A/B Test Segmentation
Split segments into mutually exclusive groups to test the performance of different targeting platforms or algorithmic models against your KPIs, so that you can ensure you’re hitting success metrics.
Don’t just test your designs, test your platforms.
Creating campaigns takes time. And money. You need to know which destinations deliver results. But testing gets complicated. At first glance, one destination might provide better results than another. The results, though, might be skewed and don’t account for disproportional variables in a group like device type.
Traditional A/B testing pits one creative against another. But the Audience Lab feature in Audience Manager does that and more by letting you test destinations (like a DSP) against each other. Powered by Adobe Sensei, Audience Lab splits segments into mutually exclusive test groups and maps them to different destinations, allowing you to find out which destination yields the best results. Audience Manager is also the only DMP that has a testing tool built into the platform, allowing you to move from insights to action faster.
See what makes it work.
Create mutually exclusive test segments to compare and measure effectiveness of different destinations to see which yields a better conversion rate or match rate for your segment.
With Audience Lab, you can text multiple creative executions going to the same destination and determine the most effective design or message.
Data providers and models
You can also use Audience Lab to test data providers and models. Using different data sets, you can create multiple look-alike models and find out which one delivers the best results.
Audience Lab also tests different ad products for publishers as well as different types of ads, like video versus display banner to determine what generates better CPM.
Learn more about A/B test segmentation in Adobe Audience Manager.
Learn from others’ mistakes.
Read our Adobe blog to find out two common A/B testing mistakes and how to avoid them.
Uncover the details of A/B tests.
Get started setting up A/B tests and see which deliver the best results in our Help section.