If you’re into A/B testing then “Experiments” in Google Analytic are a great way to go about it. However one of the drawbacks in experiments is you can only use measure one goal during the experiment. This is alright in most cases because you’re usually testing one thing like CTR or number of signups. But once in a while you need to come up with ways to track multiple goals for a given experiment. We can’t connect multiple goals to an experiment, but what we can do is adjust the goal/create a new goal to measure multiple outcomes.
Setting up a goal using regular expressions
If you want to master Google Analytic then mastering regular expressions is equally important. It can get you through those tricky situations more where filtering data can be a pain. However in this situation we’ll use it to create a goal with multiple outcomes.
Go to Admin -> Goals in Google Analytic and click “New Goal“. You can select a template or create a custom goal. Regular expressions are available only for event types “Destination’ and “Event”. In this example I’ll use a destination event type.
When filing out goal details select “Regular expression” for destination. Now you can combine multiple outcomes for the destination. For example lets say someone can signup for different plans using your signup page and all those plans have different destination pages. Then you can combine them using something like below.
If you’re not certain about the setup you can verify this by going to Behavior -> All Pages and copy and pasting expression on the search bar. If the resulting data is what you want then you’ve setup the goal properly. Now you can setup the experiment using this goal.
Drawbacks and potential problems
As you can probably guess by now this is a great way to find totals but not so great to break it down by different channels. If you take the above example you can easily find the total number of signup. But you cant measure it for individual plans.
One can argue you can setup different goals to measure them separately. This is true to certain extent but there can be complicated situations. Consider the above example, if you’re running the experiment on the plans page the new variant might show an improvement in “team5” plans. But the number of people purchasing “team1o” plans might have gone down considerably. So in situations like this you should be very careful about coming to conclusions based on experiment data.
More methods to track multiple goals
However if its critical to capture the outcomes separately without combining them then there are other methods to track multiple goals with Google experiments.
This method makes use of segments and it is very clearly explained in this article at State of Digital. The drawback of this method is that your content reports will get messes up because you’re not consolidating. Google Analytic will treat variations as different URLs and they will show up as different URLs in content reports.
You need to figure out what’s more important for you and make the compromise.
Alternatively you can use specialized A/B testing tools like Visual Website Optimizer, Optimizely etc which support multiple goals in a single experiment.