UX and research · Beginner · 12 min
Turn a feature request into a user-centred design brief
By Padding · Updated 18 Nov 2026
Any notes app or document editor · November 2026
You will turn a one-line request into a one-page brief that names the people affected, what they are trying to get done, what is known and what is guessed, ready to hand to a team.
On this page
Most design work starts with a sentence from someone else. "Add online booking." "We need a dashboard." "Can the app send reminders?" The sentence names a solution. It does not say who needs it, what they are struggling with today, or how anyone will know it worked. If a team designs straight from that sentence, they build what was asked for and find out later whether it helped.
User-centred design means the people who will use the thing shape it from the start. The cheapest place to begin is the brief: before anyone draws a screen, rewrite the request around those people. In this tutorial you do that to one request, line by line, and end with a page you can hand to designers, developers and the person who asked.
The running example is Larkfield Clinic, a fictional community clinic. Its practice manager sends this message: "Can we add online booking to the website? The phones are a nightmare." Follow along with that, or swap in a request from your own work. Every step works the same way.
Step 1 · Write down the request exactly as it arrived
Open a blank document and title it "Brief: " followed by a working name, such as "Brief: Larkfield appointment booking". Under a heading called Original request, paste the request word for word. Do not tidy it. Then add three lines:
- Who asked: the person and their role. For Larkfield, the practice manager.
- When and how: the date and the channel, such as an email on 4 November.
- What they said around it: any reason, complaint or deadline mentioned in the same message. Here, "the phones are a nightmare".
Keeping the original wording matters later. When you show the finished brief, the person who asked can see their request in it and see how it has changed. The words around the request are often more useful than the request itself: "the phones are a nightmare" is a problem, while "online booking" is one possible answer to it.
Check it worked
Your document has an Original request section with the request in quotation marks, unedited, and three filled-in lines underneath for who asked, when and how, and the surrounding words. If you cannot fill in the third line, write "Nothing else said" so the gap is visible.
Step 2 · List everyone the request touches
A request comes from one person but lands on many. Under a heading called People affected, list every group whose day changes if this request is built, not only the people who would tap a button.
Work outwards from the request. Who does the task today? Who receives the result? Who fixes it when it goes wrong? Who is left out if it moves online? For Larkfield, that gives something like:
- Patients booking for themselves
- Parents and carers booking for someone else
- Patients who rarely use the internet, or cannot easily use it
- Reception staff, who answer the phones and manage the diary
- Clinicians, whose day is shaped by what gets booked
- The practice manager, who asked
Write one line for each group saying how the request reaches them. For reception staff: "Fewer calls, but they may have to fix online bookings that went wrong." For patients who rarely go online: "Nothing improves for them unless the phone route stays good."
Keep the list to groups, not invented individuals. Personas come later and should come from research, not from a brief.
Check it worked
Your People affected section has at least four groups, each with one line saying how the request changes things for them. At least one group is someone the original request did not mention. At least one line names a risk or a cost for that group, not only a benefit.
Step 3 · Write what each group is trying to get done
For each group, write what they want to achieve in their own terms, without mentioning the feature. Use the form "When [situation], they want to [outcome], so that [reason]." Add these under each group in People affected, or in a new section called What people are trying to do.
For Larkfield:
- Working patients: when they feel unwell before work, they want to know whether they can be seen today, so that they can decide whether to go in or take the day off.
- Carers: when they book for a parent, they want to choose a time that fits around both diaries, so that they do not have to call back.
- Reception staff: when the phones open at 8am, they want to deal with urgent callers first, so that the people who most need help are not stuck in the queue.
- Clinicians: when their list fills up, they want the right kind of appointment in the right slot, so that a 10-minute slot is not booked for a 30-minute problem.
Now read the lines back and look for anything "online booking" would not fix. "Know whether they can be seen today" might be met by showing availability without full booking. "Deal with urgent callers first" might need triage, not a booking form. These gaps are the reason for writing the brief. Note them under a small heading called Things the request might not solve.
Check it worked
Every group from the previous step has at least one "When ..., they want to ..., so that ..." line, and none of those lines contain the words "online booking", "app", "website" or "button". You have at least one item under Things the request might not solve.
Step 4 · Separate what you know from what you are guessing
Most of what you have written so far is reasonable belief, not evidence. A brief that blurs the two leads a team to treat guesses as facts. Make a table with three columns: Statement, Known or guessed, Source or how to find out.
Go through your document and copy each claim into the table. Then be strict about the middle column. "Known" needs something you can point to: a phone log, an analytics report, a complaint record, a conversation with a named person. Everything else is "guessed".
A few Larkfield rows:
- Most calls arrive between 8am and 9am. Known. Phone system call log, October.
- Patients would rather book online than phone. Guessed. Short survey in the waiting room, or five patient interviews.
- Reception spends most of its time on booking calls. Guessed. One morning sitting with reception, tallying call types.
- Some patients cannot use online booking. Guessed. Ask reception who struggles with the current patient portal.
- Clinicians want control over which slots go online. Known. Said by the lead nurse in the staff meeting on 6 November.
Each line above is one row of the table: the statement, then known or guessed, then the source.
