LovelySkin redesign checked by 10 pre-launch tests: +39% conversion
Test run bySurefootCRO agency
Commentary byLiftArchiveOur analysis, not theirsLovelySkin's site redesign was followed by a 39% rise in conversion rate and a 20% rise in average order value after launch.
Before going live, Surefoot ran three rounds of user interviews and ten A/B tests on the riskiest parts of the new design.
This is a before-and-after comparison of a full redesign, not a split test of new site versus old.
- Conversion rate: up 39% after launch, compared with before.
- Average order value: up 20% after launch.
- The case study reports no performance dip after launch.
- Results of the ten individual pre-launch tests were not reported.
Control
LovelySkin's previous website.
Variant
A redesign by LovelySkin's in-house team, with new functionality and an updated brand look. It was refined through three rounds of 1:1 interviews on prototypes and ten A/B tests on its highest-risk pieces.
Why it likely won
Our reading of the result, not the practitioner's claim.
- Testing the riskiest pieces first meant changes that would have hurt sales could be caught before launch.
- Interviews on prototypes surfaced confusion the in-house designers had stopped noticing.
- Because it is before-and-after data, some of the lift could come from season, traffic mix or promotions. The case study does not say.
Caveats: Before-and-after comparison, not a split test: other factors could have affected the numbers. The ten pre-launch A/B tests are not reported individually, and sample size and time window were not shared.
Who ran it
Surefoot ran this test for LovelySkin and published the results (Surefoot case study). The numbers are theirs; the summary and commentary on this page are ours.
Read the original case study on Surefoot →
How to run this test on your store
- Hypothesis
- If we test the highest-risk parts of our redesign with A/B tests and user interviews before launch, then the new site will match or beat the old one because we fix problems before all traffic sees them.
- Primary metric
- Site-wide conversion rate
- Guardrails
- Average order value
- Revenue per visitor
- Checkout completion rate
- Who should test it
- Established stores planning a major redesign with real revenue at stake. Very small stores may not have the traffic to run ten tests before launch.
Questions
Should you A/B test a whole site redesign?
A full new-versus-old test is hard to run cleanly.
A common alternative, used here, is to test the riskiest individual changes before launch.
How do you avoid a conversion drop after a redesign?
Combine user interviews on prototypes with targeted A/B tests on the biggest changes.
That way problems show up before the whole audience sees them.
Is a before-and-after comparison reliable?
It is weaker than a split test, because season, promotions and traffic mix can shift results.
Treat it as a strong signal, not proof.
Similar tests
@oliverkenyon@oliverkenyon@oliverkenyon
@DaveDiederen@DaveDiederen
Blend Commerce