Guerrilla Usability Testing
Steve Krug's low-budget, do-it-yourself approach to usability testing: a few users, once a month, on whatever you've got — valuing frequency and simplicity over scientific rigor.
Also known as: Guerilla Usability Testing, DIY Usability Testing, Rocket Surgery Usability Testing
Category: Techniques
Tags: usability, user-experience, evaluation, research, techniques, testing, user-centered
Explanation
Guerrilla usability testing is a scrappy, lightweight way to find usability problems without the cost and ceremony of formal lab studies. Popularized by Steve Krug in *Rocket Surgery Made Easy*, it rests on the observation that almost any real testing beats no testing — and that teams that test frequently improve their products far faster than teams that wait for a perfect research study.
**Krug's Recipe**:
- **Test three users, once a month.** Three people will reveal most of the serious problems; more testing that day produces diminishing returns.
- **Test whatever you have.** Sketches, wireframes, prototypes, the live site, or a competitor's site. Don't wait for the design to be "ready".
- **Do it in the morning, debrief over lunch, decide fixes the same afternoon.** Momentum matters more than precision.
- **Recruit loosely.** Friends, colleagues from other teams, people from the street, or paid participants — usability problems are usually obvious enough that representative sampling is overkill.
- **Fix the most serious problems first.** The goal is not a tidy bug list but shipping a better product next sprint.
**Why a Few Users Is Enough**:
Jakob Nielsen's research showed that testing with just 5 users typically reveals around 85% of usability issues. Krug argues three is often sufficient because you're testing frequently — any issue missed in one round tends to surface in the next. The value compounds over months of small tests, not from one heroic large study.
**Running a Session**:
1. **Welcome (2 min)**: Explain it's the site being tested, not the user; encourage thinking aloud.
2. **Questions (2 min)**: Gather brief background.
3. **Home page tour (3 min)**: "What is this? What can you do here?" — tests first impressions.
4. **Tasks (35 min)**: Give realistic goals and let them struggle without hints.
5. **Probing (5 min)**: Ask follow-up questions.
6. **Wrap-up (5 min)**: Thank them, pay them, let them go.
**Why It Works**:
- **Cheap**: a laptop, screen recorder, and three volunteers are enough.
- **Fast**: findings land the same day, while decisions are still being made.
- **Honest**: watching real people fumble is far more persuasive than opinions in a meeting.
- **Habit-forming**: because it's easy, teams actually do it — and testing once a month beats testing perfectly once a year.
**Common Pitfalls**:
- Leading users ("try clicking here") — always wait and watch.
- Recruiting only colleagues who know the product.
- Debating findings instead of fixing them.
- Chasing minor issues while ignoring the show-stoppers.
Guerrilla usability testing is less a methodology than a mindset: stop arguing about what users will do and go watch them do it.
Related Concepts
← Back to all concepts