If you change TRUE to FALSE the report will be sorted by URL name instead. Pages with a low score will be highlighted with red text (Excel 2007).

The Feedback Form Report works best in Excel 2007 because of the Conditional Formatting, but the other functions are working in Excel 2003. I haven't tested other Excel versions.

Fetching data from Google Analytics with Excel

Speedometer/Gauge Chart

Although it can be debated how useful Speedometer Charts are for presenting data, in this context I think this type of chart is doing the best job.

The Speedometer Chart I have used is made by Chandoo.org.

Some final words about this site and how to comment

If you look around on my site you have probably noticed that everything else is written in Norwegian. Since this blog normally is in Norwegian, and I haven't had time to implement a multilingual blog system, you find some explanations of how to comment below:

Click "Legg til kommentar" below. Than fill out your name (Navn) (Mandatory), e-mail (Epost) (Mandatory) and your website (Ditt nettsted) if you have a website. HTML formatting doesn't work in my comment field, use BBCode formatting instead. You can drag the comment field around if you want to. If you want to preview what you have written, click on "Forhåndsvis". When you are done and want to save your comment, click on "Lagre".

If you haven't had any comments approved on this blog before, I will have to approve your comment manually.

Free Feedback Form for Google Analytics

Eivind Savio — 2010-09-12

Feedback Form with Horizontal Radio Button Layout and Comment Field. Text in Comment Field will be stored as Event Labels.

Feedback Form with Vertical Radio Button Layout and Comment Field. Text in Comment Field will be stored as Event Labels.

Feedback Form with Vertical Radio Button Layout and Comment Field disabled. Text supporting selected Radio Buttons will be stored as Event Labels.

Feedback Forms on web sites are getting more and more popular - and for one good reason. Web Analytics including Voice of the Customer, or Why as Avinash Kaushik calls it, can be some of the most valuable data collection and analysis you can do.

But, combining What data with Why data is even better.

Because Why + What can give so much more insight, I have made a Non-intrusive Feedback Form that can be easy customized and sends data directly into Google Analytics using Event Tracking.

Feedback Form Customization and Functions:

Ask for Feedback based on Page Level or Site Level
- Only one feedback per visitor for either Site or Page Level
- Page Level: The visitor can comment and rate different pages on your site, but not rate and comment the same page twice. After commenting the Feedback Form will be disabled for the particular page.
- Site Level: The visitor can only comment and rate once. After commenting the Feedback Form will be disabled on all pages.
Use colors that you prefer for text, background, borders and submit button.
- Both color names and hexadecimal colors can be used (ex. black or #000000).
Decide all important text on the Feedback Form like "Thank You" Messages, Error Messages, text on Submit Button etc.
Width of the Feedback Form can be changed.
Placement (left or right) can be changed
Comments can be either Mandatory or Optional.
Turn on/off tracking of error messages.
Set how many days the comment form should be disabled after feedback has been submitted.
Possibility to track feedback as conversions in Google Analytics.
Visitors can hide/show Feedback Form with a click.
Because feedback is submitted directly into Google Analytics it's easy to drill down into your data and
- find out which URL the visitor was on when feedback was given
- find out which traffic source the visitor came from or if the search phrase the user arrived from can be a reason for a negative feedback.
- etc.

If you want even more customization for this Free Feedback Form, also check out my article about How to use BTBuckets to Administrate the Google Analytics Feedback Form afterwards.

Google Analytics Event Report

Data from the Feedback Form will be reported as Events in Google Analytics. The principle is shown below.

Event Category (Name)
- Feedback.
Event Actions
- Your Feedback Question Heading.
- Error message "Feedback Radio Button not Selected" (optional).
- Error message "Feedback Text not filled Out" (only when comment field is mandatory).
- Feedback Opened (optional).
- Feedback Closed (optional).
Event Label
- Input Text in Comment Form
  - If Comment Field is disabled and Vertical Radio Button layout is selected, text that supports the Radio Buttons will be stored as Event Labels instead.
Event Value
- Value from Radio Button. The first Radio Button (left or top) will always have value 1.

If the Comment Field is activated, it's best to use a graded Radio Button scale (ex. Bad -> Good), with the best option as the last option.

The reason for storing Error Messages as Event Actions and not Labels, are simply because storing Error Messages as Labels would mess with your reported average Event Values.

The screenshot below shows an example where I have drilled down to find out which page the user was on when the question "How useful was this page?" was answered.

Some users were happy, others not. Without my Feedback Form I wouldn't know that, and I wouldn't know Why.

How to install and use the Feedback Form

There are 2 methods for installing the Feedback Form

Download the feedback form solution (ZIP-file), unzip the files and upload them to a folder on your website called feedback (ex. yourdomain.com/feedback)
Load the Feedback Form from my website

Option 1 should be your preferred solution. The reason for this is because if my server is slow, loading the Feedback Form from my website would also slow down your site.

Loading the Feedback Form from my website

Copy the code below, and paste it right before your closing </body> tag. Then adapt the settings in the code to your needs like changing question, options, colors etc.

Loading the Feedback Form from your own website

Copy the code below, and paste it right before your closing </body> tag.

Then change line 36 from
<script type="text/javascript" src="http://www.savio.no/feedback/ga-feedback-v1.js"></script>
to
<script type="text/javascript" src="/feedback/ga-feedback-v1.js"></script>

Then adapt the settings in the code to your needs like changing question, options, colors etc.

The red text is what you can change to get different settings. Happy collecting qualitative feedback directly into Google Analytics.

Remove formatting from Code

<script type="text/javascript">
// Text you can change
var feedbackHeading = "How useful was this page?"; // Feedback Heading
var feedbackCommentHeading = "Comment"; // Heading above your Comment field
var feedbackButtonText = "Submit"; // Text on Submit Button
var feedbackButtonTitle = "Give Feedback"; // Title Text on Submit Button
var feedbackHideText = "Hide Feedback"; // Title Text Hide Feedback
var feedbackShowText = "Show Feedback"; // Title Text Show Feedback
var feedbackCommentThx = "Thank You for Your Feedback."; // Thank you text after feedback is sent
var feedbackCommentGiven = "You have given Feedback."; // Text shown to those who have been given feedback on ex. reload of page
var feedbackRadioButtonError = "You must select a radio button"; // Error Text when Radio Button is not selected
var feedbackTypeText = "Bad,Good"; // Text supporting the radio buttons. Separate with comma (,). If you are using vertical aligned radio buttons, number of options must match "feedbackOptionsNumber".
// Settings you can change
var feedbackOptionsNumber = "5"; // Number of feedback options (radio buttons).
var feedbackRadioType = "horizontal"; // Should Radio Buttons be aligned "horizontal" or "vertical". (horizontal/vertical)
var feedbackPlacement = "right"; // Should the Feedback Form be placed "left" or "right" at the bottom of the screen
var feedbackCommentValidation = "optional"; // Is Comment "optional" or "mandatory"? (optional/mandatory)
var feedbackCommentVisibility = "show"; // Should Comment Field be shown or hided? (show/hide). If "hide", feedbackCommentValidation must be set to "optional" and feedbackRadioType to "vertical".
var feedbackTextColor = "black"; // Text Color
var feedbackBgColor = "white"; // Background Color
var feedbackBorderColor = "grey"; // Border Color
var feedbackButtonColor = "#333"; // Background Color on Submit Button
var feedbackButtonTextColor = "#fff"; // Text Color on Submit Button
var feedbackCharRestrictions = "200"; // Charachter Restrictions
var feedbackLevel = "page"; // Choose between "page" level or "site" level Feedback (page/site).
var feedbackTimeRestriction = "30"; // Number of days cookies should live
var feedbackWidth = "170"; // Width of Feedback Form in pixels
var feedbackGoalURL = "no"; // If set to "yes" a Fake URL will be generated when the form is successfully submitted so it can be tracked as a goal. URL: /feedback-form/feedback-submitted
var feedbackErrorTracking = "no"; // If set to "yes" Errors like "Not selected Radio Button" etc. and Open and Closing of the Feedback Form will be tracked.
</script>
<script type="text/javascript">document.write('<div id="fixeddiv">')</script>
<div id="feedbackHolder" class="feedbackshown"><a href="http://www.savio.no/blogg/a/97/free-feedback-form-for-google-analytics" title="Free Feedback Form for Google Analytics by Savio.no">Feedback Form for Google Analytics by Savio.no</a></div>
<script type="text/javascript" src="http://www.savio.no/feedback/ga-feedback-v1.js"></script>

As I mentioned earlier, if you want even more customization and easier handling of the Feedback Form, check out my article about How to use BTBuckets to Administrate the Google Analytics Feedback Form.

By using BTBuckets as an administration tool you:

Don't have to implement all the code above in your website
Don't have to upload new code to your website if you change ex. your question, color etc. You control everything from within BTBuckets
Can ask different questions to different audiences
- Only ask your returning visitors, or ask new and returning visitors different questions
Can turn the Feedback Form on and off without changing anything in your code
Only show the Feedback Form on selected pages
etc.

Just Remember

Even if this Feedback Form allows you to collect many types of feedback (it's up to you to ask the great questions), just remember this:

You are not allowed to track and store personal information in Google Analytics, so do not ask for the users e-mail, phone number etc. in the comment field.
A passive Feedback Form like this one tends to attract either very positive or negative feedback. You will probably not often get feedback from visitors that don't have a strong opinion.

Some closing comments about the Feedback Form

This Feedback Form is what I would call a Beta version and I'm not a programmer. I have tested the Feedback Form in many different browsers, but not all. I think the Feedback Form will break in Internet Explorer 6, but since my opinion is that IE6 should die I do not plan to fix that.

If I find some time I will try improve it even more depending on the popularity to the Feedback Form. For example, making it possible to load the Feedback Form Asynchronous for those who prefer to run the Feedback Form from my website is one thing I've been thinking of. For the moment the Feedback Form can't be loaded Asynchronous.

Any updates I do will be mentioned as comments to this blog post, so if you want to be updated about this I recommend you to subscribe to the RSS feed of this blog post.

The Feedback Form only works with the new Google Analytics Asynchronous Code.

If you find any errors, or think that something could be improved, please let me know.

Some final words about this site and how to comment

If you haven't had any comments approved on this blog before, I will have to approve your comment manually.

How to use BTBuckets to Administrate the Google Analytics Feedback Form

Eivind Savio — 2010-09-12

I have just released my Free Feedback Form for Google Analytics. Although I have built in lot of flexibility in the Feedback Form, it can be a little bit hassle to change settings or questions, and you can't ask different questions to different audiences if you want that.

In this blog post I'll explain how you can use BTBuckets as an Administration tool for the Feedback Form, which will give you much more flexibility and make it easier to change settings.

What is BTBuckets?

BTBuckets is a free personalization and on-site behavioral targeting tool for websites.

By using BTBuckets as an Administration tool for my Feedback Form you:

Don't have to implement all the code in your website.
Don't have to upload new code to your website if you change ex. your question, color etc. You control everything from within BTBuckets.
Can ask different questions to different audiences.
- Only ask your returning visitors, or ask new and returning visitors different questions.
Can turn the Feedback Form on and off without changing anything in your code.
Only show the Feedback Form on selected pages.
etc. It's up to you to release the power of BTBuckets.

Using BTBuckets as an Administration Tool for the Feedback Form

I will be using my own Feedback Form solution in BTBuckets as an example.

I am showing different questions depending of the visitors location. Visitors from Norway will be shown a Norwegian Feedback Form, while all others will be shown a English Feedback Form.

1. Add the Feedback Form code to your website

The 1st thing to do is to copy the code below, and paste it just before your closing </body> tag.

Remove formatting from Code

<script type="text/javascript">document.write('<div id="fixeddiv">')</script>
<div id="feedbackHolder" class="feedbackshown"><a href="http://www.savio.no/blogg/a/97/free-feedback-form-for-google-analytics" title="Free Feedback Form for Google Analytics by Savio.no">Feedback Form for Google Analytics by Savio.no</a></div>
<script type="text/javascript" src="http://www.savio.no/feedback/ga-feedback-v1.js"></script>

2. Register an BTBuckets account and Create your Bucket(s)

Head over to BTBuckets and sign up for their services (it's free). You will find many predefined "buckets" there, but using my example I have first created a bucket that includes "All Visitors".

Visitors that have seen 1 page will be bucketed as "All Visitors".

If you target that bucket, all visitors will see the same Feedback Form. I show my English Feedback Form to that target.

Bucketing Norwegian Visitors

Since I'm from Norway and most of my website is in Norwegian, my Norwegian visitors will get a Feedback Form in Norwegian.

The settings for that bucket is shown below (click on the image for a larger size).

Notice that I overwrite the "All Visitors" bucket. If I didn't do that both the English and Norwegian Feedback Form could be shown to my Norwegian visitors.

Next I have to set up a profile/behavior for my Norwegian visitors. The image below shows these setting.

3. Create your BTBuckets Targets

The next thing you must do is to create your target. Visitors belonging to the "All Visitors" bucket will be targeted with the English Feedback Form. Visitors from Norway will be targeted with the Norwegian Feedback Form.

