Tracking an A/B Test Created with Your Own Server-Side Code

What if you are already running A/B tests with code that you’ve written? It’s just a simple matter of adding Kissmetrics tracking to these pages.

Let’s quickly go over again what’s important in an A/B test:

What Every A/B Test Needs

Each visitor is randomly assigned to a variation, either A or B.
Make sure that the visitor always sees either only A or only B. Otherwise, your results are tainted if someone ends up seeing more than one variation.
Record which variation this visitor saw, so that you can refer to it when checking the results of the A/B test.
We’re assuming that your setup is already handling #1 and #2, which just leaves setting up #3.

Record the Variations as Kissmetrics Properties

Use our set command, indicating the name of the experiment and the selected variant. For example, suppose you’re testing different signup button colors. Let’s say you have a server-side variable @color that your A/B testing code provides. Here’s how you could tell Kissmetrics about that:

<script type="text/javascript">
  _kmq.push(["set", {"Signup Button Color": "<%= @color %>"}]);
</script>

Later, when you are reporting on the test’s performance, you’ll be segmenting the report by the property Signup Button Color.

Record Kissmetrics Events Normally

Remember, the point of the test is to see if it influences how your customers reach some end goal. Naturally, you’ll want to record the end goal that you are interested in.

Just make the standard API calls to Kissmetrics to record your events. For example:

<script>
  // Record the button click with Kissmetrics
  _kmq.push(["record", "Viewed Homepage"])
</script>

Then:

<script>
  // Record the button click with Kissmetrics
  _kmq.push(["trackClick", "signup_button", "Clicked Signup"])
</script>

With these events, you could report on the conversion from people who Viewed Homepage -> Clicked Signup, segmenting by the property Signup Button Color.