
OVERVIEW
THE PROBLEM
Recommendations had to drive revenue in any segment without competing with the purchase. Done carelessly, they become friction. The opportunity: the product page.
WHAT I DID
End to end: benchmark, two concepts, a 14-person usability test and the final pattern.
IMPACT
of test participants reached the cart with 2+ recommended products using the inline version
My process
01. Research
I benchmarked how e-commerce apps show recommendations, then mapped our certainties, suppositions and doubts in a CSD Matrix.

02. Ideate
concepts moved to testing
A modal, a section on the product page, a banner in the cart: I wireframed several directions and turned the strongest into flows.



03. Test
14 participants, an unmoderated test in Maze, one task: add a t-shirt to the cart, then two suggested products. I measured format, strategy and add-to-cart feedback.
Version A: pop-up carousel
2 completed · 12 didn’t complete the task
Heatmaps showed hesitation: people couldn’t tell if they were adding a product, leaving the page or interrupting checkout.

“The pop-up gives me the impression of something I can skip without looking, because it might just be an ad.”
Version B: inline recommendations
14 completed with no confusion · 1 couldn’t move forward

“The white background looks cleaner and more organized. It makes me feel safer about the purchase.”
04. Iterate
Everyone preferred B: more familiar, organized and trustworthy, and everyone reached the cart with 2+ recommended products. I refined it with clearer hierarchy, product states and spacing, and kept the confirmation people liked most.

Up to 5 products, no SKU variations
Why: a short carousel doesn’t overwhelm, and products go straight to the cart. Trade-off: fewer suggestions, and products with variants stay out.
Lower-priced complements first
Why: easy, quick decisions on top of a purchase in progress. Trade-off: a smaller add-on per item, in exchange for more people adding one.
Only design system components
Why: less development effort, and the pattern scales across every client app. Trade-off: less room for one-off visuals per brand.
In constant
motion
© 2026