Your first A/B test winner is probably where you’re stopping too early. Here’s w
By Kamaljit Singh · October 22, 2026 · Curated by George's Blog
Your first A/B test winner is probably where you’re stopping too early.
Here’s what I mean.
You generate a few Amazon images, pick your favourite, and upload it.
You’re hoping shoppers click what you like.
Running a shopper poll first is a big step forward. But picking the winner and uploading it still leaves a lot of opportunity.
The real gains come from what you do AFTER that first test.
Read the feedback.
Why did people choose one image?
What confused them?
What made the product look more trustworthy?
Use those comments to improve the image. Then challenge your winner again.
That’s the loop:Generate → test → learn → improve → test again.
Keep going until you have an image that consistently wins by a clear margin against strong alternatives, including competitor images.
THEN take it to Amazon.
Use Manage Your Experiments to see whether that preference translates into purchases.
Track ad performance too, keeping other changes in mind.
Bradley Sutton walks through this process in his podcast episode (that i think is a masterclass on ab testing) using Listing Optimization AI, Helium 10 Audience polls powered by PickFu.
If you’re using AI to generate images in 2026, use it to act on shopper feedback too. That’s where this gets really interesting.
Link to watch: https://lnkd.in/gifmiJD3