Running a Small Usability Test on a Shoestring
You do not need a lab to watch someone use your site. A quiet room, a few realistic tasks and the discipline to stay silent will show you more than a month of guessing.
What a usability test is
Usability testing evaluates a product by testing it with representative users: people try to complete realistic tasks while observers watch, listen and take notes. The point is to see where real people struggle, not to ask whether they like the design.
Recruit people who resemble your visitors
Colleagues know too much. Friends are too kind. Try to find people who match your audience in the ways that matter for the task: someone who has never used your product, or someone who uses a competitor. A handful of sessions will not give you statistics, but it will show the problems that block people outright.
Write tasks, not instructions
- Good: "You want to change the delivery address on an order you placed yesterday."
- Poor: "Click Account, then Orders, then Edit." That tests whether people can follow directions.
- Avoid using the exact words from your menus in the task; otherwise people just match labels.
Ask people to think aloud
The think-aloud protocol asks participants to say what they are thinking as they work. It reveals expectations: "I thought this would take me to the price" tells you more than a click log. Remind them gently if they fall quiet, with neutral prompts such as "What are you looking at now?"
Stay out of the way
The hardest part for the person running the session is not helping. When a participant is stuck, count to ten before saying anything. If they ask "Should I click this?", turn it back: "What would you do if I weren't here?" Every rescue hides a problem real visitors will hit alone.
Be fair to your participants
Explain what you are recording and why, and get informed consent before you start. Make it clear that you are testing the site, not them, and that they can stop at any time.
Turn notes into a short list
Prepare the room and the prototype
Check the device, the network and the test account before each session, and reset anything the previous participant changed, such as items in a basket or saved settings. A broken prototype wastes the participant's time and teaches you nothing. If you are testing remotely, ask people to share their screen and do a short practice task first, so the technology is out of the way before the real tasks begin.
Turn notes into a short list
Straight after each session, write down the three moments where the person hesitated longest. After all sessions, look for the moments that repeat. Those are your first fixes. Resist the urge to redesign everything; fix the worst blockers, then test again.
By Iris Calloway · last checked 2026-11-09