Test archive / Collection page
Do star ratings on collection listings lift conversion?
Our collection pages showed each product with an image, name and price. We tested adding star ratings under the first 16 products. Over 15 days, conversion fell 15.6% against the control, and the result did not reach our confidence bar.
Not significant Responsive 15 days Not called
The hypothesis
By adding star ratings to reviewed products on collection pages, we will build trust in each listing, resulting in more product clicks and purchases.
Shoppers on a collection page judge each product from a small card: image, name and price. Many products already had reviews, but the ratings only showed on the product page.
Star ratings are a familiar trust signal, so we expected that showing them earlier would help shoppers pick products to explore. A testing tool limit meant the stars could only appear on the first 16 products in each tested collection.
What we changed
We added star ratings under the first 16 products on collections.
In the variation, products with reviews showed a row of stars on their card on the tested collection pages, on mobile and desktop. Because of the tool limit, only the first 16 products in each collection got stars, so cards further down the grid looked the same as before. Everything else on the page stayed the same.
- Star rating row added to each product card
- Shown only on products with reviews
- Limited to the first 16 products per collection
The result
Conversion fell 15.6% with star ratings, short of our bar.
Revenue per visitor fell 27% with the stars in place. In the desktop data alone, conversion was down 6.1% and revenue per visitor was down 42.9%. The numbers swung a lot from week to week, and none of the drops reached our confidence bar, so we paused the test.
We held the test while a site navigation change went live and restarted it afterward, so the 2 changes would not blur each other. The figures leave out orders over $2,000, which mostly come from bulk buyers and can skew results.
What we would do next
Extra lines on each card may have crowded the product grid.
With stars added, each product card carried 4 lines of information, which may have made the grid harder to scan. Showing stars on only the first 16 products may also have pulled attention away from the rest of the collection. A cleaner rerun would show ratings on every reviewed product without adding height to each card.