The example below show my Norwegian target. The code in the screenshot below is the "Feedback Form Settings Code". You find an example code below the image.

If I want to disable the Norwegian Feedback Form, I just have to log into BTBuckets and uncheck the "Norwegian Visitors" bucket and click save. Voila, my Norwegian visitors will no longer see a Feedback Form, and I didn't have to change any code on my own website.

Below is some example code that you can copy and past into your BTBuckets target form. It will create a Feedback Form with Horizontal Radio Button Layout and Comment Field.

Remove formatting from Code

<script type="text/javascript">
// Text you can change
var feedbackHeading = "How useful was this page?"; // Feedback Heading
var feedbackCommentHeading = "Comment"; // Heading above your Comment field
var feedbackButtonText = "Submit"; // Text on Submit Button
var feedbackButtonTitle = "Give Feedback"; // Title Text on Submit Button
var feedbackHideText = "Hide Feedback"; // Title Text Hide Feedback
var feedbackShowText = "Show Feedback"; // Title Text Show Feedback
var feedbackCommentThx = "Thank You for Your Feedback."; // Thank you text after feedback is sent
var feedbackCommentGiven = "You have given Feedback."; // Text shown to those who have been given feedback on ex. reload of page
var feedbackRadioButtonError = "You must select a radio button"; // Error Text when Radio Button is not selected
var feedbackTypeText = "Bad,Good"; // Text supporting the radio buttons. Separate with comma (,). If you are using vertical aligned radio buttons, number of options must match "feedbackOptionsNumber".
// Settings you can change
var feedbackOptionsNumber = "5"; // Number of feedback options (radio buttons).
var feedbackRadioType = "horizontal"; // Should Radio Buttons be aligned "horizontal" or "vertical". (horizontal/vertical)
var feedbackPlacement = "right"; // Should the Feedback Form be placed "left" or "right" at the bottom of the screen
var feedbackCommentValidation = "optional"; // Is Comment "optional" or "mandatory"? (optional/mandatory)
var feedbackCommentVisibility = "show"; // Should Comment Field be shown or hided? (show/hide). If "hide", feedbackCommentValidation must be set to "optional" and feedbackRadioType to "vertical".
var feedbackTextColor = "black"; // Text Color
var feedbackBgColor = "white"; // Background Color
var feedbackBorderColor = "grey"; // Border Color
var feedbackButtonColor = "#333"; // Background Color on Submit Button
var feedbackButtonTextColor = "#fff"; // Text Color on Submit Button
var feedbackCharRestrictions = "200"; // Charachter Restrictions
var feedbackLevel = "page"; // Choose between "page" level or "site" level Feedback (page/site).
var feedbackTimeRestriction = "30"; // Number of days cookies should live
var feedbackWidth = "170"; // Width of Feedback Form in pixels
var feedbackGoalURL = "no"; // If set to "yes" a Fake URL will be generated when the form is successfully submitted so it can be tracked as a goal. URL: /feedback-form/feedback-submitted
var feedbackErrorTracking = "no"; // If set to "yes" Errors like "Not selected Radio Button" etc. and Open and Closing of the Feedback Form will be tracked.
</script>

4. Install the BTBuckets Script and track your data in Google Analytics as Custom Variables

You don't have to track your BTBuckets data in Google Analytics, but I really recommend you to do so. We are going to do so using Google Analytics Custom Variables (this method is also described on the BTBuckets community page).

Remove formatting from Code

<script type="text/javascript">
//<![CDATA[[
$BTB={s:XXXXX};
$BTB.jsh="btbuckets.com/bt.js";
//]]>
</script>
<script type="text/javascript">
$BTB.jshost = (document.location.protocol == "https:")?"https://ssl.":"http://static.";
document.write(unescape("%3Cscript src='" + $BTB.jshost + $BTB.jsh + "' type='text/javascript'%3E%3C/script%3E"));
</script>
<script type="text/javascript">
var btbucketsGA = $BTB.getAllUserBuckets().join("-");
</script>
<script type="text/javascript">
var _gaq = _gaq || [];
_gaq.push(
['_setAccount', 'UA-XXXXX-X'],
['_setCustomVar',1,'BTBuckets',btbucketsGA,1], // BTBuckets data stored as Custom Variables in Slot 1 with Scope set to Visitor Level
['_trackPageview']
);
</script>
<script type="text/javascript">
(function() {
var ga = document.createElement('script');
ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
ga.setAttribute('async', 'true');
document.documentElement.firstChild.appendChild(ga);
})();
</script>

This is it. You have now a even more flexible solution for the Feedback Form, with even greater data collected in Google Analytics.

Some closing comments about Behavioral Targeting

Used wisely Behavioral Targeting can really enhance the user experience, but used wrongly your visitors may leave your website screaming. Nobody likes to feel that someone is watching them, and used wrongly Behavioral Targeting can give that impression.

Also notice that if you use BTBuckets you will have to inform your users about it, and how they can opt out. I have this information on my privacy page.

Some final words about this site and how to comment

If you haven't had any comments approved on this blog before, I will have to approve your comment manually.

How to Measure True Time with Google Analytics

Eivind Savio — 2010-08-30

If you are familiar with how web analytics tools measure time spent on page or site, you know that time spent on the last page is (usually) counted as 0 seconds. With other words, time on last page is not measured.

The reason this isn't measured is simply that the web analytics tool need to have a start point and a end point. On the last page of the visitors visit, there isn't a end point normally to calculate time against.

To be able to get time spent on last page you have to be able measure 2 things:

Clicks on Back Button
Visitors leaving by Closing their Browser or Browser Tab

In this article I show you how you can measure "True" Time spent on your website with Google Analytics by measuring Back Button and Visitors who leave by Closing their Browser.

The time difference between Standard Time Measurment and True Time Measurement can be huge.

How to measure "True" Time by Tracking Back Button and users who leave by Closing their Browser

There are 3 steps you have to do to be able to do this tracking.

Step 1.
The first thing you should do is to create a new Google Analytics profile with a new profile ID. You shouldn't implement this measurement in your regular Google Analytics profile. This method will "kill" your Bounce Rate in addition to that your time spent reports of course will be inflated. With other words, we are going to track data to 2 different profiles in Google Analytics.

The principle for tracking to 2 different profiles is shown below using Google Analytics Asynchronous code.

Remove formatting from Code

<script type="text/javascript">
var _gaq = _gaq || [];
_gaq.push(
['_setAccount', 'UA-XXXXX-1'], //Main Google Analytics profile
['_trackPageview'],
['pageTrackerTime._setAccount', 'UA-XXXXX-2'], //Profile for True Time Tracking
['pageTrackerTime._trackPageview']
);
</script>
<script type="text/javascript">
(function() {
var ga = document.createElement('script');
ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
ga.setAttribute('async', 'true');
document.documentElement.firstChild.appendChild(ga);
})();
</script>

Step 2.
The next thing you have to do is to download this ZIP-file. The ZIP-file contains 2 files and is necessary in the measurement of the Back Button.

Blank.html - This is an empty HTML-file that you must upload to the root of your domain (yourdomain.com/blank.html).
Backfix.js - This is the javascript that does all the magic that is necessary to be able to track the Back Button

The solution I'm using to track the Back Button is a slightly modified solution borrowed from the Browser Back Button Detection article by Brook Bryan. The original solution was creating 404-errors that Google Analytics counted, so the only thing I have done is to fix that.

Upload backfix.js to for example a folder called script.

Step 3.
This is the last step. Copy and paste the code below between <head> and </head> on your website.

Remove formatting from Code

<script type="text/javascript">
leaveSiteStart = new Date;
function leftSite() {
var leaveSiteEnd = (new Date - leaveSiteStart )/1000;
var roundleaveSiteEnd = Math.round(leaveSiteEnd,1);
_gaq.push(['pageTrackerTime._trackEvent','Track Time','Closed Browser', document.location.href, roundleaveSiteEnd]);
}
window.onbeforeunload=leftSite;
</script>
<script type="text/javascript" src="/script/backfix.js"></script>
<script script type="text/javascript">
bajb_backdetect.OnBack = function() {
var leaveSiteEnd2 = (new Date - leaveSiteStart )/1000;
var roundleaveSiteEnd2 = Math.round(leaveSiteEnd2,1);
_gaq.push(['pageTrackerTime._trackEvent','Track Time','Back Button', document.location.href, roundleaveSiteEnd2]);
}
</script>

If you have done everything correctly, you should if you drill down to you Event Tracking Report find a report called Track Time. There you will find a report similar to the screenshot below.

Event Labels in that report will be the URL the visitor was on before he either clicked the back button or closed the browser. Event Value reported is the number of seconds from the visitor arrived to the page and until the browser was closed or the back button was clicked.

If you drill down even more in your numbers you will perhaps get the same feeling as me:

Time spent on page or site isn't the same as time spent interaction with your content.

To be honest, I think the image below sums up the behavior to many of my visitors. They do hopefully not fell asleep, but they leave the page open and do some other things before they close the page.

Some readers may spend long time on your website due to other reasons than reading. Image by notemily.

Things to be aware of - and Why did I do this?

First, I'm not a programmer so if you find any errors or don't like mye code, please tell it nicely.

This measurement method also shows why this method isn't a standard method in web analytics tools - this method isn't accurate. Only around 80 % of my visitors are measured using this method, which is why I have written "True" Time and not True Time in this blog post.

Since I'm normally a person that will speak out for actionable analytics, which this isn't, so why did I do this?

An important part of web analytics is to be curious. If you aren't curious you will find it more difficult to be friend with your data. With other word, the reason for me to do this was simply because I was curious and had some geek time to use.

Some final words

How to Identify if Visitors are Reading Your Articles with Google Analytics

Eivind Savio — 2010-08-09

One question I often get from webmasters running content websites is:

Do people read my articles, and is it possible to identify this with Google Analytics?

Not all visitors are reading your articles, but more visitors are maybe reading than you think. Image by Federico Scotto d'Antuono.

If you have high bounce rate and lots of single page visits, is it really so that all these visitors didn't find any value in your content? This can be true, but sometimes these visitors did read and found value in our content, but we can't identify it from a standard setup in Google Analytics.

"What if we have a low bounce rate, wouldn't time spent indicate how visitors are interacting with our content?" you may ask then. My answer would be "it depends, but a standard setup in Google Analytics would not measure interaction on the last page the visitor visited".

In this article I will dive into how you can use Google Analytics to identify if visitors are reading your articles.

Is "real" Bounce Rate a possible metric to use?

I have looked at a couple of methods to try to get an better understanding of how visitors interact with my content. For example, a solution called The Real Bounce Rate in Google Analytics is trying to get a better picture of the bounce rate.

Unless you write very short articles I would normally not recommend it (I will explain why later), since this solution doesn't tell a story about the behaviour to the visitor. In my optinion, The Real Bounce Rate solution doesn't answer if visitors are reading your content because it doesn't measure interaction.

When I was researching for a solution to measure if visitors are reading my content, I used Google and searched for terms that I assumed could be an answer. On the search result page (SERP) I got several listings that perhaps could be the solution I was looking for, but since I weren't sure I opened up several of the pages I found in different tabs in my browser.

I found what I was looking for, but my visits to all the pages were a bounce. If all those other pages where using The Real Bounce Rate solution they would see my visit as a non-bounce, which it wasn't. However, I found only value in one of the results, and I read that article from top to bottom.

My visit to that article was a bounce using standard measuring, but the Google Analytics tracking solution on that page would identify me as someone that did read the content.

The solution - if I have to scroll to read your content

The answer I found was an article called Tracking scroll depth to reveal content engagement in Google Analytics, so I implemented this to test the solution. In short what this solution is doing is to track if the visitor scrolls down 90 % of the page, and if the visitor does that a javascript kicks off an Event in Google Analytics.

After a couple of days I found that one piece of the puzzle was missing with this method, namely time used from the page loaded to the visitor had scrolled down 90 % of the page. Although it's possible to get a feeling of this from other reports in Google Analytics (but these numbers will not be the same since they also includes "normal" time spent), I wanted to spot this information easy.

Although the number of seconds isn't a perfect metric, it can give you a clue about the interaction with your content. Did the visitor just scroll quickly down your page, or did he use some time?

Using long time can be an indication of that the visitor is in fact reading your content.
Using short time can be an indication of that the visitor is just scanning your content (I especially find short scroll time on category pages).

Scroll down report using Google Analytics Events. Event Value is scroll time measured in seconds.

Below you find the code. It's identical to the code found in the original article Tracking scroll depth to reveal content engagement in Google Analytics except from 3 differences:

The script is measuring time spent from the page loads to the visitor has scrolled 90 % of the page (this difference in coding is highlighted with red text).
"My" code is using the newest Google Analytics Asynchronous code.
I am tracking scroll to a different Google Analytics profile than my main profile (called secondTracker in the code below, and highlighted with green text). The reason for this is that this script will affect both bounce rate and time spent, and tracking to 2 different profiles gives me the opportunity to compare data.
- Add a new Google Analytics profile (Add a profile for a new domain).
- See this page for information about how to send information to multiple trackers.

