Five-second test: what is it and how do you run it?

What is a five-second test

In a five-second test, a person sees a page for only 5 seconds. Then the page closes, and they are asked what they remember. What sticks is what the page actually conveys at first glance.

Christine Perfetti coined the term in the mid-2000s. Eduard Kuric writes this in his 2023 case study in Smashing Magazine. According to Kuric, Perfetti’s article, published at UIE on September 11, 2007, can count as the first source of the modern five-second test.

According to Perfetti, the point is to see whether a content page conveys its main purpose clearly and concisely. Users make their real judgement about a page in the first moments, and the test captures exactly those moments.

The method later became the topic of a book too. Paul Doncaster’s 2014 book “The UX Five-Second Rules” uses examples from more than 300 tests, according to the publisher’s page at Elsevier.

Live five-second test

Below are two landing pages for the same made-up product. Version A’s heading is vague; version B says clearly what it offers. Choose a version, press Start and answer the question as soon as the page disappears.

If five seconds is too short for you, press “Remove time limit”. The page stays visible until you are ready.

I built that button in because of WCAG 2.2 success criterion 2.2.1. Under this level A criterion, users must be able to turn off, adjust or extend a time limit.

Live five-second test

Both versions side by side

menu.example

Example MenuSolutionsAbout usContact

Taste, reimagined digitally

With a next-generation approach, we take your business into the future.

Learn more

The heading describes a feeling, not the product. The subtitle fits any company. After five seconds, probably only “something to do with food” sticks.

menu.example

Example MenuHow it worksPricingSign in

QR menus for your café or restaurant

Upload your menu once and put the code on the table. Change a price and guests see it right away.

Try it free for 14 days

The heading says what the product is and for whom. The subtitle explains in two sentences how it works. The button names the next step.

Because JavaScript is off, the test doesn’t run; above you see both versions side by side. In the live test, the page appears for 5 seconds, is then hidden, and the question “What does this website offer?” follows.

The Example Menu brand, both landing pages and the answer options are made up by me; it is not a real company. In a real test, each person sees only one version. Here you can try both, but on the second try you already know the answer.

How a five-second test works

The process in Perfetti’s article is short and repeatable. I build the test in this order.

1. Give a focused task. The article tests the Red Cross donation page. The task reads: “You’re ready to donate to the Red Cross.”
2. Announce the time in advance. Tell the person they will see the page for only 5 seconds. Ask them to remember as much as they can.
3. Show and hide. The page stays on the screen for 5 seconds and then closes.
4. Have them write down what they remember. The person writes down everything that stayed in their memory. Then ask two follow-up questions. In the article they are “What is the most important information on this page?” and “How would you donate to the Red Cross?”.

According to the article, the test takes about 10 minutes per person. That is why it can be run with many people in a short time at trade shows, conferences or in the company cafeteria.

The test works with paper sketches or PDFs too. So you can use it before the design is even on a screen.

The result of the Red Cross example is instructive. With the first design, everyone understood that the page was there to collect donations, but participants only knew about donating money.

You could also donate stock, clothes and airline miles, and that was on the page. Yet nobody noticed. Those who saw the new design with highlighted options remembered all the options after 5 seconds.

What the five-second test measures and what it doesn’t

The test measures what the page conveys at first glance. According to Kuric, it should show whether the person’s first impression is the one you want. Does the main message get through, do people remember the company name, how complex does the page seem? The test answers these.

What it doesn’t measure, Lyssna’s guide lists like this:

• It shows what is noticed, but not why.
• It gives no behavioural data such as task success or error rate.
• The page is shown without context. The person doesn’t know which ad or search brought them there.
• The person may not remember everything they noticed.

Then there is a limit on page type. Perfetti writes that the test works best for pages with a single main purpose.

Kuric reports that guides give the same warning for interfaces with several purposes. According to him, though, dropping the method altogether for such interfaces means giving up the chance to measure the first impression.

