How to Prioritize Product Features: A Repeatable Process With Examples
A repeatable way to decide what to build: the process, the methods compared, a worked example, and how to say no without losing trust.
Prioritize against a goal, not a backlog. Gather options as problems, score them with one simple shared method (RICE, ICE, value versus effort or MoSCoW), debate the top ten with design and engineering, then decide and publish what you chose not to do. Reserve capacity for bugs, and say no with a reason and a next step.
- Start from the goal, not the backlog
- A process you can repeat every quarter
- The methods, and when each one fits
- A worked example: eight items, one decision
- Handling stakeholders and “urgent” requests
- How to say no and keep trust
- When your goals conflict
- How often to reprioritize
- Stakeholder types and how to respond
- A one-page prioritization note you can copy
- Common mistakes
Every product team has more ideas than time. That is not a problem to be solved; it is the permanent condition of the job. What separates good product managers is not that they pick the perfect item every time. It is that they have a repeatable way to decide, they can explain it, and they revisit it when the facts change. This guide gives you that process, and shows how the well-known methods fit into it.
Start from the goal, not the backlog
The first mistake is to open the backlog and start ranking. Prioritization only makes sense against a goal. “What should we build next?” has no answer; “what should we build to raise week-4 retention?” does. Pick one or two goals for the period, write them at the top of the page, and reject anything that does not serve them, however good it is on its own.
A process you can repeat every quarter
Gather options from support tickets, sales requests, research, data and engineering. Write each as a problem, not a feature. Score them with a simple lens (next section). Debate the top ten with design and engineering, challenging the inputs: is that reach real? is that effort realistic? Decide and explain: publish the result together with what you chose not to do and why.
The methods, and when each one fits
| Method | How it works | Best for | Watch out for |
|---|---|---|---|
| RICE | Reach × Impact × Confidence ÷ Effort | Comparing many ideas with some data | False precision; see the RICE guide |
| ICE | Impact × Confidence × Ease, each 1–10 | Quick triage of small experiments | Subjective; ignores reach |
| Value vs effort | Plot ideas on a 2×2 | Workshops with many stakeholders | Value is a guess |
| MoSCoW | Must / Should / Could / Won’t | A release with a fixed date | Everything becomes a “Must” |
| Kano | Basic needs, performance, delighters | Deciding what delights vs what is expected | Needs survey data |
| Weighted scoring | Score against several weighted criteria | Complex decisions with multiple goals | Weights are arbitrary |
Do not agonise over the choice. Any of these is far better than none, and they usually agree on the top and bottom of the list. The differences appear in the middle, and that is where you should apply judgment.
A worked example: eight items, one decision
A learning app with the goal “raise week-4 retention”. The backlog, and the reasoning (illustrative):
| Item | Serves the goal? | Value | Effort | Decision |
|---|---|---|---|---|
| Shorter practice sets | Directly | High | Low | Do first |
| Reminder nudges | Directly | Medium | Low | Do first |
| Progress summary emails | Yes | Medium | Medium | Next |
| Peer study groups | Yes, indirectly | High if it works | High | Test small first |
| Dark mode | No | Low | Low | Not now |
| New payment option | No (revenue goal) | Medium | Medium | Not this quarter |
| Full redesign of home | Unclear | Unknown | High | Reject for now |
| Fix crash on old phones | Yes, hurts retention | High | Low | Do first (bug) |
Three things to notice. Two items were rejected simply because they do not serve this goal, not because they are bad. One large, risky idea (peer groups) got a small test rather than a yes or no. And a bug fix was included, because retention lost to crashes is retention lost. That is what a real priority list looks like: not a perfect ranking, but a set of defensible decisions.
Handling stakeholders and “urgent” requests
Everyone believes their request is the priority. Your job is not to pick a winner in private; it is to make the trade-off visible. When a senior person asks for something, answer with: what problem it solves, where it ranks on the shared list, what it would displace, and when you could do it. Most of the time, seeing what gets pushed out is enough for them to reconsider or to argue for a real reason.
What about bugs and technical debt?
Reserve capacity for them explicitly instead of making every bug fight a feature in the ranking. A steady slice of each cycle for quality keeps the product from slowing down, and stops you having to choose between “new” and “broken” every week. When a bug directly affects your goal metric, it belongs in the main list, as above.
Tie-breakers when two items look equal
- Which one teaches you more if it works or fails?
- Which one is easier to reverse if you are wrong?
- Which one unblocks other work?
- Which one do your most engaged users keep asking for?
How to say no and keep trust
A useful pattern: “I understand why this matters, especially for [problem]. Right now our goal is [goal], and this ranks below [items] because [reason]. If it moves up we would have to drop [item]. I have added it to the not-now list and will revisit it in [month].” It acknowledges the problem, gives a reason, shows the trade-off, and offers a next step. People accept a clear no far better than a vague maybe.
When your goals conflict
Sometimes two goals pull in opposite directions: grow signups and improve the quality of new users, or ship faster and reduce bugs. Do not hide the conflict in a score. Name it, decide which goal wins this quarter, and write it down. A useful test is “if we could only hit one, which would we choose?” If leadership cannot answer, you have a strategy problem and no prioritization method will fix it.
How often to reprioritize
Reprioritize deeply each quarter, lightly each month, and only for real emergencies in between. Constant reshuffling feels agile and mostly destroys focus: teams pay a switching cost every time the top item changes. A good rule is that an item in progress is only replaced if something has genuinely changed, such as a customer-critical failure or a new fact that alters the goal.
Stakeholder types and how to respond
| Stakeholder | What they usually want | Helpful response |
|---|---|---|
| Sales lead | A feature to win a specific deal | Ask how many deals share the need; add it to evidence, not to the top |
| Executive | A pet idea or a competitor’s feature | Ask what outcome it moves; show it on the shared list |
| Engineering lead | Time for technical health | Agree a standing share of capacity |
| Support | Fixes for frequent complaints | Bring the ticket counts; high-frequency pain deserves a high score |
A one-page prioritization note you can copy
Goal this period: [ ]
What we are doing, in order: 1. [ ] 2. [ ] 3. [ ]
Why these: [one line each, with the evidence]
What we are not doing, and why: [ ]
What would change this: [ ]
Circulate it before the planning meeting. People argue more usefully with a written proposal than with a blank whiteboard.
Common mistakes
Ranking without a goal. Letting the loudest voice win. Scoring precisely on guessed inputs. Never removing anything from the list. Refusing to say no. Treating prioritization as a one-time event instead of something you redo when facts change. And confusing “most requested” with “most valuable”: the number of people asking is evidence, not a verdict.
Next: put the result in a roadmap, write the top item up as a PRD, and see how this reasoning gets tested in interviews.
Frequently asked questions
How do you prioritize product features?
Start from a clear goal, gather options written as problems, score them with a simple shared method such as RICE or value versus effort, debate the top items with design and engineering, then decide and explain what you chose not to do.
What is the best prioritization framework?
There is no single best one. RICE suits comparing many ideas with data, ICE suits fast triage, MoSCoW suits fixed-date releases, and value versus effort suits workshops. All beat having no method.
How do you say no to a stakeholder's feature request?
Acknowledge the problem behind it, show where it ranks against the current goal and what it would displace, offer a timeline or alternative, and record it on a visible not-now list.
How do bugs and technical debt fit into prioritization?
Reserve a steady share of capacity for quality work so it does not compete with features every week, and include bugs in the main ranking when they directly affect your goal metric.
Want this sharpened for your specific situation?
I take on a small number of career calls and resume reviews each month.
See how we could work together