Remove formatting from Code

<script src="http://ajax.microsoft.com/ajax/jquery/jquery-1.3.2.min.js" type="text/javascript"></script>
/* <![CDATA[ */
var scrollTimeStart = new Date;
var IsDuplicateScrollEvent = 0;
$(document).ready(function () {
SetupGoogleAnalyticsTrackEvents();
});
function SetupGoogleAnalyticsTrackEvents()
{
TrackEventsForMinimumPageScroll();
}
function TrackEventsForMinimumPageScroll()
{
$(window).scroll(function(){
var scrollPercent = GetScrollPercent();
if(scrollPercent > 90)
{
if(IsDuplicateScrollEvent == 0)
{
IsDuplicateScrollEvent = 1;
// alert("Page Scrolled to 90% in " + document.location.href);
var scrollTime = (new Date - scrollTimeStart )/1000;
var roundTimeScroll = Math.round(scrollTime,1);
TrackEvent("Content Engagement", "Scrolled To 90%", document.location.href, roundTimeScroll);
}
}
});
}
function GetScrollPercent()
{
var bottom = $(window).height() + $(window).scrollTop();
var height = $(document).height();
return Math.round(100*bottom/height);
}
function TrackEvent(Category, Action, Label, Value)
{
_gaq.push(['secondTracker._trackEvent', Category, Action, Label, Value]);
}
/* ]]> */

If you don't want to copy + paste the code, just download the script (ZIP-file).

Possible drawbacks with this method

If you write short articles, and a visitor don't have to scroll to read your complete article, this method will not work. In these cases using the "real" bounce rate solution can be something to look into. I would however in these situations use the "real bounce rate solution from Immeria, it's a little bit more sophisticated.
Screen resolution can affect if the reader must scroll or not to read your complete article.
This method is only based on quantity data, not quality, so you can really not now if the visitor did read your article even if he used some time to scroll down. However, based on my experience, it can be an indicator.
If you are running the Prototype Javascript Framework on your site the prototype script may not function properly.

I did also fear that screen height would affect time spent when scrolling down the page. A visitor with a high screen resolution will not have to scroll as much as a visitor with low screen resolution. Because of this I have checked if I could find correlation between screen height and time spent scrolling.

Hypothesis: Screen height should afffect the time.
- I have checked this on different levels and for different articles, and I did not find any correlation.

Even if I didn't find any correlation, I recommend that you check this yourself also since my articles isn't your articles.

What do you think about this method for trying to identify if visitors are reading your content? Have you found other methods useful in measuring this?

Some final words

Click "Legg til kommentar" below. Than fill out your name (Navn), e-mail (Epost) and your website (Ditt nettsted) if you have a website. HTML formatting doesn't work in my comment field, use BBCode formatting instead. You can drag the comment field around if you want to. If you want to preview what you have written, click on "Forhåndsvis". When you are done and want to save your comment, click on "Lagre". If you haven't had any comments approved on this blog before, I will have to approve your comment manually.

Ingen sp�k � brekke ryggen selv p� 1. april

Eivind Savio — 2010-07-05

1. april i år bidro jeg til å øke 2 statistikker i Norge som jeg gjerne skulle vært foruten, samt fikk mine venner og bekjente til å lure på om jeg hadde veldig dårlig humoristisk sans. Dessverre var det ingen spøk.

1. april brakk jeg ryggen på telemarkski, noe som igjen har resultert i at jeg ble sykemeldt for aller første gang. I tillegg ble jeg ble en av de 60 - 70 som årlig pådrar seg en ryggmargsskade.

Jeg skriver vanligvis ikke om private ting på denne bloggen, og har ikke noe spesielt formål med å skrive om ulykken min bortsett fra å bruke anledningen til å takke alle de som har gjort en innsats for meg - både venner og alle de som har gjort en innsats for meg i helsevesenet.

Ulykken

Skaden skjedde som nevnt på telemarksski 1. april, og skyldtes en noe hard landing etter et noe mislykket hopp (for å si det pent) i Vangslia i Oppdal. Jeg landet på beina, men siden det var heller flatt der jeg landet var det ikke saft nok i beina til at jeg greide å stå. Baken var derfor det neste som fikk kontakt med den harde påskesnøen før ryggen fulgte samme veien, og påskeferien var dermed over for min del.

I det baken prøvde å sette sitt avtrykk i snøen kjente jeg en smertestråle som skjøt ut fra korsryggen og ned i høyre lår. Umiddelbart tenkte jeg at nå har jeg brukket ryggen, og da jeg var ferdig med å falle hadde jeg av en eller annen grunn lagt meg selv i et slags stabilt sideleie.

Min gode venn Gjermund som jeg var på ski sammen med kom straks til og lurte på hvordan situasjonen var, og om Røde Kors skulle kontaktes. Jeg konstaterte at jeg kunne bevege bein og tær og var i hvert fall ikke lam, men at smertene i ryggen og høyre lår tydet på at jeg ikke burde prøve å reise meg opp eller bevege meg noe særlig. Oppdal Røde Kors Hjelpekorps som hadde vakt i bakken ble tilkalt.

Personellet fra hjelpekorpset tok det for gitt at jeg hadde en ryggskade og behandlet meg deretter, det vil si med største varsomhet, men det gjorde likevel jæv... vondt. De la meg på en scoopbåre og løftet meg over i pulken, før de fraktet meg de få 100 metrene med snøscooter ned til en ventende ambulanse.

St. Olavs Hospital

Etter en kort tur innom legesenteret på Oppdal gikk turen med samme ambulanse til St. Olavs Hospital i Trondheim. En forsiktig sjåfør, samt at de ikke var kjipe på morfinen, gjorde den cirka 1,5 time lange kjøreturen utholdelig. En stor takk til dere.

På St. Olavs Hospital stod først røntgen på programmet. Dette innbar at jeg måtte bli lagt på en annen båre. Sykepleieren samlet sammen "det som var å oppdrive av folk", og om jeg ikke husker feil var det til sammen 8 personer som gjennomførte et klesløft på meg. Til tross for deres varsomhet - aldri har jeg hylt så høyt av smerte tror jeg. Samtidig synes jeg at dette viser hvor viktig det er med nok hender på sykehusene. Sykepleieren som organiserte løftet måtte tromme sammen portører og hjelp fra andre avdelinger for å få flyttet meg.

Røntgenbildene konstaterte brudd i ryggen, og jeg ble trillet inn på et enerom i påvente av operasjon. St. Olav er 1 av 3 sykehus i Norge som kan gjennomføre denne typen operasjoner. De 2 andre er Haukeland i Bergen og Ullevål i Oslo.

Etter litt venting på rom i en sånn passelig morfindøs, kom det jeg tror var avdelingssykepleieren inn og informerte om at de ikke hadde noen kirurger som kunne gjennomføre operasjonen denne kvelden. Fordi det var viktig å få operert så snart som mulig hadde de konferert med Ullevål universitetssykehus og de kunne operere. Jeg ville derfor om 2 timer, kl 22, bli hentet av en ambulanse som ville kjøre meg til Værnes til et ventende ambulansefly for å bli flydd til Gardemoen.

Luftambulansetjenesten

Ambulanseturen til Værnes er ikke så mye å skrive om egentlig, bortsett fra at ambulansepersonellet var noen hyggelige folk som hadde stor forståelse for mitt ønske om morfin.

På Værnes ble jeg lastet inn i et fly fra Luftambulansetjenesten. Jeg hadde etter forholdene en fin flytur (med prioritet i luftrommet) til Gardemoen. Ombord på flyet var det verken kaffe eller ferske aviser, men morfinen som ble servert gjorde flyturen til en etter omstendighetene hyggelig opplevelse sammen med den erfarne sykepleieren som var ombord.

Etter en myk landing ble jeg tatt i mot av 2 ambulansesjåfører som var erfarne i å ta i mot pasienter som kommer med ambulansefly. Det var litt knotete og trangt å få båren inn og ut av flyet, og jeg er derfor glad for at de hadde gjort dette mange ganger før.

Ullevål universitetssykehus - om operasjonen og skaden

På nevrokirurgisk avdeling på Ullevål universitetssykehus stod det et team og ventet på meg. Jeg tror klokken da var halv to på natta.

De satte straks i gang med undersøkelser. Det ble tatt nye røntgenbilder for å kunne se om det hadde skjedd noen endringer under transporten, samt at det ble gjennomført CT-skanning i tillegg til noen fysiske undersøkelser. For første gang skjønte jeg for alvor at ikke bare var ryggen brukket, men at ryggmargen min sannsynligvis også var skadet.

Jeg ble deretter operert under full narkose så den delen vet jeg ikke så mye om, annet enn at det visstnok tok 2 timer og 40 minutter, og at jeg ble trillet ut av operasjonssalen kl 0530.

Skaden

Jeg har ikke tenkt å gå helt i detaljer om skaden min, først og fremst fordi at jeg fremdeles å betrakte som nyskadd når det gjelder ryggmargen, og at det derfor er for tidlig å si noe om konsekvensene av skaden.

I ryggen har jeg brukket ryggvirvelen lumbal 1, eller L1 som også brukes som betegnelse. Bruddet er et kompresjonsbrudd, og ut fra hva jeg kunne se av bildene var ryggvirvelen delt i 3 deler. Under operasjonen ble det skrudd inn skruer i virvelen over og under L1. Skruene holder noen støttestag som går forbi bruddet og stiver av ryggen. Skruer og stag vil bli der for alltid om ikke de ikke gir meg smerter. Ved smerter kan de opereres ut etter 1 års tid. Mine skruer blir nok der for alltid.

I tillegg ble (deler av?) ryggtaggen til L1 fjernet for å gi nok plass til ryggmargen da den hadde ligget kraftig i klem. Kirurgen sa etter operasjonen at de ikke hadde kunnet se noen ytre skader på ryggmargen, men den lå som nevnt i klem. Jeg hadde også et lite brudd i bekkenet.

Ryggmargskaden er det som nevnt for tidlig å si noe om konsekvensene av, men jeg har en såkalt inkomplett skade, det vil si at ikke alle impulser er kuttet av. Skaden min er også så ny at det er for tidlig å si noe om ryggmargen er permanent skadet, eller om det er et spinalsjokk. Ved spinalsjokk kan funksjonalitet og følelse komme tilbake etter en (lang) stund, er ryggmargen skadet vil ikke skaden kunne helbredes.

Basert på progresjonen jeg har hatt, og tilbakemelding på undersøkelser som har vært gjort har jeg godt håp om at jeg en dag vil bli bra igjen - men det vil kunne ta tid. Vær tålmodig er noe jeg har hørt ofte.

Som nevnt innledningsvis kunne jeg etter ulykken bevege tær og føtter - jeg har tross alt hatt hell i uhellet og har ikke blitt lam. Siden operasjonen har jeg nesten kontinuerlig kunnet merke fremgang, og møter du meg i dag vil du antageligvis ikke se at jeg er skadet.

Hvilke problemer man kan oppleve etter en ryggmargskade avhenger av hvor i ryggen ryggmargen har blitt skadet, og ingen inkomplett skade vil være lik en annen selv om det finnes fellesnevnere. Dette kan være veldig frustrerende, spesielt rett etter skaden da ingen kan si hvilke skader man vil ende opp med, hvor lang tid det vil gå før man kan si noe sikkert osv. På den annen side gir dette håp for at man kan bli bra igjen.

Redusert muskelstyrke i høyre fot, litt dårligere balanse i høyre fot, redusert eller manglende hudfølelse, en heller ræva setemuskulatur og problemer med naturlige funksjoner er ting jeg har opplevd etter ulykken og som skyldes en ryggmarg som ikke er helt på nett.

Rehabiliteringen

Påske-coctail på post-operativ avdeling Ullevål

Rehabiliteringen startet umiddelbart etter operasjonen. Den første uken etter operasjonen på Ullevål lå jeg på post-operativ avdeling. Man ligger vanligvis ikke så lenge på denne avdelingen før man blir overført til vanlig sengepost, men fordi jeg fikk blodtrykksmedisin for å øke blodtrykket mitt måtte jeg ligge der. De 7 første dagene etter en ryggmargskade kan mer oksygen til skadestedet forbedre resultatet, og blodtrykksmedisinen sørger for dette.

Jeg kan ikke få rost de som tok hånd om meg på Ullevål godt nok, og da spesielt sykepleierne på post-operativ avdeling. Alltid et smil og omtanke, samtidig som de pushet meg i gang med å komme meg opp av senga. Dagen etter operasjonen stod jeg ved siden av senga, og 2 dager etter operasjonen gikk jeg mine første ustøe skritt. Begge deler med støtte vel og merke, men likevel.

Sykepleierne på post-operativ avdeling tenkte også fremover på min videre rehabilitering og sørget for det ble søkt om plass for meg på Sunnaas Sykehus.

Sunnaas sykehus

Ambulansetransport til Sunnaas sykehus
fra Ullevål

9. april ble jeg overflyttet til Sunnaas sykehus for rehabilitering, og jeg ble utskrevet 17. juni.

