Writing problem statements that guide teams
One clear sentence that saves weeks of building the wrong thing.

Zander Whitehurst
8 min read
A good problem statement is one of the most useful things a product team can have. It is a single clear sentence that describes who is struggling, with what and why it matters. Get it right and every later decision becomes easier. Get it wrong, or skip it, and teams can spend weeks building something that never quite lands.
I used to think problem statements were a box ticking exercise. Something you wrote for a slide and then forgot. Then I joined a project where three teams had been building for a month, each with a completely different idea of what problem they were solving. Nobody had written it down. The first meeting after we did was the shortest and most productive of the entire project.
What a problem statement does
A problem statement gives a team a shared target. It answers the question "what are we actually trying to fix?" in a way everyone can agree on and refer back to.
It is not a solution, a feature list or a goal. It describes the situation that needs to change. Solutions come later, and they get judged against it.
A simple structure
Most good problem statements follow a similar pattern:
[Who] struggles to [do what] because [why],
which means [consequence].
For example: "New team admins struggle to invite colleagues because the invite option is buried in settings, which means most accounts never grow beyond one user."
That one sentence tells you who to design for, what they are trying to do, what is getting in the way and why the business should care.
Be specific about who
"Users" is almost never specific enough. Different people experience the same product in very different ways. A first time visitor, a daily power user and an account admin all have different problems.
Name the specific group. "New team admins in their first week" gives designers and engineers something concrete to picture. "Users" gives them nothing.
Describe the struggle, not the solution
The most common mistake is sneaking a solution into the problem statement. "Users need a dashboard to see their data" is not a problem statement. It is a dashboard request wearing a disguise.
Ask what the person is trying to achieve and what stops them. "Managers struggle to see which projects are falling behind" leaves room for many solutions, of which a dashboard is only one.
Include the why
The "because" part is where understanding shows. It explains the cause of the struggle, and the cause is what your solution needs to address.
If you do not know the cause, that is a useful discovery in itself. It tells you that you need more research before designing, which is far better to learn now than after building.
Make the consequence matter
The final part connects the problem to something people care about. Lost revenue, wasted time, churn, support costs or frustrated customers. Without it, a problem statement can describe something real that simply is not worth solving.
This part also helps with prioritisation. When you have several problem statements competing for attention, the consequences make the comparison much easier.
Test it with the team
A problem statement only works if the team agrees with it. Share it early and ask a few questions:
- Does everyone recognise this problem?
- Is the group specific enough?
- Do we have evidence for the cause?
- Would solving this actually matter?
Disagreement at this stage is a gift. It is far cheaper to argue about a sentence than about a finished feature.
Use it to guide AI work
Problem statements are also excellent prompts. When you ask an AI agent to explore ideas or build a prototype, starting with a clear problem statement produces far more focused results than starting with a feature request.
An agent told "build a dashboard" builds a generic dashboard. An agent told "managers struggle to see which projects are falling behind because status updates are scattered across tools" builds something much closer to what you actually need.
Keep it visible
Write the problem statement at the top of every design file, document and ticket for the project. When new ideas come up, check them against it. If an idea does not help solve the problem, it can wait.
I have a slightly annoying habit of reading the problem statement aloud at the start of every review. People roll their eyes. Then, about once a month, someone realises halfway through that their design has wandered off target, and the eye rolling stops for a week or two.
Start with your current project
Take the project you are working on right now and try to write its problem statement in one sentence using the structure above. If it comes easily, share it with the team and check they agree. If it does not, that is the most important thing you will learn this week, and it is far better to learn it now. We practise writing problem statements and PRDs in the AI Product Management Bootcamp.
Memorisely