Facebook 'likes' widget on product pages: no measurable change
Commentary byLiftArchiveOur analysis, not theirsAdding a Facebook widget that showed more than 1,000 likes to product pages made no measurable difference to bounce rate or orders.
FietsPunt, a Dutch online store for bike parts and cycling wear, ran this test itself, and it was published as a VWO case study.
A second placement of the widget also made no difference.
- Bounce rate: no meaningful change. The high bounce rate the team wanted to fix stayed the same.
- Orders: no meaningful change reported.
- Second placement: also no difference. No numbers were shared for either version.
The change
A Facebook likes widget added to the product page
Simplified mockup drawn by LiftArchive from the shared screenshots.
Control
Product pages had no social proof widget.
Variant
A Facebook widget showing more than 1,000 likes was added to product pages. A later variant tried the widget in a different spot.
Why it likely came out this way
Our reading of the result, not the practitioner's claim.
- Page likes say a brand is popular, not that a shop delivers. Shoppers worried about a small store may need proof of service, not fans.
- A count of around 1,000 likes may not seem large enough to reassure anyone.
- Facebook widgets are easy to ignore and can look like page decoration.
Caveats: No numbers, sample size or duration were shared, so we can only say the widget made no clear difference. Traffic was split 50/50.
Who ran it
FietsPunt ran this test for FietsPunt 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 FietsPunt →
How to run this test on your store
- Hypothesis
- If we show our Facebook likes on product pages, then fewer visitors will bounce because they will see that many people trust the shop.
- Primary metric
- Product page bounce rate
- Guardrails
- Orders per visitor
- Page load time
- Who should test it
- Small or little-known stores where trust is the main worry. If you have customer reviews, test those first; in FietsPunt's case a reviews widget did far better than likes.
Questions
Do Facebook likes on a website increase sales?
Not in this test. FietsPunt saw no change in bounce rate or orders, in either of two placements.
What kind of social proof works better than likes?
Customer reviews tend to say more about the buying experience.
FietsPunt later tested a Trustpilot reviews widget and saw a 36.73% rise in orders, though the tests ran months apart.
Does the position of a social widget matter?
It can, but here moving the Facebook widget to a second spot made no difference either.
Similar tests
@oliverkenyon
@DaveDiederen@DaveDiederen
@CROguy_@DaveDiederen@DaveDiederen