Whether someone can complete a task from start to finish isn’t something this test tells you. For that you need a usability test.

Why 5 seconds?

According to Perfetti, with more time participants start studying the page like designers. They notice details they would otherwise miss. That distorts the first impression.

But the 5 isn’t a law. Kuric writes that this duration rests mainly on researchers’ observations. In his team’s case study, participants saw the pages for 2, 5 or 10 seconds.

With more time, fewer people gave a wrong purpose for the website. The difference was clearest between 5 and 10 seconds. Attitudinal answers, on the other hand, were largely unaffected by the duration.

Kuric recommends choosing the duration by the type of question and the page’s visual complexity.

A good question for a five-second test

A good question asks about exactly one thing and doesn’t suggest the answer. Lyssna’s guide gives two examples. “What elements did you notice?” is neutral. “Did you notice our eye-catching call-to-action button?” steers the answer.

Kuric distinguishes three kinds of questions:

• Attitudinal question: have the page rated on a scale from 1 to 7, for example from ugly to attractive.
• Target identification question: ask about a specific element, such as the company name in the logo.
• Memory recall: ask the person to tell everything they remember.

Keep the number of questions small. Lyssna recommends 5–8 questions at most, because participants get tired.

Split compound questions. Lyssna’s example: “What was this page about, and how did the visual hierarchy draw your attention to the value proposition?” is two questions.

My first question is always the same: “What does this website offer?” When the person answers in their own words, you see better what the page really conveys. In the demo above I used a multiple-choice question because I wanted to show the result right away.

How many participants you need

Perfetti’s article gives no number. Lyssna’s guide writes that five to ten participants are usually enough to see the main patterns. For a specific target audience it recommends 5–6, and for a general first impression test 8–10 people.

Because one person finishes the test in about 10 minutes, increasing the number is cheap. If you compare two versions such as A and B, give each version its own group. If the same person sees the second version, they already know the answer.

How I think about sample size in general is in the guide how many users to test.

Using it for landing pages and home pages

The five-second test works best for a landing page. A landing page usually has exactly one job: tell visitors what it offers and lead them to a step. That fits exactly the single-purpose page Perfetti describes.

The time pressure is real too. According to Jakob Nielsen’s 2011 NN/g article, the likelihood of leaving a page is very high in the first seconds. The first 10 seconds are critical for the decision to stay or leave. Nielsen’s advice: communicate the value proposition within 10 seconds.

When testing a landing page, I look for answers to these questions:

• Can the person say what is offered and for whom?
• Do they remember the main button?
• Does the product’s benefit appear in the answers, or only its name?

How I build the rest of the page is described in the landing page pattern.

Be careful with the home page. Perfetti writes that home pages and large navigation pages give less useful results because they serve many tasks and user types. Lyssna, on the other hand, considers the home page a good candidate as the first point of contact.

With the right question, both are right in my view. On the home page, “What does this website offer?” works. “Did you understand everything on this page?” doesn’t.

How to read the results

First group the answers. Put together those who say the same thing in other words, and count who names the page’s purpose correctly.

Then look at the words. If words from the heading appear in the answers, the heading is doing its job. If nobody mentions the main button, it is either not visible or doesn’t say what it does.

Change something and repeat the test with a new group. In Perfetti’s Red Cross example too, only the second round gave the real result. Those who saw the new design remembered donation options that nobody had noticed with the first design.

My note

A feeling sentence like the one in version A can seem perfectly adequate to the team building the product. Because the team knows the product well, it doesn’t notice what the heading leaves out. For me, the five-second test is one of the cheapest ways to expose this blindness with an outside view.

I like to attach the test to the first minute of a usability session. I get the first impression before the person uses the product, then move on to the real tasks. That way one session gives two kinds of data.

If you aren’t sure what your page conveys in the first 5 seconds, I can look at your landing page and one flow from start to finish in the UX audit. I look at the heading, value proposition and main button first.