Startup Ideation
Scope
Covers
- Turning vague “startup ideas” into structured opportunity theses
- Expanding your information diet to find off-the-beaten-path opportunities
- Running a Why Now analysis based on technology + behavior + distribution shifts
- Identifying tarpits (ideas that look good but are structurally hard) and pruning early
- Scoring ideas and producing a top-idea 1‑pager + 2‑week validation plan
When to use
- “Help me come up with startup ideas in/around.”
- “We have 5 ideas — help us pick one and explain why.”
- “What’s a good *Why Now* for this idea?”
- “Pressure to do AI — where are real new opportunities?”
- “How do we avoid idea tarpits and pick something differentiated?”
When NOT to use
- You already chose an idea and need a delivery-ready PRD (use
writing-prds) - You need to define the problem space for a specific user pain (use
problem-definition) - You need to execute research (recruit, interview, synthesize) rather than frame it (use
conducting-user-interviews) - You need market sizing / pricing / fundraising pitch materials (adjacent work, not covered here)
Inputs
Minimum required
- Founder/team context + constraints (time, budget, skills, regulatory constraints)
- The decision to make + timeline (e.g., “pick 1 idea to validate in the next 2 weeks”)
- Target customer type (B2B/B2C; any preferred industries or segments)
- Any starting ideas (even rough) + what prompted them
Missing-info strategy
- Ask up to 5 questions from references/INTAKE.md.
- If still missing, proceed with explicit assumptions and list Open questions that could change the recommendation.
Outputs (deliverables)
Produce a Startup Ideation Pack in Markdown (in-chat; or as files if the user requests):
- Context snapshot (goal, constraints, decision, timeline)
- Unfair advantage + off-the-beaten-path signals (what you know/see that others might not)
- Shift scan + Why Now candidates (tech/behavior/distribution/regulatory shifts)
- Opportunity theses table (15–30 ideas, each structured + testable)
- Tarpit & differentiation check (prune to a shortlist)
- Idea scorecard (score top 3–5 with evidence)
- Top idea brief (1‑pager) (clear wedge + Why Now + ICP + risks)
- 2‑week validation plan (fastest tests for the highest-risk assumptions)
- Risks / Open questions / Next steps (always included)
Templates: references/TEMPLATES.md Expanded guidance: references/WORKFLOW.md
Workflow (8 steps)
1) Intake + decision framing
- Inputs: User context; references/INTAKE.md.
- Actions: Clarify decision, time horizon, and constraints. Define success as “pick 1 idea to validate next” (or similar).
- Outputs: Context snapshot.
- Checks: You can restate the decision in one sentence (“We are deciding whether to… by ”).
2) Inventory unfair advantage + off-the-beaten-path signals
- Inputs: Founder/team background; past work; lived experience; access.
- Actions: List 5–15 unique signals: personal pain, workflows you’ve seen, niche communities, privileged distribution, proprietary data access, or operator insight.
- Outputs: Unfair advantage + signals list.
- Checks: Each signal is specific (who/where/when) and could plausibly lead to a differentiated idea.
3) Run a shift scan (“Why now?” raw material)
- Inputs: Domain + constraints; current trends the user cares about.
- Actions: Generate 10–20 “shifts” across: technology capability, buyer behavior, regulation, distribution, and cost curves. For each, write: “This enables X that was hard before.”
- Outputs: Shift scan + Why Now candidates.
- Checks: At least 5 shifts are concrete and falsifiable (not vague hype).
4) Generate opportunity theses (structured ideas)
- Inputs: Signals + shifts.
- Actions: Produce 15–30 opportunity theses using the template: *Customer → Job → Pain → Why now → Wedge → First test*.
- Outputs: Opportunity theses table.
- Checks: Every idea includes a Why Now statement and a proposed first validation test.
5) Tarpit & differentiation check (prune)
- Inputs: Opportunity theses table.
- Actions: Flag tarpits and thinly differentiated ideas. Apply “off-the-beaten-path” pressure: if an idea is widely discussed, require a strong wedge or discard.
- Outputs: Pruned list + notes on tarpits/differentiation.
- Checks: The remaining shortlist has at least one concrete advantage (distribution, insight, data, speed, regulatory, workflow depth).
6) Score + shortlist top 3–5
- Inputs: Pruned list; references/RUBRIC.md.
- Actions: Score each shortlisted idea with evidence and assumptions. Highlight the 1–2 criteria that dominate the outcome (sensitivity).
- Outputs: Idea scorecard + top 3–5 recommendation.
- Checks: Scores cite specific evidence or clearly labeled assumptions (no hand-wavy numbers).
7) Draft the top idea 1‑pager + 2‑week validation plan
- Inputs: Top idea; references/TEMPLATES.md.
- Actions: Write a crisp 1‑pager (ICP, problem, Why Now, wedge, GTM motion hypothesis). Then design the fastest validation plan focused on the riskiest assumptions.
- Outputs: Top idea brief + validation plan.
- Checks: The plan includes: who to talk to, what to build (if anything), success criteria, and a stop/pivot rule.
8) Quality gate + finalize pack
- Inputs: Full draft pack.
- Actions: Run references/CHECKLISTS.md and score with references/RUBRIC.md. Add Risks / Open questions / Next steps.
- Outputs: Final Startup Ideation Pack.
- Checks: A stakeholder can review async and decide “validate / park / discard” without a meeting.
Quality gate (required)
- Use references/CHECKLISTS.md and references/RUBRIC.md.
- Always include: Risks, Open questions, Next steps.
Examples
Example 1 (B2B): “We’re ex‑operators in logistics. Generate and score startup ideas; pick 1 to validate in 2 weeks.” Expected: opportunity theses rooted in real workflows + a shortlist + a top idea 1‑pager with a concrete validation plan.
Example 2 (AI shift): “We think new LLM capabilities enable something new in customer support; help us find a differentiated idea and Why Now.” Expected: shift scan → structured theses → tarpit check → top idea brief with a tight wedge and clear risks.
Boundary example: “Give me 100 startup ideas with no context.” Response: ask intake questions first; if the user won’t provide any, produce a small set of generic theses with explicit assumptions and advise on how to ground them in real signals.