What is A/B Testing?
A/B testing (also called split testing) means showing two or more different versions of a page to different visitors and measuring which version performs better. For example, you might test two different headlines, call-to-action buttons, or page layouts to see which one drives more conversions.
Psyke AI Variants
When Psyke generates pages for a campaign, it creates multiple AI variants — different versions of a page targeting the same keyword. These variants differ in layout, copy, or structure. You can use these variants to run split tests.
Setting Up a Split Test
In your campaign, navigate to the page you want to test.
Click Variants to see the different versions Psyke has generated.
Enable the variants you want to test (typically 2–3 at a time).
Set the Traffic Split — for example, 50% to Variant A and 50% to Variant B.
Click Publish to go live.
How Traffic is Split
Psyke distributes visitors between your active variants randomly but evenly, based on the percentages you set. Each visitor always sees the same variant on repeat visits (this is controlled by a cookie), so results are consistent and not skewed by the same person seeing different versions.
Measuring Results
Check your analytics after at least 1–2 weeks (or once each variant has received at least a few hundred visitors) to get statistically meaningful data. Look for differences in:
Conversion rate (form submissions, button clicks)
Time on page
Bounce rate (people who leave immediately)
Declaring a Winner and Iterating
Once a variant is clearly performing better, pause the others and set the winner to receive 100% of traffic.
Note what made the winner successful — was it the headline? The CTA? The image?
Use that insight to inform your next round of tests.
Tip: Change only one element at a time so you can clearly attribute the difference in performance to a single change.