Expect most rows to be guessed. That is normal at the start. The value is that each guess now has a way to check it, and the team can decide which guesses are worth checking before design begins.
Star the two or three guesses that, if wrong, would change what gets built. For Larkfield, "patients would rather book online" and "reception spends most of its time on booking calls" both qualify: if most calls are about test results, not bookings, online booking will not ease the phones much.
Check it worked
Your document has a table where every row is marked either known or guessed, every known row names a specific source, and every guessed row names a way to find out. Two or three rows are starred as the riskiest guesses.
Step 5 · Rewrite the request as a problem and an outcome
Now rewrite the request. Under a heading called The problem, write two or three sentences that describe what is going wrong, for whom, and why it matters, without proposing a fix. Under The outcome we want, write what will be true when the problem is solved.
For Larkfield:
The problem: Patients who need an appointment ring at 8am and wait in a long queue, and some give up. Reception staff spend the busiest hour of the day on routine bookings, which makes it harder to help urgent callers. We do not yet know how many calls are bookings and how many are about other things.
The outcome we want: Patients can get a suitable appointment without waiting on the phone, people who prefer to phone still can, and reception has more time for urgent callers in the first hour.
Notice what changed. The original request was a solution ("add online booking"). The rewrite is a problem the team can solve in more than one way: online booking, a call-back service, sharing same-day availability, or a mix. Online booking may still be the answer. It now has to earn its place.
Then add a How we will know line with two or three signs of success the team could actually observe. Keep them tied to the outcome and to data you can collect: "Fewer calls abandoned between 8am and 9am, compared with the October call log" is better than "patients are happier".
Check it worked
Your The problem section names at least one group of people and does not contain the proposed feature. Your The outcome we want section would still make sense if the team chose a different solution. How we will know has at least two measures, and each one names where its numbers will come from.
Step 6 · Set the edges of the work
A brief also says what is out of scope and what cannot change. Add three short sections.
In scope: what this piece of work covers. For Larkfield: "Routine appointments with GPs and practice nurses, booked by patients or carers."
Out of scope: what it does not cover, so no one assumes it does. For Larkfield: "Urgent same-day triage, repeat prescriptions, test results." Write each item with a brief reason, such as "triage needs clinical rules we have not agreed yet".
Constraints: the fixed things the design must work within. For Larkfield, these might include the clinic's existing appointment system, the rule that clinicians decide which slots are bookable, staff time to support any change, and accessibility, since a clinic serves people with a wide range of needs. Mark each constraint known or guessed, as you did in the table. "We must use the current appointment system" is often an assumption no one has questioned.
Finish with Open questions: things the team needs answered and who can answer them. Turn your starred guesses into questions here, each with a name next to it. "What share of 8am calls are routine bookings? Reception lead, by 15 November."
Check it worked
Your document has In scope, Out of scope, Constraints and Open questions sections. Every out-of-scope item has a reason, every constraint is marked known or guessed, and every open question has a named person who can answer it.
Step 7 · Cut it to one page and test it on someone new
Put the sections in this order and cut until the brief fits on one page:
- Original request (the quote and who asked)
- The problem
- The outcome we want, and how we will know
- People affected and what they are trying to do
- What we know and what we are guessing (keep only the starred and most useful rows)
- Scope, constraints and open questions
Move anything you cut into a second page called Notes. Nothing is lost, but the first page stays readable in a few minutes.
Then test it. Give the page to someone who has not seen the request, such as a developer, another designer or a colleague from a different team. Ask them to read it and then tell you, without looking: who this is for, what problem it solves, and what the team does not yet know. Write their answers down. Where their answer differs from what you meant, the brief is unclear at that point, so rewrite that section.
Finally, send it back to the person who asked. For Larkfield, the practice manager should be able to see "online booking" in the original request, follow how it became the problem statement, and agree or correct the people and outcomes. Their correction is useful too: it is the first piece of evidence the brief gains after you wrote it.
Check it worked
Your brief fits on one page with the six sections in order, and a separate Notes page holds what you cut. You have written down one reader's answers to the three questions, and each answer matches the brief, or you have revised the section that caused the mismatch. The person who asked has seen the brief, and you have recorded whether they agreed or what they changed.
Next steps
Your brief lists guesses and open questions. The next job is to decide how to answer them and when to move from exploring to designing. These tutorials pick up from here:
- Plan a project with the double diamond to turn the brief into a plan for discovery and design.
- Write a problem statement before you design to sharpen the problem section once you have research.
- Map and rank your design assumptions to decide which guesses to test first.
- Interview stakeholders at project start to fill in the open questions with the people who can answer them.
Part of these paths
Related tutorials
- Plan a project with the double diamondYou will have a two-week project plan drawn as a double diamond, with dated phases, the activities and outputs for each, two decision gates and a note of what each shortcut would cost.
- Write a problem statement before you designYou will have one problem statement your team agrees on, naming a person, a situation and a consequence, plus a log of rejected drafts with the reason each one failed.
Need this done for you?
Padding does this work for growing companies, from the first diagnosis to a system that keeps working.
Get new tutorials