Usability Testing That Shows Where People Get Stuck
Sessions in which real users attempt real tasks on your prototype or live software while we watch where they hesitate, guess or give up. GullySystem runs the testing and returns a ranked list of fixes.
What a Usability Test Actually Involves
A usability test is not a demo and not a feedback meeting. It is one participant, one facilitator, a set of tasks written from the real job, and no help offered. The participant works through each task while saying what they are thinking, and we record every point where they stop, hesitate, take the wrong path or ask a question the screen should have answered. Running the same tasks with a handful of people per role surfaces the problems worth fixing, because serious ones repeat quickly. GullySystem writes the tasks, helps recruit and schedule participants, moderates the sessions, and turns the recordings into an ordered list of what to change.
When Testing Is Worth Doing
Testing earns its cost at particular moments rather than continuously. These are the points where it pays back quickly.
Before a Build Budget Is Committed
Testing a clickable prototype is the cheapest possible moment to learn that a flow does not make sense. Changing the screen then costs a revision; changing it after release costs a release.
When Support Keeps Answering the Same Question
A repeated support question is a screen failing in public. Testing that screen with people who do not already know the answer shows exactly which word or button is doing the damage.
When Staff Have Built Workarounds
A parallel register, a WhatsApp group and a spreadsheet kept just in case are signs that part of the job is not trusted to the software. Watching the job being done shows why.
Before a Redesign, and Again After It
Testing the current software first gives you a record of where people struggle, so the redesign can be judged on whether those specific struggles disappeared rather than on whether it looks nicer.
When Two Opinions Cannot Be Settled
Two versions of a screen and a room full of disagreement. Putting both in front of people who do the job settles it with evidence instead of seniority.
When Customers Use the Software Themselves
Dealer portals, patient booking and self-service payment have no training session and no helpdesk in the room. What an untrained person does on a first attempt is the only measure that counts.
How We Run a Round
A round is a small set of sessions run to a written script, with the same tasks put to every participant in the same words.
Tasks Written in the User's Own Words
Tasks are phrased as the job, not as the software — record today's delivery, not use the delivery module — so we are testing the screens rather than the wording of our own instruction.
The Right Participants
Testing with your management team teaches you very little. We test with the people who do the work: counter staff, drivers, dealers, parents, patients. We help write the screening criteria and arrange the schedule.
Moderated, In Person or Remote
Sessions at your premises where the work happens, or over a call with screen sharing when participants are spread across cities. Recorded with the participant's consent and nothing recorded without it.
No Rescuing, No Steering
The facilitator does not help, does not explain and does not defend the design. Every rescue destroys the finding that was about to appear.
Findings Ranked by Damage
Each issue is written up with what the person was trying to do, what actually happened, how many participants hit it and what it costs you: a wrong entry, an abandoned task, a support call.
Fixes, Not Just Complaints
The report ends with specific changes to specific screens, split into what should be fixed before release and what can wait, with redrawn screens for the serious ones.
What You Receive
The output is something your team can act on in a working week, not a library of recordings nobody will open.
The Task Script and Screening Criteria
Agreed with you before any session runs, so you know exactly what will be asked and of whom, and can add a task you have been arguing about internally.
Recordings and the Moments That Matter
Session recordings kept with consent, plus short clips of the points where things went wrong — far more persuasive in a management meeting than a written summary.
A Ranked Issue List
Every issue tied to the task and screen where it appeared, ordered by how much damage it does rather than by how easy it is to fix.
Redrawn Screens for Outright Failures
Where a task could not be completed, you get a proposed screen rather than a description of the problem, so the next step is a build decision and not another design discussion.
A Readout With Your Team
A working session where we go through the findings with the people who will act on them, while the evidence is fresh and the decisions can be made in the room.
What Shapes a Testing Round
Rounds differ in cost mainly by who has to be in the room and how hard those people are to reach.
- How many user types are being tested, since each role needs its own participants and its own task script
- Who recruits the participants — your own customers and staff are quick to arrange, external recruitment is not
- Whether sessions are run in person at a site or remotely over a call
- Whether we test a prototype, which means the tested paths must be built as clickable screens first, or live software, which needs a safe environment and disposable data
- The languages participants work in, since moderation and transcription follow the participant, not the office
- Whether a second round after the fixes is part of the scope
Testing as Part of a Larger Design Project
Testing is most useful when there is a prototype to test and a team ready to act on what it finds. Both come from the wider design service.
Frequently asked questions
How many people do you need to test with?
Fewer than most owners expect, and more than one. The same serious problems repeat quickly across participants, so a small group per role usually reveals them, and adding many more mostly confirms what you already saw. We agree the number per role during planning, based on how many distinct types of user you have.
Can you test software that is already live, or only prototypes?
Both, and they answer different questions. A prototype tells you whether a planned flow makes sense before it is built. Live software tells you where your current screens are costing you today. For live testing we ask for a test environment and data participants may safely change.
Our users are spread across several cities. Does testing still work?
Yes. Remote sessions over a call with screen sharing work well for anything done at a desk, and they widen the pool of participants considerably. Work that involves a counter, a warehouse or a vehicle is better observed in person, because half the finding is in what is happening around the screen.
What does a usability testing round cost depend on?
The number of user roles being tested, how many sessions each needs, whether we travel to sites or run remotely, who finds the participants, and whether participants receive a token of appreciation. Preparing a prototype for testing, when the paths are not yet clickable, is quoted separately.
How quickly can a round be arranged and reported?
Scheduling participants is almost always the longest part, especially when they are your customers rather than your staff. Once sessions are booked, the testing itself runs quickly and the report follows the last session closely. If you need results before a specific release date, tell us that first and we plan backwards from it.
Do we get the recordings and the raw notes?
Yes, along with the report. Recordings are shared with you under whatever consent the participants gave, and the notes and clips are yours to use internally. Participant consent forms are agreed with you before the round, so you know what may be shown to whom.
What do you need from your side to run the sessions?
Access to genuine users, which is the one thing we cannot substitute. Beyond that: a quiet room or a working call setup, a test login with usable data, agreement on the tasks, and somebody from your team watching at least one session. Watching a real user for twenty minutes changes more minds than any report.
Tell us what you need.
Send a short brief and one of our engineers will come back to you — usually the same day.
- No obligation
- We reply the same working day
- Your details stay private