A five-second test is a quick usability method that measures first impressions. Participants see a page, usually a homepage or landing page, for about five seconds. The page is then hidden and they answer questions such as what the company does, who it is for and what they could do next. If most people cannot answer, the page's message is not landing quickly enough.
What it measures
The test checks clarity, not persuasion. It tells you whether the headline, visual and layout communicate the core idea within a glance. It does not tell you whether people would buy, or whether the page performs well for visitors who read carefully.
How to run one
Use a screenshot or live page with no other context, and ask open questions rather than leading ones. Online research tools offer five-second tests with recruited panels, or you can run them informally with people who match your audience.
- What does this company or product do?
- Who do you think it is for?
- What was the main thing you could do on this page?
- What, if anything, do you remember about the design or images?
Using the results
Compare participants' answers with what you intended. If people describe a different product, or say “something to do with data”, rewrite the headline and supporting line and test again. Five-second tests are cheap enough to run on several headline drafts before choosing one to A/B test.
Example
Checking a new homepage headline
Hypothetical background-check API Marlspike runs five-second tests on two homepage drafts. With the first, most participants guess it is an HR tool for recruiters. With the second, which mentions developers and shows a code snippet, most correctly say it is an API for adding checks to an app.
Common mistakes
Testing with your own team
Anyone who knows the product will fill in the gaps. Use people who have never seen it.
Testing a page that depends on context
A landing page for an ad assumes the visitor saw the ad. Show participants the ad first, or the test will undersell the page.
Questions
It is a convention long enough to take in a page's headline and main visual but too short to read detail. The exact time matters less than keeping it consistent between versions.
