Integrate Convert Experiences with Google Analytics

Introduction

You can now have your experiments' data and statistics sent from Convert Experiences inside Google Analytics (Classic, Universal or Global Site Tag) and this way you are able to build rich reports that can bring valuable insights about how specific experiments perform on certain segments of users.

For example, you could see how are your variations performing in a certain browser, or how are they performing for visitors that are coming from Adwords.

What is the difference between Classic Analytics, Universal Analytics and Global Site Tag?

Before we start with the integration, let's clarify the different versions of tracking codes in Google Analytics.

In November 2005 Google launched Google Analytics, known also as 'Classic Analytics'. The tracking code can be found by the name ga.js. From 2016 it is considered Legacy and most of the accounts have been migrated to Universal Analytics.

In October 2012 another new version of Google Analytics was announced, called 'Universal Analytics'. The tracking code can found by the name analytics.js. The key differences from the previous version were: cross-platform tracking, flexible tracking code to collect data from any device, and the introduction of custom dimensions and custom metrics. 

In October 2017 the newest version of Google Analytics was announced, called 'Global Site Tag'. The tracking code can be found by the name gtag.js. Its purpose is to unify the tagging system to simplify implementation. 

Enable the Integration in Convert

To enable the Google Analytics integration on your Convert Experiments tested website, you need to do the following:

  1. Make sure that Google Analytics tracking code is placed under Convert Experiments code, in your website's header. You can use the Classic Analytics (ga.js) or the Universal Analytics code (analytics.js) or the Global Site Tag (gtag.js) version. Any custom Google Analytics code could break the integration.
  2. After setting up the experiment variations, go to Experiment Summary by editing the experiment. Once you are in the Experiment Summary, go to Settings section and then select Add/Edit Integrations: 
    You can either select Classic or Universal Analytics in order to activate the integration. If you use Global Site Tag in your website select Universal Analytics as it works for both Universal and Global Site Tag tracking codes:

     

Integrate with Classic Analytics 

Integrate through the use of Google Analytics Custom Variables - this is the choice if you are using the Google Analytics Classic tracking (ga.js). In that case you need to select one of the five available Google Analytics slots. 

Make sure you do not have multiple experiments running in the same time using the same slot. Reusing a slot for experiments that run in different periods is fine though.

Integrate with Universal Analytics

Integrate through the use of Google Analytics Custom Dimensions - this is the option needed if you are using the Google Analytics Universal tracking (analytics.js). In that case, you need to first create inside your Analytics account the Dimension you would like to use here. 

Create Custom Dimension in GA for Universal Analytics

In Google Analytics you can create Custom Dimensions. Go to Admin, select the Account that you are interested in and then select Property. Under Property select Custom Definitions and then Custom Dimensions:

Select the +New Custom Dimension:

Give a descriptive name to your Custom Dimension since that will be used inside your reports (preferably give the Convert experiment_id as name to find it easier). Select User for Scope and then hit Save. (If you want to have more information on why we select User for Scope check the different values here: https://support.google.com/analytics/answer/2709828?hl=en).

Notice that in the table with Custom Dimensions, your new Custom Dimension is now listed and it has an Index value. 

The index attached to each Custom Dimension should be the same as the index we use in Convert (e.g. if you want to use Experiment #10023753 Custom Dimension which has Index value 1, you should select Dimension 1 in Convert):

Integrate with Global Site Tag

Use the steps as described in the previous section for Universal Analytics.

See Convert data in GA reports  

After setting up this integration, the test data will start appearing inside Google Analytics as follows.
 

Convert Data in Classic Analytics

If you are using the ga.js tracking and therefore the custom variables, the data will be visible in two locations:
  1. Under Audience --> Custom --> Custom Variables, there will be a variable named CONVERT-XXXX where XXXXX will be the experiment ID inside Convert Experiments. You can see the Convert experiment id in the address bar when you edit your experiment. 
  2. Under that variable, you will see your testing combinations. 
  3. On each pageview of a tested page we also send an event to Google Analytics. You can find those under Content --> Events inside your Google Analytics. You can use the events to see reports of how many times a variation has been seen (not just by how many visitors). 

Convert Data in Universal Analytics

If you are using the Universal Analytics tracking method and therefore the Custom Dimensions, you can use the custom dimensions inside custom reports or inside segments (to filter standard reports).

Go to Customisation and then Custom Reports:

Create a new Custom Report, and under dimensions select the Custom Dimension you created in the previous step:

Once the report has been created, you should wait until you see the variations and data coming in. Google says that there could be a 24-28 latency.
But eventually it should look like this. Each variation of the Convert experiment that you use should have a corresponding row in the Custom Report:

Convert Data in Global Site Tag

If you are using the Global Site Tag tracking method and therefore the Custom Dimensions, you can use the custom dimensions inside custom reports or inside segments (to filter standard reports).

You can follow the steps as described in the previous section for Universal Analytics.

 
 
Have more questions? Submit a request

Comments