Sunnaas sykehus er ikke som et helt vanlig sykehus, da oppgaven de har er å rehabilitere deg og klargjøre deg for livet som ryggmargsskadd etter din tid på Sunnaas. Ved ankomst får man et team rundt seg som skal kunne ivareta de fleste ting som dukker opp. Alt fra praktiske ting som du plutselig ikke behersker selv, til trening, undersøkelser, hjelpemidler, papirer og ikke minst støtte og menneskelig omsorg.

De jeg hadde mest med å gjøre var sykepleierne og fysioterapeutene. Sykepleierne fordi de naturlig nok var de som måtte hjelpe meg alt jeg ikke fikk til, og fysioterapautene gjennom treningen. Jeg har aldri trent så mye som på Sunnaas, opptil 3-4 økter om dagen på enkelte dager - men det hjalp virkelig!

Jeg kunne skrevet mye om Sunnaas og min tid der, og hvordan det føltes å oppnå de små seirene som kom etterhvert, men jeg skal nøye meg med ett eksempel. Har du tenkt over hvilken glede det kunne gi å kunne bøye seg ned mot vasken igjen og kunne kaste kaldt vann i ansiktet om morgenen?

Ute og triller prekestol på Sunnaas sykehus

Til alle de som var mine medpasienter - lykke til med deres rehabilitering. Det var hyggelig å bli kjent med dere.

Til alle dere som jobber på Sunnaas - takk for den fantastiske jobben dere gjorde for meg, og all den støtten dere viste. Når ting først gikk galt er jeg glad for at det var til dere jeg kom, og på vegne av fremtidige pasienter er jeg også glad for at Sunnaas fremdeles kommmer til å være et eget helseforetak.

Erfaringer og hva nå?

Når dette skrives er jeg fremdeles sykemeldt, men satser på å starte så smått og jobbe igjen i begynnelsen av august. Ryggmargen er fremdeles ikke helt samarbeidsvillig så jeg starter antageligvis med aktiv sykemelding.

Som jeg skrev innledningsvis er dette min første sykemelding noensinne. I og med at jeg aldri har vært sykemeldt har jeg egentlig aldri tenkt over hva slags konsekvenser det kan få å bli langtidssykemeldt - og da spesielt de økonomiske. I tillegg til alle utfordringene med det å bli skadd (som i mitt tilfelle) vil man også kunne få en skikkelig økonomisk baksmell. Selv om livet mitt ble satt på pause ble det ingen pause på faste utgifter og regninger - de kom trillende inn som før.

Heldigvis for meg dekket min arbeidsgiver mesteparten av differansen mellom hva jeg får i sykepenger og lønnen min så dette har ikke blitt et problem for meg, men hva med deg?

Hva skjer med økonomien din om du blir langtidssykemeldt?

Man gjør seg også mange erfaringer med venner og kjente når man plutselig blir borte fra det vanlige livet over en lengre periode. For min del var dette i hovedsak en positiv opplevelse. Takk til alle som har tatt i et tak for meg, besøkt meg, invitert meg på middag og sendt meg varme tanker.

Video Analytics using Google Analytics API and Excel

Eivind Savio — 2010-03-17

Google Analytics Event Tracking is a flexible solution for measuring user interaction with your website - including videos. I have demonstrated this before with my YouTube Chromeless Player with Google Analytics Event Tracking (you can download the player from that article).

Since Event Tracking isn't made especially for video analytics, it can be a little bit difficult to understand what the data means as the image to the right illustrates (click for larger image).

To make the video analytics data more understandable I have made an Excel 2007 spreadsheet that is connected directly to Google Analytics using the Google Analytics API.

Different Video reports

The Excel spreadsheet contains 3 different reports.

Video Dashboard

The dashboard contains an overview of your video data from the selected month and includes video KPI's for 3 different Stakeholder levels (levels borrowed from Eric T. Petersons Big Book of KPIs).

Senior strategists
Mid-tier strategists
Tactical resources

The dashboard shows information about this month compared to last month, and this month compared to the same month last year plus a graph showing trends for the last 12 months.

The dashboard and the calculations are a rewritten version of Web Analytics Dashboard fun by Stéphane Hamel.

Video KPIs are inspired from Online Video Analytics - KPIs written by Dennis Mortensen.

Videos

This sheet contains information about which videos that have been watched the month you have chosen in the month field on the dashboard. The report contains the following information:

Video views (Videos that have been started)
Completed videos
Percent completed videos
Visits
Video length (If none of the visitors have completed the video the selected month this report will be blank)
Average time viewed
Percent of video viewed on average
Total time viewed the selected month
Conversion Rate (E-commerce Conversion Rate)
Number of Conversions

Single Video Analytics

This report shows a day to day report for the selected month, and contains the same information as the "Videos" report, but for the single video selected. The graph shows information about views, completed views, visits and e-commerce conversions.

Getting started

The Excel spreadsheet is adapted to my custom YouTube player, so if you are using this player you should be able to use the Excel spreadsheet as it is. If not you will have to modify the Excel spreadsheet to your needs.

Requirements

You will need Excel 2007. It will unfortunately not work in older versions of Excel or on Mac. Sorry.

Quick start

Download the zipped Excel 2007 spreadsheet and unzip it.
Open the Excel file and activate macros.
In the field "Month" on the Dashboard, choose what should be the last month in the report.
Go to the Settings sheet, and change "Sites" from YourWebsiteName to your site name.
On the same line as your site name, insert your Google Analytics ID number. This is not your UA-XXXX-X ID, it's the ID in the Google Analytics URL (ex. https://www.google.com/analytics/reporting/dashboard?id=123456)
Insert your Google Analytics username (email) and password in the Settings sheet.
Get yourself a cup of coffee (too many calculations going on in Excel) and wait for the result.

Since all the calculations can make Excel slow it can be wise to change Calculation Options from Automatic to Manual. You find this setting under the Formulas menu. Then if you need to update your sheet, hit Shift + F9 if you only want to calculate the sheet you are on, or F9 if you want to calculate your entire workbook.

More information of the sheets in the Workbook

Dashboard

The KPIs on the Dashboard are flexible, which means that the same calculations can be reused for different KPIs. Each stakeholder can have up to 4 KPIs.

Settings

You will also find a table on this sheet with the columns EventCategory and EventAction. If you are using different category and action namings than me you will have to change this.

Naming of Stakeholders are also done on this sheet. The numbers in the Num. column is used to attach the different KPIs to the different stakeholders.

Segment1

All the monthly data are reported here. You shouldn't normally touch these data unless you want other data reported or calculated.

Lookup

All the calculation of average values and standard deviation is done on this sheet. If you add or changes column titles on the Segment1 sheet you must also add the same columns to this sheet (use the same column title). Except from that you shouldn't normally have to change anything on this sheet.

KPIs

On this sheet you can edit the KPI names and description. Column KPI1 will match Stakeholder 1 on the Settings sheet.

KPI names must match headings used in the Segment1 sheet and the Lookup sheet.

GA Metrics and Dimensions

On this sheet you find all Google Analytics Metrics and Dimensions available with explanations.

How Excel is fetching data from Google Analytics

Excel is fetching the data from the Google Analytics API using VBA coding. These codes are written by Mikael Thuneberg, and he has explained on his blog how to use the codes. You will also find other Excel examples there.

Some final words about this Excel Video Analytics spreadsheet

Everything about this Excel Video Analytics spreadsheet isn't explained in this article - it's just too much to explain. Therefor

Download Excel 2007 Video Analytics spreadsheet (120 Kb)

and then make yourself familiar with the spreadsheet and try to understand how things work.

I'm using Excel Tables, Arrays and Name(ing of ceels) heavily in the spreadsheet, so these are some things that should try to understand how works.

And since an article about Video Analytics isn't complete without a video, below is a video where the developers from Google explains the Google Analytics API.

Happy Playing and Analyzing.

If you make any improvements to the Excel video sheet I would love to hear about it.

YouTube Chromeless Player with Google Analytics Event Tracking upgrade

Eivind Savio — 2010-03-14

This is just a short notice of an upgrade of my YouTube Chromeless Player with Google Analytics Event Tracking.

The player has been updated with a bug fix and a improvement to the Google Analytics Event Tracking.

Tracking time when somebody leaves the page or closes the browser window
Video - Stop Left Page is added to the player/Google Analytics Event Tracking. Earlier if a visitor just left the page or closed the browser windon, the time would be recorded as 0 seconds no matter how long the video had been playing. By using ununload and unbeforeunload the time is now tracked. The time will only be tracked if the video is either playing, buffering or is paused. This is done by checking getPlayerState.

If you are using a previous version and want to upgrade, look for the function leave() in the new youtube.js, and update the <body> tag in default.asp or index.php to this:
<body onload="embedPlayer();" onbeforeunload="leave();" onunload="leave();">

Google Analytics Asynchronous Tracking

The player is using the new Google Analytics Asynchronous Tracking. All the Google Analytics codes are placed inside youtube.js. You will have to change the Google Analytics UA-code (UA-XXXXX-X) to your account number.

Below you find a video about Tracking events using Advanced Segments in Google Analytics. I thought it was appropriate to integrate this video since it's a little bit about video segmentation in Google Analytics.

If you make any improvements to the player I would love to hear about it.

If you are familiar with my previous version of the player, Download my YouTube Chromeless Player Example (zip). If not I recommend you read my article YouTube Chromeless Player with Google Analytics Event Tracking first.

Veien til Effektiv Webanalyse

Eivind Savio — 2010-02-08

Poenget med webanalyse er optimalisering og handling, ikke innsamling av data, verktøy og flotte rapporter. Gjennom optimalisering og handling vil webanalysen bidra til økonomisk verdiskaping og forbedring av brukskvaliteten.

En rapport fra Econsultancy fra en tid tilbake tilsier at investeringen i webanalyse vil øke. I rapporten kommer det også frem at investeringene som gjøres er hovedsakelig i humankapital og ikke i verktøy - fordi det er menneskene som skaper verdiene, og ikke verktøyene.

For å kunne optimalisere og handle må man gjøre et forarbeid. I dette innlegget tar jeg for meg veien frem til optimalisering og handling gjennom en modell kalt

Effektiv webanalyse

Modell for Effektiv Webanalyse.

Strategisk fundament

Webanalyse uten en strategisk forankring i bedriften eller organisasjonen gjør arbeidet mye tyngre. Uten forankring hos ledelsen er det vanskelig å få gjennomført forbedringer (handling). I stedet for å bruke tiden til verdiskapning for egen bedrift/organisasjon eller for de besøkende, vil mye av tiden gå med til å misjonere webanalyse innad.

Webanalyse kan med andre ord også innebære organisasjonsendringer, politiske prosesser, endringer av tankesett og annerledes allokering av budsjetter.

Definer forretningsmål og ønsket utbytte

Det første man bør gjøre er å definere hvilke forretningsmål webanalysen skal gi svar på, og hva som er ønsket utbytte. Utarbeidelse av KPIer (Key Performance Indicators) vil bidra til å gi svar på dette, samt kunne avdekke hvilke målinger som bør gjøres.

Datainnsamlingsbehov og metode

Gjennom utarbeidelsen av KPIen bør det komme frem hvilke databehov man har, og gjennom dette også hvilke metoder som bør benyttes. Med andre ord, utarbeidelse av KPIer vil indirekte kunne si noe om valg av webanalyseverktøy.

Er det et krav at dataene må lagres in-house?
Er det behov for å lagre personinformasjon i webanalyseverktøyet?
CRM-integrasjon eller integrasjon mot andre systemer?
Innsamling av kvalitative data?
Økonomiske forhold?

Selv om jeg skriver en del om Google Analytics er jeg verktøynøytral. For enkelte vil ikke gratisverktøy slik som Google Analytics eller Yahoo! Analytics være tilstrekkelig, men valg av verktøy bør avgjøres ut i fra hvilke behov man har, og ikke ut i fra en salgspitch fra en verktøyleverandør.

Operasjonell infrastruktur

Oppsett av webanalyse i henhold til behov

Denne delen har til hensikt å gi svar på hvordan informasjonen skal samles inn. Vi har kommet frem til hvilke(t) verktøy som er mest formålstjenlig for vårt behov.

Oppsett og implementering kan i blant være en forferdelig tung prosess, men et så riktig oppsett som mulig er nødvendig for at man skal kunne ta riktige beslutninger basert på dataene som kommer ut i andre enden. Med vilje har jeg skrevet "et så riktig oppsett som mulig", fordi at det vil også være et spørsmål om kost/nytte.

Mange installeringer jeg har sett er ikke satt opp i henhold til behov. Årsaken til dette er ofte at man ikke har definert forretningsmål eller ønsket utbytte. Dette punktet handler også om datakvalitet. Hvilke data er det viktig å samle inn, og hvilke data bør filtreres bort?

Dataene vil aldri bli perfekte, men de bør være så gode at man kan gjøre kvalifiserte analyser.

