Skip to content

The annotated journal

What should you change after a small usability test?

One observation can justify a repair. A hypothetical library test shows how to decide what to fix, what to investigate and when to test a new label.

Chris Taylor

A hand lifts a paper arrow pointing towards the loan-card pocket in an open library book.

If you can verify the problem and the fix has limited consequences, one observation is enough to act. I’d apply that rule in the debrief, while everyone can still remember what the person was trying to do. A decision to wait should be just as specific. What do you still need to know, and would the answer change what you do?

Hypothetical example — renewing a library loan

Imagine a borrower trying to keep a book for another week. They follow Renew your loans and land on the catalogue search. In the same round, three other borrowers go straight to the catalogue and search for the titles they have at home. Someone suggests renaming My account to My loans.

Start with the link. Sign in as a borrower with a book on loan and follow Renew your loans. If you land in the catalogue too, check the intended destination. Can you find the book there and complete a renewal? Check where else the link is used. If it belongs to one page, correct the destination and try the journey again from that page. If editing it would change a shared menu, check which journeys use that menu before you make the change.

Why act after seeing one borrower struggle? The link promises a particular action, and you can check for yourself where it leads. You can also tell whether the repair worked by following it through to the renewed loan.

Why did the other borrowers search for the book?

“Three borrowers searched for their book” is a useful debrief note. “Borrowers don’t understand My account” adds an explanation that the observation alone cannot support. GOV.UK’s research guidance recommends recording what participants did or said before interpreting it and deciding what to change. Analyse a research session

Go back to the first screen and the moderator’s exact words. If the moderator mentioned the catalogue when introducing the task, try a neutral prompt before editing the page. If My account was out of view when the task began, check whether the link came into view before the borrowers searched. The count tells you the same route came up three times in this round. How often borrowers take it across the service is a separate question.

Give the new label a fair test

If the wording still looks like a plausible cause, try My loans on a copy of the same screen. Keep the link in the same place. Give fresh borrowers a goal they might recognise. “Your book is due back tomorrow. You’d like to keep it for another week. Show me what you’d do.” Let them choose where to start. Watch whether they reach their existing loan without a hint. GOV.UK guidance on choosing participants and writing test tasks

There’s a cost to narrowing the label. The account also holds reservations for books the borrower has yet to collect. Start a reservation task from the same opening screen. Give some borrowers that task first, so finding their loans hasn’t already shown them the way. If they can find their loans but overlook their reservations, you have a reason to reconsider the wording before changing the shared menu.

What to put in the debrief notes

For this hypothetical round, I’d leave the team with two entries.

  • Renewal link — content owner. Confirm where the link is used, correct its destination and check that a borrower can renew a book starting from the original page.
  • Account label — researcher. Check the prompt and what was visible when borrowers began searching. If the wording still looks like the issue, test My loans with both a renewal and a reservation task before deciding whether to rename it.