Before you launch: test with a few people, then make it simpler

The launch trap

When a new website, app or online service is nearly ready, there is a strong pull to keep adding and polishing. One more feature, one more page, one more round of colour changes. All of it feels productive. Very little of it tells you whether people can actually use what you have built.

Two simple habits make a bigger difference than almost anything else: test with a small number of real people, and make every important journey shorter.

You do not need a big study

Usability testing sounds expensive, but it does not have to be. In his well known article Why you only need to test with 5 users, Jakob Nielsen of Nielsen Norman Group explains that a study with five participants found about 85% of the usability problems in a design. [S19] His advice is to run several small tests rather than one large one, fixing what you find in between.

The article was published in 2000 and is based on his own modelling, so treat 85% as a rule of thumb rather than a promise. The principle has held up well: a handful of people, watched carefully, will show you most of the big problems.

A simple test looks like this:

1. Pick the three or four tasks that matter most. For example, find a price, book a call, sign up or pay.
2. Find five people who are like your real users. Not your team, and ideally not your friends.
3. Ask them to try each task while thinking out loud. Do not help, and do not explain the screen.
4. Write down where they hesitate, get stuck or give up.
5. Fix the worst problems, then test again with a new group.

Every extra step costs you

The second habit is to look hard at your forms and checkouts. Baymard Institute, which researches online shopping usability, keeps a list of cart abandonment rate statistics. Across the 50 studies it lists, the average documented online shopping cart abandonment rate is 70.22%. [S20]

Some of that is people simply browsing. Baymard's own survey of US online shoppers found that 42% had abandoned a cart because they were just browsing or not ready to buy. [S20] Looking at the other reasons, 18% said the site wanted them to create an account, and 17% said the checkout process was too long or complicated. [S20]

Baymard's checkout research also suggests that an ideal checkout flow can be as short as 12 to 14 form elements, while the average US checkout it has benchmarked shows 23.48 form elements by default. [S20] These are figures from US shopping sites, but the lesson applies to any form: every field you ask for is a chance for someone to give up.

Making journeys shorter

A few questions help with any form, booking flow or sign up:

• Do we really need this field now? If you can ask for it later, or not at all, remove it.
• Can people continue without an account? Offer to save details at the end instead.
• Is the total cost clear early? Surprises at the last step lose people.
• Does it work well on a phone? Large tap targets, the right keyboard for each field and clear error messages.
• Is the language plain? Labels and buttons should say exactly what happens next.

Launch smaller, learn faster

Launching does not have to mean launching everything. A simpler first version that people can actually use teaches you far more than a bigger one that confuses them. You can always add features once you know what your users need.

If you are getting ready to launch a website, app or digital product, our New product design service includes testing with real users, and our Bespoke web apps and iOS apps team builds simple, accessible journeys from the start. If you are planning how to take it to market, see our Go to market strategy service, or book a discovery call.

Sources

1. Jakob Nielsen, Nielsen Norman Group, Why you only need to test with 5 users, 18 March 2000. [S19]
2. Baymard Institute, 50 cart abandonment rate statistics 2026, last updated 22 September 2025. [S20]