Noen enkle eksempler på dette kan være:

Er kontoen riktig organisert?
Spores konverteringsmål?
Er egne ansatte ekskludert fra dataene?
Spores kjøp og salg i nettbutikken? (Jeg overraskes stadig over nettbutikker som ikke tracker kjøp og salg).
Spores søk fra "norske søkemotorer" slik som ABC Startsiden?
Spores feilmeldinger, og er webanalysen satt opp til å spore brukerproblemer?
Internsøk og søk med null treff
Kampanjesporing
Segmentering, for eksempel kunde/ikke-kunde.
Sporing på tvers av domener og subdomener
Sporing av RIA (Rich Internet Applications) slik som Flash og AJAX

Men, det er heller ikke noe poeng i at man skal samle alle typer data bare fordi man kan. Som Avinash Kaushik sier:

Det er ikke noe vits i å være rik på data, men fattig på informasjon.

Et forhold som mange ikke tenker over, er at om man bytter til en annen publiseringsløsning, må man også påberegne at webanalysen må settes opp på nytt.

Lag og kommuniser standardiserte analyserapporter

Når data har blitt samlet inn, hvordan får vi denne informasjonen til de riktige personene til riktig tid i riktig format?

Kun de mest interesserte vil logge seg inn i webanalyseverktøyet, og det er kanskje heller ikke riktig at alle skal ha tilgang. Det kan derfor vær nødvendig å utvikle standardiserte rapporter tilpasset mottageren, og sørge for distribusjon av rapportene. Forhåpentligvis har det vært gjennomført en KPI-workshop som har avdekket hvem som skal motta rapportene, og hvilke KPIer som skal rapporteres til hvem.

Rapportene må inneholde kontekst, det vil si at man kan ikke bare generere noen rapporter automatisk og sende de til mottageren. For at rapporten skal være verdiskapende og ikke bare en tidstyv for mottageren, må rapporten inneholde tolkninger av informasjonen og forslag til handlinger gjeldende for denne personens fagområde - med andre ord rapporter med action!

Verdiskapende prosess

Det er først gjennom analyse av dataene og handling basert på disse at webanalysen fører til verdiskaping. Det er dette som er webanalyse, men det er her minst ressurser brukes.

Avinash Kaushik mener at 10 % skal investeres i verktøy, og 90 % skal investeres i noen som kan analysere dataene. Eric T. Petersson mener en riktigere tilnærming er at 50 % bør investeres i verktøy, og 50 % bør investeres i mennesker.

Selv om de har en sprikende prosentsats er poenget likevel dette:
Uten en investering i noen som kan tolke dataene og komme med anbefalinger til optimalisering med bakgrunn i disse, gir webanalysen liten verdi.

For at webanalysen skal føre til verdiskaping fordrer det også at man ikke er en rigid organisasjon med en nettløsning som det er umulig å endre på. For som jeg skrev innledningsvis, først når webanalysen benyttes til optimalisering og handling vil den føre til økonomisk verdiskaping og forbedring av brukskvaliteten.

Analyser data og identifiser forbedringsområder

Det er mye og mangt som kan analyseres og antageligvis forbedres. Jeg er tilhenger av å dele dette opp i 4 områder basert på en modell som kalles REAN:

Reach: Markedsføring, annonsering og all kommunikasjon som skjer utenfor nettstedet
Engagement: Brukerens mottagelse på nettstedet (eks. kommunikasjon/innhold, brukervennlighet, design)
Activation: Konverteringen
Nurture: Vedlikeholde kontakten med den besøkende/kunden (eks. CRM, nyhetsbrev, sosiale medier)

Optimalisering

Basert på tolkning av data, hvordan kan vi optimalisere og forbedre det vi gjør? Dette punktet er det viktigste med hele webanalysen fordi at det er her verdien skapes og læring skjer.

Det kan kanskje synes som om det er en lang vei å gå før man får verdi ut av webanalysen, men det er fullt mulig å optimalisere og skape verdi selv om ikke alle forhold er på plass. Noe optimalisering er bedre enn ingen optimalisering.

Men jeg er likevel av den oppfatning av at har man ikke vært i gjennom en øvelse slik som jeg har beskrevet i dette innlegget, og har etablert forankring og prosesser, vil læringen utebli og verdiskapingen bli mindre.

Har du noen kommentarer, prosesser, erfaringer eller læringer som du har funnet nyttige - del de gjerne.

PS. Om du lurer på hvorfor to av stegene i modellen er grå, mens de andre har svart bakgrunn så er det fordi at min oppfatning er at de aller fleste befinner seg på disse nivåene. De har gått rett på innsamling av data og oppsett - men da ikke ut i fra behov siden de har hoppet bukk over forretningsmål og ønsket utbytte.

En KPI er en m�ling, men en m�ling er ikke n�dvendigvis en KPI

Eivind Savio — 2010-01-31

Ordene i overskriften tilhører Dennis R. Mortensen, Director of Data Insights hos Yahoo!.

Dennis R. Mortensen uttaler også:

Å utvikle feil KPI vil gjøre større skade enn nytte!

Men hva er så forskjellen på en KPI og en måling, og hva kjennetegner gode KPIer?

I dette innlegget har jeg samlet noe av det som er blitt sagt og skrevet av andre, i tillegg til noen av mine egne tanker og erfaringer.

Mye som kan måles, men ikke alle målinger kvalifiserer til å være en KPI.
Bilde: D Sharon Pruitt

Hva er en KPI, og hva er hensikten med de?

Det finnes mange definisjoner på hva en KPI er, men jeg forholder meg ofte til Avinash Kaushik sin enkle definisjon:

Målinger som hjelper deg til å forstå hvor bra du gjør det i forhold til dine mål.

KPIer er forenklinger av virkeligheten, og en av de viktigste grunnene til at KPIer benyttes er at man i det daglige har altfor mye informasjon å forholde seg til. Denne forenklingen er både styrken og svakheten til en KPI, og er også grunnen til at det er viktig å identifisere de riktige KPIene.

Hvordan utforme KPIer, og hva kjennetegner disse?

Eric T. Peterson fra Web Analytics Demystified sier dette om KPIer i The Big Book of Key Performance Indicators:

En Key Performance Indicator er alltid andel, gjennomsnitt, prosent eller verdi, de er aldri ubehandlede tall. Ubehandlede tall sier ikke noe om sammenheng, men de fører til dialog slik de hører med som bakgrunnsinformasjon.

Ved utforming av KPIer sier Avinash Kaushik at man skal starte enkelt, og at man skal starte med utbytte. Videre sier Avinash at den eneste grunnen til at vi har et nettsted er dette:

Redusere kostnader
Øke inntekter
Forbedre kundetilfredshet

Jeg finner det veldig hensiktsmessig å ha disse 3 punktene i bakhodet ved utforming av KPIer, fordi det fører til fokus på at KPIen skal si noe om kvalitet og ikke kvantitet. KPIer som fokuserer på kvalitet vil også føre til fokus på de besøkende sine behov eller problemer. For det er jo slik at for at vi skal få oppfylt våre mål, må våre besøkende oppnå sine mål først.

Noen kjennetegn på KPIer som ofte nevnes er disse:

KPIer må være forankret i ledelsen og gjenspeile organisasjonens mål
KPIene må være basert på troverdige data
Begrens antall KPIer, og vær relevant
- For eksempel vil antageligvis Markedsavdelingen og Service/support ha forskjellige KPIer.
- KPIene må knyttes til riktig person i organisasjonen. Denne personen må være inneforstått med hva disse KPIene skal besvare, noe som igjen betyr at denne personen må/bør ha vært med på utformingen av KPIene.
Alle gode KPIer fører til handling!

Om ikke minimum punktene over er oppfylt, vil ikke KPIer hjelpe bedriften/organisasjonen til å yte bedre.

KPI-rapportering

Den/de som er ansvarlig for webanalysen, bør også være ansvarlig for å oversende KPI-rapportene til de personene som skal motta slike rapporter. Man bør med andre ord være proaktiv, og ikke overlate det til den enkelte å logge seg inn i et eller annet webanalyseverktøy og hente ut KPI-rapportene selv.

Grunnen til dette er at en KPI-rapport uten kontekst, altså en forklaring på hva en eventuell endring skyldes, vil ikke så lett føre til handling. Man kan heller ikke forvente at disse personene selv skal kunne finne årsaken til endringene i webanalyseverktøyet.

En god KPI-rapport skal

beskrive sammenhenger
fokusere på endringer
ikke være en presentasjon av tabeller med data

Unike Brukere - mulig KPI eller..?

Til slutt en liten digresjon om temaet KPIer. Jeg støter forholdsvis ofte borti at Unike Brukere foreslås som en KPI. For en god stund siden ble Unike Brukere nevnt som et eksempel på en KPI på den norske Adwords-bloggen, og jeg har også sett andre blogger nevne dette som et KPI-eksempel.

Om du har lest helt ned hit så har du nok skjønt det at dette ikke er en KPI. De færreste nettsteder har etter min mening liten nytte av å bruke Unike Brukere som en overordnet målsetting heller, siden dette ikke er et kvalitetsmål.

I stedet for å argumentere for dette vil jeg anbefale å lese Avinash Kaushik sitt innlegg "Standard Metrics Revisited: Daily, Weekly, Monthly Unique Visitors", samt fortelle en historie fra eMetrics i fjor høst.

Sony Ericsson hadde inntil i fjor vår definert Unike Brukere som en av sine KPIer. Om de oppnådde X-antall unike besøkende, ble det utløst en bonus til de som jobbet med web. I mars i fjor byttet de webanalyseverktøy, samt forbedret kvaliteten på dataene som ble samlet inn.

Resultat: Ingen bonus!

Unike Brukere er ikke en KPI hos Sony Ericsson lengre.

Avslutning

Noe av det som jeg har skrevet om her har jeg også skrevet om tidligere i artikkelen 9 tips for å lykkes med KPIer og webanalyse.

Har du noen erfaringer med KPIer som du ønsker å dele, eller har du noen synspunkter på det jeg har skrevet, legg gjerne igjen en kommentar.

Fargen p� en link skal v�re bl�. Eller? En test av farger p� lenker.

Eivind Savio — 2009-12-06

For lenge siden ble det opplest og vedtatt at en link skal være blå, og den skal være understreket. Disse retningslinjen har blitt kritisert, og mange nettsider har blitt laget siden den tid med all verdens farger på linker. Jeg ønsket derfor å teste denne teorien ved hjelp av Google Analytics og Google Website Optimizer.

I denne artikkelen beskriver jeg hva resultatet av testen ble, og hvordan testen ble gjennomført. Det spesielle med denne testen er at jeg ikke bruker Google Website Optimizer til å evaluere suksess, men baserer meg på dataene som Google Analytics samler inn.

Testens hypotese

Min hypotese før jeg startet å teste var at blå linker burde føre til lavere bounce rate og flere sidevisninger per besøk, siden de burde "stå frem" tydeligere som lenker, og dermed føre til flere klikk. Jeg satte 5 % som suksessmål for testen. Om gjennomsnittlige sidevisninger per besøk ville øke med 5 %, eller om bounce rate ble redusert med 5 % ville jeg anse testen som en suksess.

I testen har jeg testet linker med fargene brun, grønn og blå som vist av bildene under. Testen har pågått i mer enn 3 måneder.

Tekst med brune linker.

Tekst med grønne linker.

Tekst med blå linker.

Bildene av teksten over er hentet fra artikkelen Hvilke linker klikker dine besøkende på?

Testresultat

Resultatet av testen er at jeg velger å forkaste min teori om at blå lenker vil gi lavere bounce rate, men velger å beholde teorien om at blå lenker gir i gjennomsnitt flere sidevisninger per besøk (selv om forskjellen er veldig liten).

Tabellen under viser testens data fra November. Grunnen til at jeg ikke bruker dataene fra hele perioden, er at tallene fra November viser en ting med mine blå lenker mye tydeligere enn hvis jeg hadde vist tallene fra hele perioden (trenden er den samme for alle måneder, bare ikke så tydelig).

Besøkende som så blå lenker brukte i gjennomsnitt lengre tid. Hvorfor det er slik vil bare være spekulasjoner. Klikkdata som jeg har samlet inn gjennom denne testen forteller bare hva som har skjedd, ikke hvorfor.

	Brune linker	Grønne linker	Blå linker
Besøk:	1 214	1 247	1 220
Bounce Rate:	78,43 %	76,66 %	76,80 %
Sider/besøk:	1,46	1,43	1,48
Gj. snitt tid	00:01:08	00:02:13	00:03:35
Nye besøk:	85,67 %	83,88 %	84,26 %

Hvordan ble testen gjennomført?

Data fra Google Website Optimizer dukker ikke automatisk opp i Google Analytics. For å få dette til har jeg benyttet en løsning utviklet av ROI Revolution.

Dette innebærer at man legger inn deres "spesialkode", og setter deretter opp en multivariasjonstest i Google Website Optimizer. Denne koden medfører at det legges på parametere på nettadressene som genereres for Google Analytics. Parameteret som genereres for alle som for eksempel får "servert" brune lenker er ?combo=0.

