Skip to main content

UX and research · Beginner · 14 min

Interview stakeholders at the start of a project

By Padding · Updated 18 Nov 2026
Any document editor and a spreadsheet · November 2026

You will have run short interviews with the people who own, fund and operate a product, and written a one-page summary of where they agree, where they conflict and what success means to each.

On this page

Stakeholders are the people who own, fund or run the thing you are about to design. They are not your users, though some of them may also use the product. Each one carries a picture of what the project is for, and those pictures rarely match. If nobody compares them at the start, the differences show up at the first design review, when a doctor and a practice manager disagree in front of you about a screen that took two weeks to make.

A stakeholder interview is a short, one-to-one conversation that drags those pictures into the open while changing course is still cheap. This tutorial uses a running example: Larkfield Clinic, a fictional community clinic that wants an appointment booking app where patients choose a service, pick a time, confirm and get a reminder. You will interview three people there: a receptionist, a doctor and the practice manager. Swap in your own project's people as you go.

By the end you will have a stakeholder list, a question sheet, notes from each conversation in one shared format, a comparison grid, and a one-page summary you can send to everyone you spoke to.

Step 1 · List who has a stake and why

Open a new document and make a table with four columns: Name and role, What they own or decide, How the project changes their day, What they could block.

Fill one row per person. Think in three groups:

  • Owns: the person who will be held responsible if the product fails. At Larkfield, the practice manager.
  • Funds or approves: whoever signs off budget, legal or clinical safety. Sometimes the same person as the owner; write them twice if they wear both hats, so you remember to ask about each.
  • Operates: people whose daily work changes when the product ships. At Larkfield, the receptionist (fewer phone bookings, more no-show follow-ups) and the doctor (a schedule that fills differently).

The last column matters most. A stakeholder who can block a launch but was never asked is the usual source of late surprises. If you are unsure whether someone belongs on the list, ask the project lead "who would be annoyed to find out about this after it launched?"

Aim for three to six interviews. Fewer than three and you have nothing to compare; more than six at this stage and the summary stops fitting on one page.

Check it worked

Your table has at least three rows, every row has something written in all four columns, and at least one row covers each of the groups owns, funds or approves, and operates. If a "What they could block" cell says "nothing", ask yourself again whether they need an interview at all.

Step 2 · Write one question sheet for everyone

Everyone gets the same core questions. That is what lets you compare answers later; if you improvise a different conversation with each person, you cannot tell a real disagreement from a difference in what you asked.

Write the sheet in five short sections, about 30 minutes in total:

  1. Their role (3 minutes). "Walk me through how a booking happens today, from your side." Ask them to describe the last real one, not the usual one.
  2. Goals (7 minutes). "If this app works, what is different for you in six months?" Then: "What would make you say it was a waste of money?"
  3. Success (5 minutes). "How would you know it is working? What would you look at?" Push for something observable: a number, a report, a complaint that stops arriving.
  4. Constraints (8 minutes). "What must this app never do?" "What rules, systems or contracts does it have to fit around?" "What has been tried before?"
  5. Other people (5 minutes). "Who else should I talk to?" and "Where do you expect other people here to see this differently?"

The last question is the one that surfaces disagreement early. People often know exactly where the friction is and will tell you if asked directly in private, when they would never raise it in a group meeting.

Keep the wording neutral. "Should patients be able to cancel online?" invites a yes or no and hints at your own view. "What should happen when a patient can't make their slot?" lets them describe it.

Check it worked

Your sheet fits on one page, has no more than about 12 questions, and every question can be answered by any of your stakeholders. Read each one aloud: if any contains "don't you think", "would it help if" or an answer option, rewrite it as an open question.

Step 3 · Book the conversations and set them up

Send each person a short message. Say what the project is, why you want their view, how long it takes (30 minutes), and that their individual answers will not be quoted by name in the summary. That last line is what makes people candid about each other.

A message for Larkfield's receptionist might read:

We're starting work on an online booking app for the clinic. Before any design work, I'm talking to a few people about what it needs to do and what could go wrong. You handle more bookings than anyone, so your view matters. It's 30 minutes, one to one. The summary I share afterwards groups views together and won't name who said what.

Book them one to one, never as a group. In a group, the most senior person sets the answer and everyone else nods. Book them close together, ideally within the same week, so the project does not move between the first and last interview.

Before each conversation, ask if you may record it for your own notes. If they say no, take written notes and do not push.

Check it worked

Every person on your stakeholder list has a confirmed time in your calendar, each slot is one to one, and the first and last interviews are no more than about a week apart. Your invitation text says how the answers will be used.

Step 4 · Run each interview and take notes in a fixed format

Make a notes template with the same five headings as your question sheet, plus two more at the bottom: Words they used and Surprises. Use a fresh copy for each person.

