Create segments for experiment variations
You can create segments of the traffic to each of your variations. While the primary focus of an experiment is to find the variation that gives you the specific improvement you want (e.g., more goal conversions, lower bounce rate), you can also examine how behavior differs for each variation to see whether there are other advantages or disadvantages to a variation. For example, while the focus of an experiment might be to increase conversions for a specific goal, you might find that users of another variation spend more money overall.
When you create the segment, you need to configure it so that the experiment and variation are included in the same hit, as outlined in the example below.
To create a segment for an experiment variation:
- Open the experiment list, and click the name of the experiment for which you want segments.
- At the top of the experiment report, click Edit settings.
- Under Configure your experiment, you can see the order in which the different variations were added. This order corresponds to the variation number you need when you create your segments. The original page is 0, and the subsequent variations are 1, 2, 3, etc.
- Under Setting up your experiment code, click View the experiment code.
- Below the code, find the Experiment ID, and make a copy of the ID.
- Click Back to experiment.
- In the experiment report, open the segment list, then click + Create New Segment.
- Create the following type of segment for each variation:
- Select Advanced > Sequences.
- Filter: Include Sessions or Users (depending on the information you want)
- Sequence start: Any user interaction or First user interaction (depending on the information you want)
- Step 1:
- Experiment ID exactly matches <your experiment ID>
- Variant exactly matches <the experiment variation you want to match>