Dette parameteret kan brukes til segmentering, og jeg har satt opp 3 forskjellige Google Analytics profiler, en for hver lenkefarge.

For å kunne evaluere lenkefargene best mulig har jeg ekskludert en del data fra disse profilene:

Egen IP-adresse.
Data generert fra YouTube-filmer avspilt gjennom min YouTube-spiller.
Utgående klikk på lenker. Å ekskludere utgående klikk kan alltids diskuteres om var riktig, siden de genererer virtuelle sidevisninger og påvirker bounce rate, men dette var et valg jeg tok.

Siden testen ikke er implementert på alle sider på Savio.no, er bounce rate mye høyere enn vanlig. For eksempel er ikke John Savio en del av eksperimentet, og heller ikke eksemplene knyttet til artikkelen The only CSS layout you need?

Var testen en suksess?

Om bare data fra Google Website Optimizer hadde blitt lagt til grunn hadde svaret på dette blitt nei. Konverteringsmålet som jeg satte opp i Google Website Optimizer er bloggkommentarer, og det er ikke statistisk signifikant forskjell i antall kommentarer.

Min hypotese var ikke at det ville bli flere kommentarer med blå linker, men at blå linker ville føre til flere sidevisninger per besøk og lavere bounce rate. Jeg vil derfor kalle testen en delvis suksess siden siden økningen i gjennomsnittlige sidevisninger per besøk økte noe, men ikke nok til at målet på 5 % forbedring ble nådd.

Suksess ved testing kan bli målt på mange forskjellige måter (Kilde: Successful Web Site Testing Practices av Eric T. Peterson):

For noen er "suksess" en dramatisk økning i målinger knyttet til fortjeneste, fordi de vet at de fleste ledere vil respondere på inkrementell fortjeneste.
For andre er "suksess" en liten økning i målinger av de besøkendes engasjement, fordi man vet at små seire over tid til sammen vil være en stor seier.
For andre er "suksess" en reduksjon i antall problemer de besøkende kan oppleve på nettstedet, fordi man man vet at ved å redusere barrierer vil brukeropplevelsen forbedres.
For andre er "suksess" rett og slett å være i stand til å lansere et nytt utseende på nettstedet uten det slår ut negativt på eksisterende Key Performance Indicators.

Har du gjennomført testen riktig, det vil si at du blant annet har en plan, har utviklet en hypotese og har et mål for testen, vil også en test som feiler gi nyttig lærdom. Som Avinash Kaushik sier:

Det fine med eksperimentering og testing er at du lærer å ta feil - fort!

Testing videre

I eksperimentet testet jeg bare lenker som var understreket. Det har blitt sagt at understrekingen av lenkene er viktigere enn fargen på lenkene, slik at det kunne vært neste test.

Jeg har likevel ingen planer om å gjennomføre en slik test siden jeg ikke har noen planer om å innføre lenker uten understreking. Testing for testingens skyld er ikke spesielt nyttig eller god bruk av ressurser.

Google Website Optimizer + Google Analytics = mer lærdom

I dette blogginnlegget har jeg skrevet om en test som Google Website Optimizer ikke kunne gi svaret på fordi testen ikke baserte seg på konverteringer. Men også i de tilfellene hvor du er ute etter et svar fra Google Website Optimizer, kan det være hensiktsmessig å benytte denne metoden. Noen eksempler på dette kan være:

Nye versus Returnerende besøkende: Viser de forskjellig atferd i forhold til testversjonene? Segmentering av dataene i Google Analytics kan gi oss mer innsikt.
Testing av skjemaer: Kanskje fører ikke det nye skjemaet som testes til signifikant flere konverteringer, men i webanalysen ser vi at de bruker kortere tid på å fylle ut skjemaet. En forbedring i brukeropplevelsen kan være grunn god nok.

Kan du komme på andre grunner til at det kan være hensiktmessig å kombinere Google Website Optimizer og Google Analytics, eller har du spørsmål eller hva det nå enn måtte være - legg gjerne igjen en kommentar.

Avslutningsvis vil jeg bare si at husk at denne testen er gjennomført på mine besøkende som har lest mine artikler innenfor min layout med mine farger. Det er ikke sikkert du ville fått det samme resultatet.

Google Analytics Individual Qualification (IQ)

Eivind Savio — 2009-10-03

Mesteparten av bloggingen min her på Savio.no handler om Google Analytics. Grunnen til dette er at jeg er for deling av kunnskap, og det hjelper meg selv å huske spesielle ting bedre. Med andre ord, bloggen fungerer også som en huskelapp for meg selv i mitt daglige arbeid.

At jeg har valgt å skrive så mye om Google Analytics gjør at jeg også stiller meg "selv for hogg". Den 3. mars ble Google Analytics Individual Qualification (IQ) lansert. Samtidig med testen lanserte også Google Conversion University for å hjelpe folk til å forberede seg til testen eller bli en mer kunnskapsrik Google Analytics bruker.

Ut av det blå i dag dukker dermed følgende spørsmål opp på bloggen til Metronet:

...var vi først i Norge eller tok du den i helgen Savio, E?

Om de var først i Norge vet jeg ikke, men de var i hvert fall ferdig med testen lenge før meg. Jeg fikk ikke somlet meg til å ta testen før i kveld, men jeg passerte testen heldigvis med god margin siden jeg følte presset fra Metronet.

All honnør til de som har tatt testen og bestått (klapper meg selv på skulderen). Som Kenneth Gvein fra Metronet skriver i bloggen var testen relativt utfordrende.

Testen koster $50, og består av 70 multiple choice spørsmål som må besvares innen 90 minutter. Du må ha 75 % riktige svar for å bestå. Består du testen er resultatet ditt gyldig i 18 måneder.

Slik jeg ser det vil testen forhåpentligvis føre til at man får etablert en minimumsstandard i forhold til kvalifikasjoner hos personer som leverer tjenester knyttet til Google Analytics.

Har du tatt testen, vurderer du å ta den eller vet du om andre som har tatt den i Norge?

PS. Nei, jeg røper ikke spørsmålene og svaralternativene.

Webanalyse for ledere med Eric T. Peterson og Aur�lie Pols

Eivind Savio — 2009-10-03

Jeg bruker vanligvis ikke denne bloggen til å markedsføre arrangement i regi av jobben min hos Creuna, selv om bloggen hovedsakelig omhandler fagområdene som jeg jobber med der. Det er bare en personlig beslutning jeg har gjort, da jeg synes det er greit å ikke blande inn jobben for mye på min private blogg.

Denne gangen gjør jeg likevel et unntak, fordi temaet er for viktig til å la være og nevne, og det er ikke hvem som helst som er foredragsholder.

Fredag 9. oktober kl 0830 arrangerer Creuna Online Performance Seminar, eller Webanalyse for ledere som det også har blitt kalt. Seminaret er gratis.

Foredragsholdere er Eric T. Peterson og Aurélie Pols, og i tillegg vil Flust.no fortelle om hvordan de bruker webanalyse som et sentralt styringsverktøy.

Hvorfor et seminar om webanalyse for ledere?

Jim Sterne (President of the Web Analytics Association) uttalte nylig:

It was a mistake to call it web analytics, it's really business analysis.

Webanalyse har opp gjennom tiden vært teknologidrevet, og ikke forretningsdrevet. Dette til tross for at webanalyse egentlig handler om å optimalisere businessen. Firmaer som leverer tjenester innen webanalyse har bidratt til at utviklingen har gått i denne retningen - de har vært (og noen er fremdeles) teknologi og verktøyorientert, og ikke forretningsorientert.

For at webanalysen skal bli forretningsorientert og innebære en forbedring av businessen, krever det at webanalysen er forankret i ledelsen.

I artikkelen 9 tips for å lykkes med KPIer og webanalyse har jeg skrevet noen noen praktiske tips knyttet til temaet.

Hva er din oppfattelse av webanalyse i Norge? Handler det hovedsakelig om "klikk og trafikk" fremdeles, eller har bedrifter og organisasjoner begynt å forstå at det handler om å forbedre businessen?

Google Analytics Campaign URL builder with Short URL and Twitter posting

Eivind Savio — 2009-05-13

One of the biggest problems with online marketing campaigns are that the campaigns are not tracked as campaigns. If you do not track your campaigns, you will not know how efficient they are.

Another reason for bad campaign data are that the marketers are creating infunctional campaign URLs because they don't know much about URLs.

In my humble opinion, the Google Analytics tool builder isn't the best tool in the world for creating campaign URLs. Because of this I have created what I humbly believe can be an easier and more time saving solution to create Google Analytics Campaign URLs.

This tool supports

Excel: Export Campaign URLs to Excel to build larger campaigns easier
Short URLs: Create short URLs with Goo.gl and Bit.ly
Twitter: Post shortened URL to Twitter
QR Codes: Create QR Codes of shortened URLs

Installing the Campaign URL builder

Make sure the "Bookmarks Toolbar/Links Toolbar" is visible. If it is not, go to menu View > Toolbars (or something similar, depending on your browser type).
Drag this link: Google Analytics URL builder up to your Bookmarks Toolbar. Alternatively, right click on the link and save to bookmarks.

You are now ready to create campaign URLs, shorten the campaign URL (and share it on Twitter), or export the Campaign URL to Excel.

How to use the Google Analytics Campaign URL builder

Visit your site campaign landing page, click on the Google Analytics URL builder link, and a page looking like the image above will load in front of your page.

Creating the campaign URL

The URL field will be automatically filled out with the URL to your landing page. The Source, Medium and Campaign field are prefilled for a Twitter campaign, and the final Campaign URL is created on the fly. This means that if you are changing some of the values in the different fields, the Campaign URL is automatically updated.

One problem I have experienced with people not very familiar with URLs are that they sometimes uses several ? (question marks) when they create campaing URLs. This tool adresses that particular problem.

Exampel of incorrect URL:
http://domain.com?id=123?utm_source=src&utm_medium=med&utm_campaign=cmp

One URL should only contain one question mark, if not the URL will/may break, and therefore this tool handles this automatically.

Exporting Campaign URLs to Excel

Sometimes it can be more handy to work with your campaign URLs in Excel, for exampel if your campaign includes several sources.

Choose your Microsoft Office language (the tool supports English and Norwegian), and click the Excel button. The Excel spreadsheet created makes it easy to continue to build your campaign, since it is pre-programmed with an Excel formula.

The Excel spreadsheet and the Excel formula is inspired by The Link Tagging Tool (Google Docs tool) from Epikone.

Creating Short URLs with bit.ly or Goo.gl

If you click the Shorten button, you can shorten your Campaign URL using the URL shortener from bit.ly or Goo.gl.

If you want to shorten a URL without the Google Analytics campaign variables, just delete the value in the Source field. All the campaign variables will then be automatically removed.

By clicking the Tweet button you can share your shortended URL on Twitter.

Creating short URLs with bit.ly

If you want to use bit.ly, you will have to register and get your API key. Then do the following:

Copy your API key.
Choose bit.ly from the URL shortener drop down menu.
Paste your API key into the API text field and your username into the username field.
If you click the Save button, your username and API key will be saved into a cookie on your computer, so you don't have to enter this information the next time you uses my Google Analytics URL tool.
Click the Shorten button, and your short URL will be created from the URL in the Campaign URL field. The Short URL created will be visible in the Short URL text field.

Creating short URLs with Goo.gl

The Goo.gl URL shortening integration is not 100 % perfect yet, but I have despite of this launched this option.

If you want to get the shortened URLs listed in your Google Account on the Goo.gl domain, then you have to do the following:

Click the Authorize Access link. The tool is using OAuth-verification.
Allow access to this application if you are logged in to your Google account. This tool will then be granted access for 1 hour.
- If you don't allow access you can still shorten URLs with this tool, but the URLs will not be listed in your Google account.
After you have granted access you will be redirected back to the tool, and the Goo.gl Authorization Token field should automatically be filled out. If the field is still empty, click on the Authorize Accces link again.
If the URL you are going to shorten is missing from the Campaign URL Field after you have been redirected back from the authorization process, just set your cursor in one of the other fields and press for example space (and then delete the space again) or edit one of the campaign parameters.
Click the Shorten button, and your short URL will be created from the URL in the Campaign URL field. The Short URL created will be visible in the Short URL text field.

Things to be aware of

Browsers: The tool should work well in different browsers, but I have experienced some problems with Internet Explorer 6. This is related to the Zero Clipboard Copy function.
Pop-up blockers: Some pop-up blockers (for exampel the pop-up blocker in Google Toolbar) may prevent the tool from loading. If this happens, allow pop-ups from this domain.
Export to Excel: You may see a warning when you try to open the Excel spreadsheet. This happens because I'm forcing a HTML page to act like a Excel spreadsheet. It is totally safe to open the Excel spreadsheet created.
Goo.gl URL shortening: For some reason some times Goo.gl is creating 3 different shortened URLs when you shorten the URL with this tool. I'm working on a fix to this.

