Prioritize Your Feature Backlog with RICE Scoring and Flagged Weak Assumptions

Get a ranked RICE scoring table for your pasted backlog, with the math shown, weak assumptions flagged, and a short list of what to build next.

You getScoring table, top 3 to build next and weak assumptions to validate

Prompt

Prompt
You are a senior product manager who runs quarterly planning with RICE scoring.

Job: prioritize the feature backlog I will paste right after this prompt. Before scoring, ask me these questions in one numbered message and wait for my answers:
1. How many users or accounts will each feature reach per quarter, and where does each number come from (analytics, survey, guess)?
2. How many person-months will each feature take, and who estimated it?
3. What is this quarter's main goal (e.g., activation, retention, revenue) for judging Impact?
4. Any deadlines, dependencies or commitments that should override the score?

Rules:
- Use standard scales: Impact 3, 2, 1, 0.5, 0.25; Confidence 100%, 80%, 50%. RICE = Reach x Impact x Confidence / Effort.
- Use only my Reach and Effort numbers. If one is missing, mark the feature "missing" and leave it unranked.
- Flag with ⚠ any score resting on a guess, a single person's estimate, or an Impact you inferred, plus a one-line reason.
- Show the calculation for every score so I can check it.

Output format:
1. Scoring table: Rank | Feature | Reach | Impact | Confidence | Effort | Calculation | RICE | Weak assumption
2. Top 3 to build next, one sentence each on why.
3. Weak assumptions to validate, with the cheapest way to check each.
4. Overrides from deadlines or dependencies.

End with one line naming the biggest risk in this ranking.

Example output

At a glanceRICE scores by featureOffline mobile mode unranked until effort is estimated
  • Slack notifications: 2,400
  • Onboarding checklist: 1,667
  • Bulk CSV import: 960
  • Dark mode: 533
  • Custom dashboards: 200
  • SAML SSO: 150

Example for: a B2B project-management SaaS, 7-item backlog, quarter goal is new-user activation.

1. Scoring table
RankFeatureReachImpactConfidenceEffortCalculationRICEWeak assumption
1Slack notifications3,000180%13000x1x0.8/12,400None
2Onboarding checklist2,500250%1.52500x2x0.5/1.51,667⚠ Impact inferred; no test data on checklists
3Bulk CSV import1,200280%21200x2x0.8/2960None
4Dark mode4,0000.2580%1.54000x0.25x0.8/1.5533None
5Custom dashboards800250%4800x2x0.5/4200⚠ Reach is a guess from 5 sales calls; effort from one engineer
6SAML SSO1503100%3150x3x1/3150None
–Offline mobile mode1,800––missing–missingEffort not estimated
2. Top 3 to build next
  1. Slack notifications – highest score, solid usage data, and only one person-month.
  2. Onboarding checklist – directly targets activation, but validate Impact before committing the full 1.5 months.
  3. Bulk CSV import – removes the most-cited setup blocker in trial exit surveys.
3. Weak assumptions to validate
  • Onboarding checklist Impact (2): ship a static checklist to 10% of new signups for two weeks and compare day-7 activation.
  • Custom dashboards Reach (800): add a "request dashboards" button in the reports page and count clicks for 14 days.
  • Custom dashboards Effort (4): ask a second engineer for an independent estimate.
4. Overrides
  • SAML SSO ranks 6th, but the Northfield account (renewal due in 10 weeks) requires it. Schedule it now despite the score.
  • Offline mobile mode cannot be ranked until engineering estimates effort.

Biggest risk: the #2 item rests on an inferred Impact; if the checklist test shows no lift, CSV import moves up.

Written by AI from this prompt. Your results will vary.

Tips and common edits

Why this works

  • Asking for reach and effort sources first keeps the model from inventing numbers and exposes which inputs are guesses.
  • Showing the calculation in each row lets you audit every score in seconds.
  • Flagging weak assumptions with a cheap validation step turns the ranking into an action plan, not a false certainty.

Use before quarterly or sprint planning. Paste your backlog list into ChatGPT, Claude or Gemini, then have rough reach numbers, effort estimates and your quarter's goal ready to answer.

Common edits

  • Swap RICE for ICE scoring if you lack reliable reach data.
  • Ask for the table as CSV so you can paste it into a spreadsheet or Jira.

Prompt details

By
Prompts-GPT editors
Published

Did this prompt work?

Keep exploring

More in Product Management