Guerrilla research brings a small research activity to people who can take part there and then. A team might approach visitors in a library to try a catalogue search or ask attendees at a relevant event to react to a prototype. The convenience of the setting makes the work possible, but also shapes whose experience it captures.

The approach belongs to a broader tradition of lightweight usability work, discussed in Jakob Nielsen’s Guerrilla HCI article. In practice, the label covers several informal arrangements rather than a single fixed protocol.

Make the setting relevant to the question

A public location is useful only if the people there plausibly match the study’s purpose. Library visitors may help investigate a general catalogue task; they cannot stand in for specialist archivists simply because both use search tools.

Ask a few eligibility questions before the task and record the characteristics that matter. This is still recruitment, even when it takes place without advance appointments. The resulting convenience sample should not be described as random or representative.

Keep the activity short enough for the situation, and consider noise, privacy, device familiarity, and interruptions. A quick test in a busy lobby may reveal a confusing label while offering little opportunity to explore a sensitive or complicated experience.

Give people a real choice about taking part

Explain the activity, duration, recording arrangements, and any incentive before beginning. Permission from a venue does not replace the participant’s informed consent, and a friendly approach should leave an easy route to decline.

Use a realistic task rather than asking whether the design looks good. In the illustrative library example, finding a book available to borrow today tests an actual goal. Telling someone to open the availability filter would remove an important part of the question.

Avoid rescuing people too quickly. Their attempts and expectations are useful evidence, provided the session remains comfortable and they understand that the product is being examined.

Use the findings at their proper scale

Record who took part, where, what they attempted, and what happened. A short round can identify a problem worth fixing or a question worth investigating further. It cannot establish how common that problem is across the intended audience.

Follow up with more suitable recruitment or moderated testing when the decision needs deeper context. The speed of guerrilla research is valuable when it helps a team encounter evidence early, with enough discipline to keep an accessible group of participants from becoming an imagined picture of all users.