Update

05.08.2011: Support for Campaign Term added to the tool. If you are a previous user, please drag the new link:
Google Analytics URL builder
up to your Bookmarks Toolbar.
22.06.2011: Create QR Codes of shortened URLs.
03.04.2011: Support for Goo.gl URL shortening added using jsonlib.
15.01.2011: Support for Cli.gs removed.
13.02.2010: Support for Conversion Attribution added. A conversion can be attributed to the First Click if you want, and not the last one (which is default in Google Analytics). This is done by adding the parameter utm_nooverride=1 to the campaign URL.
See this help information from Google for more information.
13.02.2010: Support for Parameter Separator added. # (hash) separator can be used instead of ? (question mark) in the Campaign URL. This is a more SEO friendly solution. See Google for information about implementation.
09.12.2009: URL shortening with bit.ly added. If you are using my previous version update your bookmark since the new version is having a larger window size.

Some final words

If you find my Google Analytics Campaign URL builder tool useful, please rate it in the Google Analytics App Gallery.

If you have any comments or suggestions, I appreciate your feedback.

Google Analytics: How to track Google Keyword Ranking from your Visitors

Eivind Savio — 2009-04-16

In a blog post from the Google Analytics team a couple of days ago, Google announced some changes to their URLs. Blogstorm did investigate these new URLs further, and asked the question: Google Adds Ranking Data to Referrer String? They were correct about their assumptions.

So - here is how you can track keyword and keyword rankings with Google Analytics, and a couple of things you should be aware of.

The first thing you must do is create a new Google Analytics profile. When this is done we will have to create some filters.

Creating the Google Analytics filters

There may be a simpler method for doing this, but I have only had under a day of testing my solution, so please bear with me.

The first thing to do is to create som filters that includes only visitors from Google organic search.

The Google Analytics filter above includes only organic traffic.

The Google Analytics filter above includes only traffic from Google.

Grab the keyword and add the ranking

The filter below is the vital filter. This filter will grab the keyword, and add the ranking the keyword had when the visitor clicked on it.

The filter above is a modified filter "inspired" from a blog post from Yoast from January about tracking SEO rankings.

This is "all" you have to do to track keywords and rankings from your visitors from Google. After a while the "User Defined" report should look like this.

There are however a couple of things you should be aware of.

1. (For the moment?) not all visitors from Google organic search are sending ranking information. This means that ranking for all of your keywords will not be tracked.

2. If elements from Universal Search (images, videos etc.) is part of the SERP (Search Engine Ranking Page), they will also be counted.

This means that if there for example are 2 images at the top of the result, and you are listed as the last result on the page, your rank will be 12 (2 images + 10 organic listings). However, if you go to page 2 in the result, the first result on that page will have rank 11.

This means that if your rank is 11, you could be listed on page 1, but you could also be listed on page 2. To identify which page the ranking parameter comes from I use Advanced Segmentation.

To be able to to make this segmentation work, you will have to implement a filter that can track Full Referral Path. Don't be confused by the title "How to track Google Searchwiki", the filter is shown in that article.

Google Analytics Advanced Segmentation

Below you will find two different Advanced Segments.

The first segment will identify visitors that came from page 1 in the Google search result.

Matches regular expression:	https?:\/\/(w{3})?\.?google.(&\|\?)q=.
Does not match regular expression:	https?:\/\/(w{3})?\.?google.(&\|\?)q=.&start=[1-9]

The Advanced Segment below will identify visitors that came from page 2 in the Google search result.

Matches regular expression:

https?:\/\/(w{3})?\.?google.*(&|\?)q=.*&start=10

If you want to identify visitors from page 3, just replace start=10 with start=20 in the regular expression above.

These regular expressions are written by Maarten Berge, and has been published here in a Norwegian blog post earlier.

Some final words

Just so I have mentioned it again. I have less than a day with testing, so I don't guarantee that everything are tested good enough. If you spot any faults or make any improvements to the solution, please add a comment.

Happy tracking.

8 CSS (Cascading Style Sheets) eksempler

Eivind Savio — 2009-04-08

På slutten av 2006, begynnelsen av 2007 lagde jeg noen eksempler/skrev noen artikler om CSS (Cascading Style Sheets). Disse artiklene ble publisert på et nettsted dedikert til CSS. Siden dette nettstedet nå er i ferd med "å pakke sammen", har jeg valgt å overføre alle artiklene og eksemplene hit til Savio.no.

Alle artiklene og eksemplene er på engelsk, og er organisert under temaene CSS layout, CSS menyer og CSS tips.

Bakgrunnen for CSS artiklene og eksemplene

Det ligger en tredelt forklaring bak artiklene. Det ene var ønsket om å forstå og lære meg CSS godt nok, og det andre var at jeg ønsket å dele mine eksperimenter med eventuelt andre som kunne finne de nyttig. Det siste punktet var om jeg kunne utforske Digg.com på egne vegne som markedsføringskanal, og lagde derfor artikkelen The only CSS layout you need(?) med tilhørende eksempler. Målet, eller kanskje det er mer riktig å skrive forhåpningen, var todelt:

Se om jeg kunne få egenprodusert stoff til å bli populært på Digg.com
Søkemotoroptimalisering (lenking til artikkelen)

Til tross for at jeg ikke fulgte alle "best practices" i forhold til Digg.com på den tiden, endte det opp med at artikkelen gikk "hot". Den fikk ca. 1 500 stemmer, et ras av besøkende fulgte, og deretter fulgte bølger med besøkende fra Delicious, Reddit og Stumbleupon siden artikkelen også gikk "hot" der.

En kjapp sjekk hos Yahoo! indikerer at ca. 1 100 lenker peker til denne artikkelen fra andre nettsider.

Hvilke konsekvenser har det å videresende disse artiklene til Savio.no?

Jeg skriver hovedsakelig på norsk, og da først og fremst om webstatistikk. Selv om jeg etterhvert har fått noen regulære lesere, kommer trafikken hit hovedsakelig fra søkemotorer. At jeg nå har videresendt disse artiklene til Savio.no innebærer at det blir mye "støy" i webanalysen, disse 8 artiklene sender en god dose besøkende.

De viktigste besøkende for meg er norske lesere, det de søker etter og deres atferd. Disse er det nå vanskeligere å identifisere gjennom all "CSS støyen". Jeg har derfor satt opp et avansert segment (som jeg sikkert må jobbe litt med) som ekskluderer CSS trafikken.

Søkemotorer, videresendinger og påskenøtter

Det er alltid en viss spenning knyttet til videresendinger fra et domene til et annet, og hvordan søkemotorene håndterer dette. Om du ikke har gjort dette før er du herved velkommen til å kunne observere hvordan det går.

Først, artiklene er videresendt permanent (301-videresending) hit til Savio.no. Det vil spøkefullt si at jeg har sendt en flyttemelding til Google og de andre søkemotorene om at Savio.no "skal" overta de eksisterende rangeringene. Du kan følge med på søkefrasen css submenu som jeg innehar førsteplassen på med det gamle domenet for å se hvordan det går.

Påskenøtten er ikke et vanskelig spørsmål, men en oppfordring om at om du har laget noen egne CSS løsninger, legg gjerne igjen en kommentar under.

PS. Siden jeg allerede har begynt å få spørsmål om dette; Nei det gamle domenet er ikke til salgs.

Hvilke linker klikker dine bes�kende p�?

Eivind Savio — 2009-03-30

Klikker dine besøkende på overskriften eller les mer linken når de klikker seg inn til en artikkel, eller hvilke av menyelementene på nettsiden din benyttes til navigering? Beklager, det kan ikke Google Analytics fortelle deg, i hvert fall ikke "rett ut av boksen".

Bildet nedenfor av nettstedsoverlegget her på Savio.no illustrerer poenget. Fra forsiden av nettstedet er det tre menylenker som fører til bloggen, Google Analytics rapporterer at alle tre har like stor prosentvis andel av klikkene.

Så - hvordan identifisere hvilke lenker de besøkende har klikket på?

For at du skal vite hvilken lenke som har blitt klikket på må lage unike lenker. I artikkelen Hvordan spore Hash-tegn i URLen med Google Analytics nevnte jeg at hash-tegn kan benyttes til å lage unike lenker, og at dette kan være en måte å gjøre det på siden dette ikke skaper problemer for søkemotorene i forhold til problematikken rundt duplisert innhold. Problemet med hash-tagging er at det ikke fungerer med Google Analytics sitt nettstedsoverlegg.

Metoden for å identifisere hvilke linker som har blitt klikket på, og som fungerer med Google Analytics sitt nettstedsoverlegg, krever at hver lenke tagges opp med et parameter.

Dette er en metode som jeg allerede ser at benyttes i Norge, men alle eksempler jeg har sett er utført på feil måte siden de skaper (potensielle) problemer for søkemotorene på grunn av duplisert innhold (noen "ledende" SEO selskaper som føler seg kallet til å rekke opp handa?).

Problemet gjelder ikke bare for nettsteder som benytter Google Analytics. Det finnes flere eksempler i Norge på nettsteder som har investert i dyr webanalysesoftware, tagget opp rubb og stubb, og dermed sikret seg en dårligere organisk synlighet.

Derfor, her er hvordan jeg mener at dette skal gjøres.

I bildet av nettstedsoverlegget i begynnelsen av artikkelen er det tre lenker som fører til bloggen. Jeg velger å benytte parameteret linkID for å separere lenkene fra hverandre. Toppmenyen får parameteret toppmeny, kategorilenkene i venstre kolonne får parameteret venstrekolonne, mens lenkene i høyre kolonne får parameteret hoyrekolonne. Det gir meg da tre forskjellige lenker:

http://www.savio.no/blogg/?linkID=toppmeny
http://www.savio.no/blogg/?linkID=venstrekolonne
http://www.savio.no/blogg/?linkID=hoyrekolonne

Jeg har nå tre forskjellige lenker til samme destinasjon - lenker som vil fungere med nettstedsoverlegget til Google Analytics. Men som nevnt, disse lenkene vil kunne føre til problemer med duplisert innhold og dårligere rangering i søkemotorene.

Hvordan unngå problemer med duplisert innhold med denne metoden?

Løsningen er å benytte en spesiell <link> tagg kalt canonical. Denne taggen er et samarbeid mellom søkemotorene for å redusere problematikken med duplisert innhold, og støttes av både Google, Yahoo!, Live Search (MSN) og Ask.com. Taggen fungerer for søkemotorene omtrent som en 301-videresending.

For å fortsette eksemplet mitt ovenfor. På nettadressen til bloggen legger jeg følgende kode inn:

<link rel="canonical" href="http://www.savio.no/blogg/" />

i <head> seksjonen på nettsiden.

Da gjenstår bare en ting. Det er ikke bare for søkemotorene at duplisert innhold skaper problemer. Forskjellige URLer som fører til samme destinasjon skaper også dårligere kvalitet på dataene dine i webanalysen. Du bør derfor ekskludere denne informasjonen fra de fleste nettstedsprofilene dine, og for eksempel bare la disse dataene være tilgjengelig i din ufiltrerte nettstedsprofil (Se artikkelen Hvordan organisere en Google Analytics konto).

For å ekskludere linkID informasjonen kan man enten benytte filtrering, eller man kan legge inn denne ekskluderingen i profilinformasjonen. Bildet under illustrerer hvordan dette gjøres i profilinformasjonen.

Du har nå en metode for å spore hvilke lenker de besøkende klikker på som fungerer med Google Analytics sitt nettstedsoverlegg, og som ikke lager problemer for søkemotorene på grunn av duplisert innhold. I tillegg har du ivaretatt kvaliteten på dataene dine i Google Analytics.

Siden metoden krever noe jobb og kanskje involvering av flere selskaper (eks. webleverandøren), er det viktig at man har tenkt igjennom om denne metoden vil gi de svarene man ønsker. Kanskje er det mer hensiktmessig å benytte analyseverktøy beregnet for klikk-tracking slik som CrazyEgg, ClickTale, ClickDensity eller Pagealizer.

Noen ord til slutt om nettstedoverlegget i Google Analytics

Jeg er ingen aktiv bruker av nettstedoverlegget i Google Analytics, rett og slett fordi jeg føler at det er for mange "svakheter". Mange liker likevel å bruke nettstedoverlegget da det kan være lettere å få en forståelse av blant annet navigeringen på nettstedet. I tillegg finner man også informasjon om transaksjoner, målkonvertering og inntekter som kan være veldig nyttig.

Likevel mener jeg at nettstedoverlegget har svakheter. For å illustrere hva jeg mener.

Om du legger igjen en kommentar til denne artikkelen (setter pris på tilbakemeldinger så gjør gjerne det), vil ikke dette vises i nettstedoverlegget. Grunnen til dette er at kommentarfeltet benytter AJAX. For å kunne spore dette i Google Analytics lages det virtuelle sidevisninger. Virtuelle sidevisninger vil ikke vises i nettstedoverlegget, ei heller for eksempel Flash-lenker. Filtrering kan også føre til at nettstedoverlegget ikke fungerer slik det skal.

Synspunkter på artikkelen, sporing av klikk på lenker, nettstedoverlegget eller gode verktøy for klikk-tracking? Lag en virtuell sidevisning og legg igjen en kommentar.

