You hand a customer a task, they try it on your site, and you watch where they get stuck. That is the whole method. The discipline is in how you set it up and what you do afterward.
What is usability testing?
Usability testing is watching people who match your customers try to complete specific tasks with your product, so you see where it fails them. It works on a live website, a checkout, a loan application, a phone app, a printed form, or a clickable prototype that is a week old. The point is behavior. You learn what people do, which is often different from what they would tell you in a survey.
How many people do I need?
Five per round is the common practice, and it finds most of the problems big enough to cost you money. By the fourth or fifth person you start seeing the same stumbles again. Fix those, then run another five. Two small rounds beat one big one, because the second round tests your fixes.
Recruit people who match the buyer: the same kind of job, the same reason for coming to the site. Your staff, your family, and your developers are the wrong testers. They already know where everything is.
What do I have them do?
Give them tasks, never opinions to offer. "Find a mortgage rate for a house around the price you would buy" tells you something. "What do you think of the homepage?" tells you nothing you can act on.
- Write four to six tasks that match why people come to you: buy, book, apply, compare, find the hours.
- Use their words. If the menu says "Solutions," the task should not.
- Ask them to think out loud while they work.
- Note where they hesitate, backtrack, or give up, and what they said at that moment.
Do I help them when they get stuck?
No. Watch, do not help. The moment you point at the button, you have lost the finding, because your real customers will not have you sitting next to them. If someone is completely stuck, note it, say "that is useful, let's move on," and start the next task.
Remote or in person?
Both work. Remote sessions on a video call with screen sharing are faster to schedule and let you test people anywhere in the country. In person is better for physical products, kiosks, older customers, or anything where you need to see hands and faces. Record every session, with permission, so the people who could not attend can watch the worst two minutes.
What do I do with the findings?
Rank the problems by how many people hit them and how much each one costs you, then fix the top three before anything else. A good report is a short list: the problem, what people did, what they said, and the change to make. Show the decision-makers the video clips. One clip of a customer failing to find the "apply" button settles arguments that a slide deck never will.
At U.S. Bank I led UX on the consumer mortgage business with a team of six to seven. Testing with real borrowers was a regular part of the work, alongside funnel analysis and prototyping, and every tracked metric went up 5% to 45%. The details are in the U.S. Bank mortgage case study.
What does usability testing cost?
Start with a one-hour second opinion for $150. We go through the site or product together, and I tell you which tasks to test first and who to recruit. A full study (recruiting, tasks, sessions, and a ranked list of fixes) is quoted as one fixed study price on UX research and leadership, so you know the number before it starts. Anything over $3,000 can be paid by the month.
Answered by Jared Lukes, Minneapolis. Twenty-five years in brand, UX, and technology, currently holding fractional AI and technology seats at three companies. About Jared.