Usability testing

The usability test checks whether a person can perform a specific task in the interface, rather than whether they like the design. Give a realistic context without prompting a solution, observe actions and words separately, and evaluate the severity of the problem by frequency, impact, and remediation.

Why you need usability testing in Discovery

What is usability testing

Usability testing is a live observation of how a real user solves problems in a product or prototype. It is performed with minimal scenarios to see natural difficulties, errors, misunderstanding of the interface.

In Discovery, it’s a quick way:

• Find the user’s real barriers;
• Make sure the new idea or prototype is understood and works live
• Justify that it is necessary to refine before the release of the development in full implementation.

Example: testing a new prototype of the checkout page. Give the task: place an order in 2 minutes. The user is confused with the address and cannot proceed to payment. This signal is too early to transmit to the development.

When to run tests in Discovery

Usability testing is particularly useful:

• When the idea first appeared, but it is unclear whether people will understand it.
• If the team is arguing about the solution;
• After the basic drawing of the prototype, before investing in the development.

How to organize a usability test in Discovery

Minimum set: scenario, tasks, metrics

Discovery doesn’t need a complicated protocol. Enough:

• Formulate 1-3 scenarios from the user’s life.
• Prepare short introductions: what does the user do, what is the purpose.
• Think about how to measure success: did you complete the task, where you stuttered, how much time you spent.

Example scenario: a new prototype landing for registration. The user must find and fill out the form. The metric is the number of questions where the user clearly does not understand what to do.

Recruitment and number of tests

Discovery doesn’t chase a large sample. It often takes 5-7 respondents to find 80% of the major problems. The main thing is that these are relevant users from your target audience.

Example: B2B is testing new navigation. We took 6 managers from different businesses and gave them a “find a report” scenario. All participants did not find the right section on the first attempt. The team revised the menu logic before release.

How to analyze and apply results

What to Look for in Results

The most important thing is to find not one-sided impressions, but stable patterns of difficulties:

• Places where all or almost all users are lost, misunderstood, confused
• Extra clicks and steps that no one can explain
• Misunderstood wording, unrecognizable buttons, strange float.

Example: Half of the participants ask how to cancel the action. This means that the interface is not transparent enough or there is not enough clear cancel button.

How to document conclusions

The recommendation is to fix not just bugs, but usability problems: invisible CTA, non-obvious step order, complex texts. The format is short, with quotations from respondents.

It is a mistake to document only subjective reviews (such as “comfortable/inconvenient”). It is important to link the conclusion to a specific behavior.

Major Errors and Anti-Patterns

What not to do

• Test only on yourself or close colleagues. The developer’s view is always biased.
• Seek confirmation of your hypothesis, ignoring non-obvious difficulties.
• Giving lengthy explanations at first is a way to grease the user’s real reaction.

Example: The marketing team asked users to take a test after detailing the product’s capabilities. 90% of the participants showed a good result. Intuitively, the interface turned out to be more complicated than it seemed on the test: the number of complaints on feature increased on the sale.

How to Use Results to Decide on Next Steps

The usability test helps to prioritize what to change urgently and what to postpone. If most people get entangled in one section, they fix it first. If there are many problems, 2-3 main ones are identified, tested again after the update.

The usual approach: after editing, repeat the quick usability test with a new set of scenarios, make sure the problems are gone.

Sources and supplementary materials

• NNgroup: Usability Testing 101
• Atlassian Guide: Usability testing

FAQ

What is the number of respondents for Discovery?

More often than not, 5-7 real users are already finding the main problems. If no new errors appear after the changes, you can proceed to the next stage.

How to distinguish an interface bug from a usability problem?

The bug is a technical error. The usability problem is if the action works, but the user cannot understand how to use it or reach it.

Do I need to record a usability test video?

Preferably. The video helps to notice the details of the behavior and argue conclusions for the team.

How to analyze conflicting reviews?

Look not at words, but at real patterns of error. If most people do the same thing wrong, it is a clear signal to correct.

How many scripts do you need to test at a time?

Usually 2-3 key scenarios are enough to find the main blockers in the first stage.

Do I need to do usability tests on every sprint?

No, Discovery takes 1-2 rounds to test a new idea or make big changes.