SAAS USABILITY TEST PLAN AND MODERATOR SCRIPT
Original fictional workflow and editable study plan. Checked 2026-10-27.
No real participant study was conducted for the example observations.

DECISION
Can relevant agency operators find and create a single-client export without
help? Use observations to review labels, selection controls, and feedback before
a wider pilot. This qualitative study is not a population conversion estimate.

RECRUITMENT
Relevant experience: prepares recurring client reports and understands which
rows belong to each client.
Record variation relevant to the decision: new/returning, role, device, and
accessibility needs. Choose a manageable first round, then review gaps.
No universal participant count makes every audience and risk adequately studied.

PREPARATION
Use an isolated account with fictional report data.
Report has ROW-1 and ROW-3 for Client A; ROW-2 for Client B.
Confirm tasks are possible, reset state, and verify any prototype limitations.
Do not expose customer records, tokens, billing details, or private notifications.
Pilot the script to catch impossible tasks and accidental clues.

CONSENT DISCUSSION — ADAPT TO THE ACTUAL STUDY
Explain purpose, expected duration, optional recording, material collected,
who can see it, retention, compensation if any, and how to stop.
Obtain appropriate agreement before recording. Participation is voluntary.
Use participant codes in the working notes and respect the agreed data controls.

INTRODUCTION
We are looking at how the product supports this task.
Difficulties help us understand what needs work.
Please say what you are looking for or expecting as you go.
You can stop or take a break whenever you need.
I may leave a question until after your attempt so I can see what happens.

TASK 1
You are preparing the monthly update for Client A.
Save a file containing only that client's rows so you can send it later.
Observer: note route, scope interpretation, completion, and any help.
Do not name the filter control in the task prompt.

TASK 2
Before saving, check which client will be included in the file.
Observer: note what evidence the participant uses to establish scope.

TASK 3
The test environment will simulate a generation failure.
Continue as you would if you still needed the file.
Observer: note whether failure is noticed and what next action seems available.
Tell the participant about simulation after the attempt where appropriate;
do not misrepresent prototype limitations as their mistake.

NEUTRAL FOLLOW-UPS
What are you looking for now?
What did you expect to happen?
What tells you the task is complete?
Avoid telling the participant which control to use.

STOPPING AND ASSISTANCE
Stop if the participant would give up, the planned task boundary is reached, or
continuing creates an inappropriate burden.
Record the exact assistance and mark subsequent completion as assisted.
Do not pressure someone to finish to make the interface appear successful.

DEBRIEF
What did you expect at the point you paused?
How would you normally complete this task?
What would you check before sharing the result?
A friendly comment does not erase an observed barrier.

OBSERVATION AND ANALYSIS
Separate behavior, words, interpretation, outcome, assistance, and next check.
Group issues by task consequence. Preserve uncertainty about cause.
A proposed redesign is an idea to test, not a verified improvement.
Keep severe one-off failures visible without claiming a prevalence estimate.

YOUR SESSION
Decision / participant criteria / test environment:
Consent and recording plan / retention:
Task prompts / expected result / limitations:
Stopping boundary:
Observer / moderator:
Evidence location / issue review date / next decision: