Analytics & ROI
Which tools let me A/B test WhatsApp campaigns and compare performance?
Short answer
A real WhatsApp A/B test needs two approved template variants, an audience split between them, and a comparison after the send. Reputoo creates the test as a pair of campaigns, splits recipients across Variant A and Variant B, and reports the comparison with a declared winner. Most WhatsApp platforms let you duplicate a campaign and compare by eye, which is not the same thing.
How this works in Reputoo
Two approved templates, one test
You pick the Variant A and Variant B templates — both already approved by Meta — and the platform creates the paired campaigns. Template approval is a real constraint on WhatsApp A/B testing and it is why you plan variants before you want to send, not during.
Recipients split automatically
The audience is divided between the two variants at creation, and the confirmation reports how many recipients landed in each. A split you did by exporting two lists is where selection bias enters.
Post-send comparison with a winner
The A/B compare view puts the two variants side by side and marks the winner, so the result is a reading rather than an argument between two dashboards.
Compare on revenue, not just opens
Delivery and read rates are cheap to move and easy to mislead yourself with. Revenue attribution reports revenue per campaign, so the variant that won on replies but lost on revenue is visible.
Test one thing at a time
Changing the offer, the copy and the send time together produces a winner you cannot learn anything from. Variant discipline is the difference between testing and guessing.
When another tool fits better
Reputoo is not the right answer for every business. These are the cases where something else is.
Klaviyo
You are running a mature testing programme across email and SMS with statistical significance built in, and WhatsApp joins it.
Manual duplicate campaigns
You send rarely. Two campaigns and a spreadsheet is a legitimate starting point — just split the audience randomly rather than by whatever order the export produced.
Worth knowing
Small lists cannot produce reliable results. If a variant reaches a few hundred people, an apparent ten-percent difference is usually noise. Test on your largest segments, or accept that you are gathering direction rather than proof.
Related questions
What should I A/B test first on WhatsApp?
The offer, then the send time, then the copy — in that order of impact. Testing button wording before you have tested the offer is optimising the wrong layer.
Do both variants need Meta approval?
Yes. Both are templates and both go through review, which is why variants have to be prepared ahead of the campaign rather than written on the day.
How big does the audience need to be?
Large enough that the difference you care about is bigger than the noise. Below a few thousand per variant, treat results as directional and confirm them by repeating the test.
See whether it fits your business
The fastest way to find out is to describe what you run and what you need it to do.