During the interview:

  • Ask the question, then wait. Silence after an answer often produces the most useful second sentence.
  • When you hear a goal, ask what it looks like in practice. "Fewer no-shows" becomes "the Monday afternoon list is full by Friday".
  • When they mention another person, write the name down and ask "how do you think they would see it?"
  • Write their own words in quotes under Words they used. "Slot", "appointment", "session" and "booking" may mean different things to different people, and you need that vocabulary later for screen labels such as Pick a time.

Do not defend ideas, pitch a solution or correct them. You are there to collect their view, not change it.

Fill the gaps in your notes straight after each call while it is fresh. Ten minutes then saves an hour of guessing later.

Here is an excerpt from the doctor's notes at Larkfield:

Success: "I'd look at how many slots go empty. Right now Tuesday mornings are a mess."Constraints: Urgent cases must always keep two slots a day that the app cannot book.Other people: Thinks reception will want patients to call for anything other than a routine check-up.

Check it worked

You have one notes document per stakeholder, each with all seven headings filled in, and every Words they used section has at least two quoted phrases. Any heading left empty has a note saying why (for example "ran out of time; follow up by email").

Step 5 · Compare answers in a grid

Open a spreadsheet. Put the stakeholders across the top as columns. Down the side, put one row per theme: Main goal, Success measure, Must never do, Constraints, Words for a booking, Expected conflict.

Fill each cell with a short paraphrase from that person's notes, a few words long. Then add one more column on the right called Status and mark each row:

  • Agree where everyone says roughly the same thing.
  • Conflict where two or more people want things that cannot both be true.
  • Gap where only one person mentioned something, so nobody else has had a chance to agree or object.

At Larkfield, the grid might show:

  • Main goal is a conflict. The practice manager wants fewer phone calls; the receptionist wants fewer no-shows and is worried online booking makes people book and forget.
  • Must never do has a gap. Only the doctor mentioned keeping urgent slots out of the app.
  • Success measure looks like agreement once you read past the wording: all three care about empty slots, even though they describe it differently.

Treat "conflict" carefully. Two people using different words for the same thing is not a conflict; two people wanting the same slot to behave in two different ways is.

Check it worked

Every cell in your grid is filled or marked "not mentioned", every row has a status, and you can point to at least one row marked Conflict or Gap. If every row says Agree, re-read your notes for the "Other people" answers; it is rare for three stakeholders to agree on everything.

Step 6 · Test each conflict before you write it down

A conflict in your grid is still your interpretation. Before it goes in front of everyone, check it with the people involved.

For each row marked Conflict or Gap, send a short follow-up to the people concerned, one at a time. Describe the other view without naming the person, and ask a question:

One view I heard is that keeping booking on the phone for anything beyond a routine check-up would protect reception's time. Another is that the app should take most bookings to cut call volume. Where would you draw the line?

Three things tend to happen:

  • The conflict was a misunderstanding, and it disappears. Mark it Agree.
  • The conflict is real, but they suggest a middle ground. Note it as a possible resolution, not a decision.
  • The conflict is real and they hold firm. Keep it as Conflict; it needs a decision from whoever owns the project.

Gaps work the same way. Ask the others whether they agree with the one person who raised it. At Larkfield, asking the practice manager about reserved urgent slots turns a gap into an agreed constraint before anyone designs the Pick a time screen.

Check it worked

Every Conflict and Gap row in your grid has a note of who you followed up with and what they said, or a dated note that you asked and are waiting. No row is still marked Conflict only on the strength of your own reading.

Step 7 · Write the one-page summary

Start a new document. Give it four sections, in this order:

  1. What we agree on. The rows marked Agree, written as plain statements. "Success means fewer empty slots. Urgent slots stay off the app."
  2. Where views differ. Each remaining conflict as a neutral question, with the options heard. "How much booking moves online? Option A: routine check-ups only. Option B: all services except urgent care." Do not name who holds which view.
  3. What success looks like. The measures people named, grouped. Say who would track each one by role if that came up.
  4. Decisions needed, and by whom. One line per open conflict: the question, the role that decides, and a date. This is the section that stops disagreements resurfacing at the first design review.

Add one line at the bottom listing who you spoke to by role and when.

Keep it to one page. If it runs over, cut detail from the agreement section first; people read agreements quickly, and the conflicts are the part that needs attention.

Send it to every stakeholder you interviewed, with one question: "Does this match what you told me, and is anything missing?" Give them a reply date.

Check it worked

Your summary fits on one page, contains no stakeholder names attached to opinions, and has at least one item in Decisions needed with a role and a date beside it. Every stakeholder you interviewed is on the recipient list.

Next steps

You now have a written record of what the people around the project want and where they pull in different directions, before any screen exists. Use it as an input to the research and design work that follows:

Need this done for you?

Padding does this work for growing companies, from the first diagnosis to a system that keeps working.

Digital Transformation

Get new tutorials

New tutorials, about twice a month.