Microsoft adCenter Analytics legges ned

Eivind Savio — 2009-03-12

Microsoft har akkurat annonsert at Microsoft adCenter Analytics legges ned.

Jeg har hatt løsningen installert her på Savio.no omtrent siden starten. Inntrykket mitt fra den første artikkelen om verktøyet har ikke endret seg siden, det ble aldri et verktøy jeg fant verdt å bruke tiden på - funksjoner og brukergrensesnitt passet aldri mitt behov.

Jeg har likevel et godt minne fra verktøyet, som jeg også delte med noen av de som skriver på Frilunsj på en en uformell samling i kveld, og det er at verktøyet opplyser om demografiske data.

Om jeg ser bort i fra at demografiske data er svært mangelfulle, er det hyggelig ut i fra mitt eget ego at Microsoft adCenter Analytics rapporterer at flesteparten av mine lesere er kvinner mellom 20 og 26 år.

Dennis Mortensen, Director of Data Insights hos Yahoo, har skrevet noen interessante betraktninger i forbindelse med nedleggelsen i bloggen sin. Disse tar blant annet for seg forskjellen mellom de gjenværende aktørene som det er verdt å snakke om på området gratis webstatistikk; Google Analytics og Yahoo Analytics.

Why do you think Google Analytics provides “Last Click Campaign Attribution“ and Yahoo! Web Analytics provides multiple models including “Original Source Campaign Attribution”? - Because last click favors Search and Google owns search and thus present a positive picture of their channel, Original Source on the other hand favors Display and Yahoo! owns Display and thus present a positive picture of their channel - same data, two different stories.

Jeg sier som Ole Brumm: Ja takk, begge deler - jeg vil vite både Original Source og Original Source Campaign Attribution. Jeg vet det finnes en mulighet for å finne ut av dette i Google Analytics, men har ikke testet det ut enda da det krever en del programmering.

I stedet for å snakke om Microsoft adCenter Analytics sin nedleggelse, lurer jeg derfor på om hva du mener er mest riktig om du måtte velge:
Last Source Campaign Attribution eller Original Source Campaign Attribution?

Hvordan spore Hash-tegn i URLen med Google Analytics

Eivind Savio — 2009-03-09

Hash-tegn (#), av noen kalt firkant eller anker, benyttes i blant i nettadresser, for eksempel slik som vist i eksemplet under:

http://www.savio.no/nettadresse#comment

Eksemplet over viser en anker-lenke, men det er ikke bare anker-lenker som benytter hash-tegn. Enkelte nettsider som benytter Flash eller AJAX kan også benytte hash-tegn i nettadressene til å kalle opp spesielle funksjoner.

Google Analytics sporer ikke hash-tegn i nettadresser til vanlig. Denne artikkelen tar for seg hvordan man kan spore hash-tegn i nettadresser med Google Analytics.

For å kunne spore hash-tegn i nettadresser med Google Analytics, må vi endre Google Analytics koden på nettsiden til koden under. Bytt ut XXXXX-1 med din Google Analytics konto- og profilID.

<script type="text/javascript">
var gapathname = window.location.pathname, gasearch = window.location.search, gahash = escape(window.location.hash);
try {
var pageTracker = _gat._getTracker("UA-XXXXX-1");
pageTracker._trackPageview(gapathname + gasearch + gahash);
} catch(err) {}
</script>

Denne koden er hentet fra Google Analytics sitt diskusjonsforum.

Det som er verdt å merke seg er når du sporer hash-tegn i nettadressene, vil du få duplisert innhold i rapportene dine i Google Analytics. Samme nettadresse vil nå kunne forekomme flere ganger i rapportene. Dette gir dårligere kvalitet på dataene, og bør som regel filtreres bort fra (minimum) hovedprofilen din i Google Analytics. Du bør kun la en profil dedikert til formålet samle inn hash-tegn dataene.

Om du ikke forstår hva jeg mener med hovedprofil og flere profiler, les artikkelen min om Hvordan organisere en Google Analytics konto.

For å filtrere bort hash-tegnene og all informasjon som kommer etter disse i URLen, legg inn filteret under i profilene i Google Analytics hvor hash-informasjonen skal ekskluderes.

Om du er litt kodekyndig så du kanskje koden benytter escape. Det vil si at # vil vises som %23 i rapportene. Fra et rapporteringsperspektiv er ikke dette så gunstig, og jeg foretrekker derfor å kjøre et søk og erstatt filter på dette. Søk og erstatt filteret er vist under.

Det er verdt å merke seg at selv om dine rapporter nå vil vise hash-tegn, vil ikke disse vises i nettstedsoverlegget i Google Analytics.

Søkemotorer og hash-tegn i nettadresser

Det er verdt å merke seg at søkemotorer ignorer det som kommer etter hash-tegnet i nettadressen når de indekserer nettsider. Hvis for eksempel nettadressen din er http://www.domene.com/no/#/nettadresse, vil søkemotorene ignorere alt etter no/ i nettadressen. Å benytte hash-variabler i nettadressen skaper dermed ikke problemer med duplisert innhold for søkemotorene.

På grunn av dette blir hash-variabler også benyttet til å spore hvilke lenker de besøkende klikker på. For eksempel, klikker dine besøkende på overskriften eller les mer linken når de klikker seg inn til en artikkel? Selv om man kan benytte hash-variabler til å spore dette, er metoden er likevel ikke helt ideell, siden som nevnt hash-variablene ikke vises i nettstedsoverlegget.

For å kunne se hvilke lenker de besøkende har klikket på i nettstedoverlegget, se artikkelen Hvilke linker klikker dine besøkende på?

SEM konferansen 2009

Eivind Savio — 2009-02-09

Den 23. og 24. september drar INMA og ANFO i gang årets utgave av SEM konferansen. Selv om jeg er litt inhabil i min anbefaling av denne konferansen (jeg sitter i Fagpanelet) mener jeg dette helt oppriktig - dette er årets viktigste arrangement innenfor internettmarkedsføring i Norge.

Under har jeg sakset noe om temaet og konferansen fra INMA.

Søk-bransjens rockestjerner til Norge!

Årets arrangement har samlet de fleste av verdens og nordens eksperter innenfor fagfeltet, og vi ønsker alle de 260 deltakere fra 2008 samt nye deltakere velkommen.Eliten av britiske, amerikanske og nordiske eksperter er på plass!

Noen som kommer er:

Bryan Eisenberg, guru, forfatter og en av iMedia top 25 markedsførere.
- I tillegg har han skrevet flere bøker, blant annet "Always be testing" som jeg anbefaler på det sterkeste. Om noen kan få konvertert noen på nett så er det han. Selv om konverteringer ikke bare gjelder nettbutikker, bør han være grunn nok til at i hvert fall alle nettbutikkeiere bør kjenne sin besøkelsestid.
Rand Fishkin, grunnlegger og høvding i SEOmoz og en av de største stjernene i SEO-bransjen, og holder alltid engasjerende og lærerike foredrag.
- Jeg har fulgt med på hva Rand Fishkin har holdt på med siden 2004, så når jeg anbefaler å dra på konferansen på grunn av han er det ikke uten grunn.
Mikkel deMid Svendsen, "Jeg er en super søkenerd" sier han om seg selv. Mikkel har skrevet SEO 2.0 - den mest omfattende boken om markedsføring i søkemotorer.
- Jeg hadde gleden av å høre og møte Mikkel på SMX London i 2007. Han er en fargerik (også bokstavelig talt) og dyktig fyr. I tillegg vil han nok kanskje utfordre noen holdninger knyttet til SEO om han holder vanlig "standard".
Lisa D. Myers er en av Storbritannias flinkeste SEO-eksperter og har vært en anerkjent blogger i flere år. Hun holder regelmessig foredrag på de største konferansene ute i verden.
- Og i tillegg er hun trønder- og en knakende dyktig en og!
Marc Saarde, Head of Online Business Optimization, Creuna Danmark.
- Marc kjenner jeg godt. Er det noen som kan webanalyse i Norden så er det Marc.
Andrew Goodman. Andrew er blogger, forfatter og SEM-ekspert fra Canada og har skrevet den beste boken om Google AdWords: "Winning Results with Google AdWords".
Adam Lasnik, Google sin første Search Evangelist.
Anders Hjorth, CEO, Outrider EMEA.
Jesper Kärrbrink, President og CEO, ENIRO.

Noen av temaene og foredragene på SEM konferansen er:

25 tips til bedre SEM.
En full gjennomgang av SEO.
Lær hvordan du får til en effektiv forankring, bruk og optimalisering av webanalyse.
Kunnskap om innhold som både brukerne og søkemotorene elsker.
Norges "Beste Søkestrategi". En detaljert gjennomgang og presentasjonen av Gulltaggens og Norges "Beste Søkestrategi".
En gjennomgang og presentasjon av "Beste Internettstrategi", som gjennom bruk av Søk, sosiale medier og viralkampanjer gikk av med seieren i kategorien Beste Søkestrategi på Gulltaggen 2009.

I 2008 hadde online annonsering en vekst på 20% i Europa - De beste fremtidsutsiktene er for søk

I følge IAB AdEx, en årlig annonseomsetningsrapport fra 19 nasjonale IAB land tilrettelagt i samarbeid med PWC, er søkemotormarkedsføring den grenen innenfor online annonsering som vil fortsette å dominere i tiden fremover. Søk hadde en vekstrate på 8% i Europa i 2009, og representerer nesten halvparten av den totale annonseomsetningen (46 %).

Samme analyse antyder at display ad annonsering vil fortsette å minske sin vekst, med en total vekstrate på 0,4 %. Mer informasjon om dette finnes i artikkelen Internett opp 20 % i Europa på INMA.no.

Selv om vi er inne i en noe ruskete økonomisk tid, er jeg av den oppfatning at og ikke finne midler til å delta på SEM konferansen er synonymt med å spare seg til fant. Søkemotormarkedsføring er den mest kostnadseffektive og brukerorienterte markedsføringsformen som finnes, og SEM konferansen gir en unik mulighet til å tilegne seg kunnskap fra de beste.

Sees vi på SEM konferansen?

Savio.no - Blog

Tracking Content Scrollers, Scanners & Readers in Google Analytics

Differences between this script & Justin's script

Bonus: Analyzing your content in Excel

Using the Excel spreadsheet

The Google Analytics script for tracking Readers & Scrollers

Some final words

Related articles

Behavioral Targeting using Facebook Connect & Google Analytics Custom Variables

This blog post covers:

Create a Facebook App

Facebook Connect

Tracking Age Groups & Gender as Custom Variables in Google Analytics

Behavioral Targeting using Google Analytics Custom Variables

Reactivating Facebook Connect Tracking

BTBuckets Behavioral Targeting

Google Analytics Events

Facebook Connect Demo

Download Example Code

Some final words

Related articles

Tracking Internal Search using POST request method in Google Analytics

Some example HTML code

The Javascript/jQuery solution for tracking Internal Site Search

Filter Name - Exclude Page: search.aspx

Site Search Setting

Javascript Redirect & Google Analytics Cross Domain Tracking

Tracking Facebook Likes, Share, Send & Comments as Social Engagement in Google Analytics

Facebook and Open Graph Meta Tags

Facebook AppID and AdminID Meta Tag

Open Graph Protocol (Meta Tags)

Facebook XFBML version (Javascript SDK)

Facebook channelURL

Google Analytics Social Media Tracking Script for Facebook

Facebook Share & Send Button

Facebook Like Box

Facebook Comment Box

Related articles

How to Track Traffic from Facebook Likes, Send or Comments in Google Analytics

The road to better insight

Facebook Like & Send

Where on Facebook did the Link Get Clicked?

Facebook Comments

The Google Analytics script

Google Analytics Filters for Facebook Likes, Send & Comments

Facebook Like, Send & Comments Campaign Medium Filter

Facebook Like, Send & Comments Campaign Source Filter

Facebook Comments Campaign Name Filter

Google Analytics Report Explanation

Facebook Campaign Name Report in Google Analytics

Facebook Campaign Ad Content Report in Google Analytics

Cleaning up your data

Some final recommendations

Related articles

How to Track Facebook iFrames with Google Analytics

Handling the Traffic Source Problem

Google Analytics tracking code for Facebook iFrame Tab Pages

Why Internet Explorer Privacy Settings can make your life miserable

The solution: P3P (Platform for Privacy Preferences)

P3P Header for ASP.NET using global.asax

P3P Header for "old" ASP

P3P Header for PHP

P3P Header for PHP using .htaccess

Related articles

How to Track Multiple Links to the same URL automatically in Google Analytics

How to Automatically track Multiple Links to the same URL

Implementing the script

The Link Click Report

Visualising the data

Why not just use a Link Analysis/Heat Map tool?

What to be aware of

The Javascript

Link ID

Anchor Text

Bounce Rate and Time Spent

Download the Google Analytics Script and Excel Report

Website Optimizer Asynchronous Integration with Google Analytics

Test Page Code

Conversion Page Code

Page Sections