Google Content Experiments – Tracking Multiple Goals
You may be familiar with Google Analytics Content experiments since the projects transition from the former “Google Website Optimizer” project. If you’re not I would suggest reading one of the articles below and bookmarking this Article for later:
- http://www.stateofdigital.com/google-analytics-content-experiments/– By Michel Wester
- http://analytics.blogspot.co.uk/2012/06/helping-to-create-better-websites.html – By Nir Tzemah
One of the current limitations with Google’s content experiments is that you can only track one goal. If your like me then this can cause potential issues as I work with a many sites which contain pages with multiple goals i.e.
- News letter sign ups
- Different Forms
- Affiliate links
- E commerce Products
Tracking Multiple Goals using Content Experiments
After much research I came across a Google Plus post by +Justin Cutroni (Google) on Multi-Goal Measurement in Google Analytics Content Experiments. This was enlightening to me and so I thought I’d put the principles into practice.
Step 1 – Goals
The most important part of Analytics is making sure your goals are set up correctly. In my experiment I’m using an Affiliate website and so there is no actual goal other than users clicking on one of the ads. So I’ve used event tracking and then created goals based on the events.
Not familiar with this? Then please read:
- Event Tracking: http://www.seobook.com/tracking-micro-conversions-with-event-tracking
- Using Event Tracking for Goals: https://www.blastam.com/blog/index.php/2011/03/how-to-use-events-goals-google-analytics/
Step 2 – Pages to test.
In order to test conversion changes you need to create multiple pages against an original, if you unfamiliar with what to test, this article is probably too far down your journey, you will find some useful insights about testing conversion below:
Step 3 – The Content Experiment
If you read the article from Michel Wester you’ll know how to set up a standard experiment, however this time we are going to make a slight change as follows:
There are four parts to setting up a new content experiment of:
- Choose Experiment Pages
- Set Experiment options
- Add and check experiment code
- Review experiment
In “Set experiment options” you need to ensure that “Rewrite variation URLs to original in Content reports” is left un-ticked.
If you tick the option then what happens is Google will treat all visits to each individual page (i.e. the variations) as just being to the original page, in our instance we want to make sure they are separated so we can use advanced segments to compare the data.
Step 4 – Setting up Advanced Segments
Now all we do is simply create three custom advanced segments i.e. one for each page:
- The Original
- Version One
- Version Two
If you not familiar with Advanced segments you can create the above as follows:
- Include: Page
- Exactly Matching: Your Page location
You’ll need to do this for as many variation pages as you are testing (In this example two variations).
Step 5 – Done
Your set up should now be complete and you simply have to let the content experiment run.
Note: You do have to start the content experiment, although you may not actual use the report it gives you. My personal advise is choose your key goal for the experiment as we will see the overall goals using segments.
Step 6 – Analysis
Now is where the fun starts, can choose your segments against a date range.
Note: Any conclusion you may have will be manual; Google will not tell you “A WINNER” you have to make the choice yourself, however you will see different data to what you ascertain is the overall winning format.
The Results – Step 1
- Choose your segments as above.
- Choose Goals >> Overview
You will see the goals completed for each experiment page against the original and the overall conversion rate (i.e. Visits v Conversions)
Note: In the example it is showing two days worth of data, this experiment is still running, but illustrate how it works.
Overall Conversion Rate
Analytics will give you an overview of the conversion rate for all Goals. As you can see from our initial data Variation two shows signs of being a winner. (It is too early to be precise).
However remember that this is the overall conversion, you probably want to dig into specific conversion of your individual goals for that page. Well Guess what? We can now do that.
- It won’t give you the individual conversion rate, but can work this out by exporting the traffic to those pages and the conversions. (Or I’m sure you can create a custom report)
- We can quickly visualise how many conversions each goal has for each page. (Also we setup in the original experiment to split 100% of the traffic.)
I think being able to track multiple goals is a very good advancement for Google Content experiments even if this is a workaround. Using the above techniques with event tracking will allow you to test more webpages for conversion.
Now its your time to try, let me know how you get on.