Research Prioritize

Use this when the team has more research questions than capacity to answer them. Prevents "we should do more research" without focus. Produces a prioritized research roadmap: questions mapped to business decisions, scored on impact and uncertainty, sequenced with recommended methods.

Related skills: Complements /assumption-map (assumptions vs. research questions -- related but distinct lenses). Upstream: /discovery-questions generates questions. Downstream: prioritized questions feed into /interview-plan, /survey-design, /usability-test-plan, or /concept-test-plan depending on question type. Uses /research-repository to check what's already known.

Process

Step 1: Gather inputs

Ask the user to provide:

1. Open research questions -- what does the team want to learn? Collect from PMs, designers, engineers, leadership. Include questions from retros, strategy reviews, customer escalations, and roadmap debates.
2. Business context -- what decisions are coming up in the next 1-3 months? (Roadmap bets, pricing changes, new market entry, feature launches, platform migrations.)
3. Existing knowledge -- what do you already know? (Prior research, analytics, customer feedback. Reference /research-repository if one exists.)
4. Research capacity -- how many studies can you run per month? Who runs them? What methods are available?
5. Stakeholders -- who is asking these questions, and what will they do with the answers?

Step 2: Map questions to business decisions

For each research question, identify the business decision it informs. Questions that don't connect to a decision are interesting but not actionable.

## Research Prioritization -- (Team/product, date)

### Question-to-decision map

| # | Research question | Business decision it informs | Decision timeline | Stakeholder |
|---|---|---|---|---|
| 1 | (e.g., "Do enterprise users need SSO to convert?") | (Whether to invest in SSO integration for Q3) | (Q2 planning -- 6 weeks) | (Head of Product) |
| 2 | (e.g., "Why do users drop off after onboarding?") | (Whether to redesign onboarding or focus on activation) | (Next sprint planning -- 2 weeks) | (Growth PM) |
| 3 | (e.g., "How do power users organize their projects?") | (Information architecture for v2 redesign) | (Q3 -- 3 months) | (Design lead) |
| 4 | (e.g., "What's our competitive position on pricing?") | (Pricing model change for annual planning) | (Annual planning -- 4 months) | (CEO) |

Mapping rules:

• If a question doesn't connect to a decision, ask: "If we knew the answer, what would we do differently?" If the answer is "nothing," deprioritize it.
• Multiple questions can inform the same decision. Group them.
• Decision timeline matters: research that arrives after the decision is made is wasted.

Step 3: Score impact and uncertainty

Rate each question on two axes:

### Impact x Uncertainty matrix

| # | Research question | Impact if answered | Current uncertainty | Priority score |
|---|---|---|---|---|
| 1 | (Question) | High | High | **Research now** |
| 2 | (Question) | High | Medium | **Research soon** |
| 3 | (Question) | Medium | High | **Research when capacity allows** |
| 4 | (Question) | Low | High | **Skip -- low impact** |
| 5 | (Question) | High | Low | **Skip -- already know enough** |

### Scoring definitions

**Impact if answered (what changes if we know this?):**
- **High:** Directly changes a major product or business decision. Gets mentioned in leadership discussions.
- **Medium:** Informs a feature-level decision or validates a direction. Useful but not decisive.
- **Low:** Satisfies curiosity or provides background. No specific decision at stake.

**Current uncertainty (how much do we already know?):**
- **High:** No data, conflicting signals, or complete guesswork. We're making assumptions.
- **Medium:** Some signals (support tickets, analytics, a few interviews) but not confident enough to act.
- **Low:** Strong evidence from multiple sources. We know enough to decide.

Step 4: Sequence into a research roadmap

### Research roadmap

**Now (next 2-4 weeks)**

| Question | Recommended method | Why this method | Effort | Decision it unblocks |
|---|---|---|---|---|
| (Highest priority question) | (e.g., 6 user interviews) | (Need qualitative depth on a behavioral question) | (2 weeks) | (Decision and timeline) |
| (Second priority question) | (e.g., Analytics deep-dive) | (Quantitative answer exists in data, no need for primary research) | (3 days) | (Decision and timeline) |

**Next (4-8 weeks)**

| Question | Recommended method | Why this method | Effort | Decision it unblocks |
|---|---|---|---|---|
| (Question) | (Method) | (Rationale) | (Effort) | (Decision) |

**Later (8+ weeks)**

| Question | Recommended method | Why this method | Effort | Decision it unblocks |
|---|---|---|---|---|
| (Question) | (Method) | (Rationale) | (Effort) | (Decision) |

**Parked (not pursuing now)**

| Question | Why parked |
|---|---|
| (Question) | (Low impact / already know enough / decision timeline too far out) |

Sequencing rules:

• Sequence by decision timeline, not just priority score. A medium-priority question with a 2-week decision deadline beats a high-priority question with a 3-month horizon.
• Batch related questions into a single study when possible. (e.g., interview questions about onboarding and activation can be covered in the same sessions.)
• The "Now" bucket should contain no more than 2-3 studies. If it has more, you haven't prioritized hard enough.
• "Parked" is not "rejected." It means "not now." Review parked questions quarterly.

Step 5: Review and validate

Ask the user:

• Are the impact ratings honest? (Teams often rate everything as "high impact" to justify research they want to do.)
• Are decision timelines real? (If the decision will happen regardless of research, the research doesn't matter.)
• Is the "Now" bucket achievable with current capacity? (2-3 studies max in a 2-4 week window for most teams.)
• Who will communicate the roadmap to stakeholders? (Stakeholders who submitted questions need to know what's being prioritized and what's parked.)
• When will you re-prioritize? (Monthly is typical. Research questions evolve as decisions get made.)

Output location

Present the research roadmap as formatted text in the conversation. The user copies it into their team planning documents and shares with stakeholders.