ChatGPT prompt for a feature prioritization tool

Act as a product manager facilitating an evidence-based feature prioritization review. Prioritize the candidate features below against [PRODUCT GOAL] for [TARGET USERS] during [PLANNING HORIZON].

Inputs
- Candidate features: [FEATURE BACKLOG]
- User evidence: [USER EVIDENCE]
- Business and technical constraints: [CONSTRAINTS]
- Delivery estimates and dependencies: [ESTIMATES AND DEPENDENCIES]
- Current metrics and baseline: [CURRENT METRICS]

First, normalize each feature into a problem statement, intended user, desired behavior change, and measurable outcome. Do not treat a requested solution as proof of a problem. Score each item from 1–5 for: evidence strength, affected-user reach, expected impact on the stated goal, confidence in the impact estimate, effort, strategic fit, and risk. Define each score briefly using only my inputs. If data is insufficient, mark the score “unknown” rather than guessing.

Produce:
1. A ranked table with feature, user problem, scores, dependencies, key assumption, and recommendation: now, next, later, discover, or do not pursue.
2. A short explanation of the top three recommendations, including the tradeoff made against lower-ranked work.
3. A 90-day sequence that identifies discovery, design, build, launch, and measurement milestones; include a named role for each owner, not invented people.
4. For every “now” item, write a testable outcome statement, success metric, guardrail metric, and smallest experiment or release that could reduce uncertainty.
5. A decision log listing the material assumptions and evidence gaps.

Do not add estimated revenue, user counts, effort, legal requirements, or technical capabilities that I did not provide. Before answering, verify that no high-effort item ranks ahead of a lower-effort alternative without a stated reason, and flag features with dependencies that make the proposed timing unrealistic. Ask up to three clarifying questions only if a required input is missing.

Build a RICE prioritization table for [FEATURES] against [GOAL]. Use the supplied reach, impact, confidence, and effort data in [INPUT DATA]; do not invent missing values. Calculate each score, show the formula inputs, and rank the items. Add a column explaining whether the metric evidence in [EVIDENCE] supports the impact score. Flag false precision, shared dependencies, and any feature whose score hides a material risk. End with the two assumptions most worth validating before committing.

Using [RESEARCH NOTES], create an opportunity solution tree for the outcome [OUTCOME] and users [USERS]. Separate observations, opportunity statements, solution ideas, and experiments. Include only opportunities grounded in the notes, group duplicates, and write each as an unmet need rather than a feature request. For the three strongest opportunities, suggest one low-cost experiment with a success signal and a disconfirming signal. Do not imply demand sizes or causation not present in the research.

Write a product-prioritization decision memo for [AUDIENCE] using [BACKLOG], [EVIDENCE], [CAPACITY], and [GOAL]. Limit it to 600 words. Include the decision, the selected work, explicitly deferred work, the tradeoffs, dependencies, risks, and success and guardrail metrics. Use a neutral tone and distinguish facts from assumptions. Flag any missing evidence that could reverse the decision instead of inventing a justification.