Automated product bundles: +11% basket value at La Bécanerie
Commentary byLiftArchiveOur analysis, not theirsAdding automatically generated product bundles went with an 11% rise in average basket value at La Bécanerie, a French motorcycle parts and gear retailer.
AB Tasty set up the bundles with its own tool, and the result was published as a VWO case study.
This was a product rollout measured before and after, not a split test.
- Average basket value: +11%. Reported as a before-and-after change, not against a control group.
- Bundle engagement: 7% of visitors who saw a bundle used it.
- Bundle users bought about three times more than non-users. This is a correlation: bigger buyers may simply be more likely to use bundles.
The change
An automated bundle of items often bought together is added to the product page.
Simplified mockup drawn by LiftArchive from the shared screenshots.
Control
Before the change, the site showed no automated bundles. Shoppers had to find related parts and gear themselves.
Variant
Bundles were built automatically from what customers tend to buy together. They were styled to match the existing site design.
Why it likely won
Our reading of the result, not the practitioner's claim.
- Motorcycle buyers often need several matching items, such as parts plus gear. Bundles save them the search.
- Bundles based on real purchase patterns are more likely to be relevant than hand-picked ones across a large catalog.
- Matching the site's design probably made the bundles feel like part of the store, not an ad.
Caveats: This is a before-and-after result from a vendor's product rollout, not an A/B test, so seasonality or other changes could explain part of the lift. No sample size or duration was shared.
Who ran it
AB Tasty ran this test for La Bécanerie and published the results (VWO case study). The numbers are theirs; the summary and commentary on this page are ours.
Read the original case study on AB Tasty →
How to run this test on your store
- Hypothesis
- If we show bundles of items often bought together on product pages, then average order value will rise, because shoppers who need several matching items can add them in one step.
- Primary metric
- Average order value
- Guardrails
- Conversion rate
- Revenue per visitor
- Return rate
- Who should test it
- Stores with large catalogs where items are often bought together, such as parts, gear or beauty routines. Stores that sell one standalone product per order will see less from this.
Questions
Do product bundles increase average order value?
They often can. La Bécanerie reported an 11% rise in basket value after adding automated bundles, though that was a before-and-after measure, not a split test.
Should bundles be built by hand or automatically?
For a large catalog, building them from purchase data scales better and keeps them relevant.
Hand-picked bundles can work well for a small range.
How do I test bundles properly?
Show bundles to a random half of visitors and compare revenue per visitor with the other half.
Comparing bundle users with non-users mixes up cause and effect.
Similar tests
@oliverkenyon@oliverkenyon@oliverkenyon@oliverkenyon
@DaveDiederen@DaveDiederen