The first move on a page reveals what the page appears to promise. Someone looking for a returns policy may open Help, scroll to the footer or select an order from their account. Each choice expresses an expectation about where the answer belongs, even before the person sees whether that expectation is right.
First-click testing records where participants first click or tap when given a task. Researchers can use a static image, prototype or working page, depending on the study. The result helps evaluate the starting point of a journey: whether people recognise a useful route, which alternatives attract them and where further investigation is needed.
Define the task and the acceptable starting points
Imagine a fictional museum website redesigning its visitor information. The team wants to know where someone would begin if they needed to leave a large bag during a visit. Asking them to “find the cloakroom” may supply language used on the website. Describing the situation lets the team investigate whether its labels connect with the visitor’s need.
Decide beforehand which starting points could genuinely lead to an answer. Help, visitor information and a search box might all be defensible. Counting only the designer’s preferred route as correct would make the scoring rule narrower than the actual service.
The first-click method is one of the approaches described in Nielsen Norman Group’s guide to testing visual design. It is useful where a visitor has a goal and the design needs to make an initial action understandable.
Choose a stimulus that contains the relevant clues
A screenshot is useful before development, but it changes the available experience. It may omit hover menus, scrolling, responsive behaviour and feedback that would help someone on the working site. A screenshot of an entire long page also presents information differently from a normal viewport.
Match the stimulus to the question and record its limitations. If the navigation depends on opening a menu, an interactive version may be necessary. If the question is whether a visible label suggests the right destination, a static design may be sufficient. Ballpark’s first-click test provides a way to collect starting choices on a design.
Keep exposure consistent when comparing versions. A participant who has already explored one layout brings knowledge to the next. Separate groups, randomised assignment or carefully balanced presentation can help, but they answer different study-design needs; choose the arrangement before collecting responses.
Read the distribution, not just the winning area
A click map can show a strong concentration on one element, several plausible routes or scattered uncertainty. Review these patterns alongside the task and any follow-up answers. An open question about what participants expected at their chosen destination can help distinguish a sensible alternative from a misleading label.
Timing may add context, but a slow click is not automatically confusion. Someone may read carefully, be interrupted or use an input method that takes longer. Report how timing was captured and investigate unusual values before using speed to rank designs.
When reporting correct-click percentages, include the number of participants and appropriate confidence intervals. A small numerical lead can be too uncertain to justify a redesign. The practical value also depends on what the alternative route costs: a recoverable detour differs from entering an irreversible process.
Follow the journey when the decision requires it
A correct first click does not establish successful completion. The museum visitor may open the right page and still fail to discover the bag-size restriction. Equally, someone who starts elsewhere may recover quickly through a helpful link.
Use usability testing to examine the complete attempt. If the uncertainty is specifically about the organisation of a hierarchy, tree testing can help investigate it without the visual layout. First-click testing is most useful when its narrow result leads to a proportionate next question, rather than standing in for the whole experience.