Recommendations cut to 5 in a carousel: +64% sales conversions
Test run byConversion FanaticsCRO agency
Commentary byLiftArchiveOur analysis, not theirsCutting recommended products from 9-12 large images to five text-only items in a carousel lifted sales conversions by 64%.
Conversion Fanatics ran this test for an ecommerce store with thousands of SKUs.
Revenue per visitor rose 151% and average order value went from $246 to $359.
- Sales conversions: up 64%.
- Revenue per visitor: up 151%.
- Average order value: $246 to $359. Shoppers bought more per order even with fewer suggestions.
- No sample size or significance was shared.
The change
The variant shrinks recommendations to five text-only items in a carousel.
Simplified mockup drawn by LiftArchive from the shared screenshots.
Control
Each product page showed nine to twelve recommended products with large images. The section cluttered the page, especially on mobile.
Variant
Recommendations were cut to five, shown without images in a carousel.
Why it likely won
Our reading of the result, not the practitioner's claim.
- Fewer, smaller recommendations kept attention on the main product instead of pulling shoppers into a wall of options.
- The page became easier to scan on mobile, where the old section was hardest to use.
- With fewer choices, shoppers may have found it easier to pick an add-on, which fits the rise in order value.
Caveats: No sample size, duration or significance was given. Removing images and cutting the count changed at the same time, so the result can't be split between them.
Who ran it
Conversion Fanatics ran this test for an ecommerce store with thousands of SKUs and published the results (Conversion Fanatics blog case study). The numbers are theirs; the summary and commentary on this page are ours.
Read the original case study on Conversion Fanatics →
How to run this test on your store
- Hypothesis
- If we cut our recommended products to a small carousel, then more shoppers will buy because the main product stays in focus and the page is easier to use.
- Primary metric
- Sales conversion rate
- Guardrails
- Average order value
- Clicks on recommended products
- Revenue per visitor
- Who should test it
- Stores with large catalogs whose product pages show long or image-heavy recommendation blocks. It probably won't help if your recommendations are already short and tidy.
Questions
How many recommended products should a product page show?
There is no fixed number, but fewer can work better.
Here, five items beat nine to twelve by a wide margin.
Do product recommendations need images?
Not always. This test removed images and still raised order value, though image removal was not tested on its own.
Will fewer recommendations lower order value?
It did not here: order value rose from $246 to $359. Track it as a guardrail in your own test.
Similar tests
@oliverkenyon@oliverkenyon@oliverkenyon@oliverkenyon@oliverkenyon@oliverkenyon