Blue form background lost: 0.69% to 0.54% orders
Test run byMarketingExperimentsResearch publisher
Commentary byLiftArchiveOur analysis, not theirsGiving the subscription form a blue background cut conversion from 0.69% to 0.54% (42 vs 33 orders), about 22% fewer orders per view.
MarketingExperiments (MECLABS) ran this as part of a multivariate test for a subscription site that lets people search registered sex offenders.
The results are early and the order counts are small.
- Conversion: 0.69% to 0.54% (42 vs 33 orders), about -22% relative. A negative result.
- Each form got roughly 6,100 page views over eight days.
- A blue-button form did worse still (0.39%), and the standard form beat this one by 28%.
- No significance was reported, and these were early numbers from a test still running.
The change
Only the form background changed, to blue.
Simplified mockup drawn by LiftArchive from the shared screenshots.
Control
The standard subscription form, with the site's usual button and plain form styling.
Variant
The same form, but set on a blue background.
Why it likely lost
Our reading of the result, not the practitioner's claim.
- A coloured panel can make a form look like an ad or a separate box, which may reduce trust.
- Colour on its own rarely matters much. What counts is contrast with the rest of the page, and the blue version may have had less.
- With only a few dozen orders per form, part of this gap may be chance.
Caveats: Early results after eight days of an ongoing multivariate test from 2006, with only 42 vs 33 orders. Headline and logo were varied at the same time, and the per-form numbers are averaged across those versions.
Who ran it
MarketingExperiments ran this test for a subscription site for searching registered sex offenders and published the results (MarketingExperiments). The numbers are theirs; the summary and commentary on this page are ours.
Read the original case study on MarketingExperiments →
How to run this test on your site
- Hypothesis
- If we put the subscription form on a coloured background, then more visitors will order because the form will draw the eye.
- Primary metric
- Orders per form page view
- Guardrails
- Form start rate
- Revenue per visitor
- Who should test it
- SaaS and subscription sites whose sign-up form is hard to spot or blends into the page. If your form already has a clear, high-contrast button, a colour change is unlikely to help, and you need a few hundred orders per version to trust the result.
Questions
Does button colour affect conversion?
Sometimes, but usually through contrast, not the colour itself.
A button that stands out from the page tends to do better than one that blends in.
Is there a best colour for a sign-up button?
No. Results depend on the rest of the page.
In this test the standard form beat both blue versions.
How many orders do I need before trusting a colour test?
Far more than a few dozen. Small design changes tend to have small effects, so you need hundreds of conversions per version to tell a real gap from noise.