Rekko offers two types of split-testing options: Self-Optimizing and Manual A/B Testing.
Self-Optimzing is your “set it and forget it” option. Create a campaign with multiple creatives and let Rekko find which ones works best. The creative variant that performs best will be delivered to the majority of your visitors.
How does this work? Rekko will take approximately 30% of your visitors at random and display one of the variants. The variant that performs best of that test group will be displayed to the rest of your visitors in the non test group.
The great thing is that this will account for any trends you might see over time. That is, if Variant A does well for awhile, but we start to see Variant B come up on the rise, then Variant B may be the creative that starts to get delivered to the non test group.
Manual A/B Testing allows you to input the percentages yourself. If you’re more hands on, you can make sure to set your splits the way you want. Use Rekko’s schedule feature to have these campaigns only run for a period of time, or just monitor them as you need.
Why isn’t the delivered variant for Self-Optimizing consistent from session to session?
The nature of the Self-Optimizing feature is to find the winningest variant and deliver it to your visitors. As we’ve mentioned above, Self-Optimizing is always testing to find the best variant – as it gathers results, the variant that’s delivered more often may change from time to time. This is especially true for new campaigns as it starts to learn quickly what the visitors prefer.