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.
1---2name: startup-ideation3description: Generate and evaluate startup ideas using off-the-beaten-path insights + Why-Now shift analysis. Produces a Startup Ideation Pack (opportunity theses table, scorecard, top idea brief, validation plan). Use for startup ideation, idea selection, “what should we build?”, why now, tarpit avoidance, information diet planning.4---5
6# Startup Ideation
7
8## Scope
9
10**Covers**
11- Turning vague “startup ideas” into structured **opportunity theses**
12- Expanding your **information diet** to find off-the-beaten-path opportunities
13- Running a **Why Now** analysis based on technology + behavior + distribution shifts
14- Identifying **tarpits** (ideas that look good but are structurally hard) and pruning early
15- Scoring ideas and producing a **top-idea 1‑pager** + **2‑week validation plan**
16
17**When to use**
18- “Help me come up with startup ideas in/around <domain>.”
19- “We have 5 ideas — help us pick one and explain why.”
20- “What’s a good *Why Now* for this idea?”
21- “Pressure to do AI — where are real new opportunities?”
22- “How do we avoid idea tarpits and pick something differentiated?”
23
24**When NOT to use**
25- You already chose an idea and need a delivery-ready PRD (use `writing-prds`)
26- You need to define the problem space for a specific user pain (use `problem-definition`)
27- You need to execute research (recruit, interview, synthesize) rather than frame it (use `conducting-user-interviews`)
28- You need market sizing / pricing / fundraising pitch materials (adjacent work, not covered here)
29
30## Inputs
31
32**Minimum required**
33- Founder/team context + constraints (time, budget, skills, regulatory constraints)
34- The decision to make + timeline (e.g., “pick 1 idea to validate in the next 2 weeks”)
35- Target customer type (B2B/B2C; any preferred industries or segments)
36- Any starting ideas (even rough) + what prompted them
37
38**Missing-info strategy**
39- Ask up to 5 questions from [references/INTAKE.md](references/INTAKE.md).
40- If still missing, proceed with explicit assumptions and list **Open questions** that could change the recommendation.
41
42## Outputs (deliverables)
43
44Produce a **Startup Ideation Pack** in Markdown (in-chat; or as files if the user requests):
45
461) **Context snapshot** (goal, constraints, decision, timeline)
472) **Unfair advantage + off-the-beaten-path signals** (what you know/see that others might not)
483) **Shift scan + Why Now candidates** (tech/behavior/distribution/regulatory shifts)
494) **Opportunity theses table** (15–30 ideas, each structured + testable)
505) **Tarpit & differentiation check** (prune to a shortlist)
516) **Idea scorecard** (score top 3–5 with evidence)
527) **Top idea brief (1‑pager)** (clear wedge + Why Now + ICP + risks)
538) **2‑week validation plan** (fastest tests for the highest-risk assumptions)
549) **Risks / Open questions / Next steps** (always included)
55
56Templates: [references/TEMPLATES.md](references/TEMPLATES.md)
57Expanded guidance: [references/WORKFLOW.md](references/WORKFLOW.md)
58
59## Workflow (8 steps)
60
61### 1) Intake + decision framing
62- **Inputs:** User context; [references/INTAKE.md](references/INTAKE.md).
63- **Actions:** Clarify decision, time horizon, and constraints. Define success as “pick 1 idea to validate next” (or similar).
64- **Outputs:** Context snapshot.
65- **Checks:** You can restate the decision in one sentence (“We are deciding whether to… by <date>”).
66
67### 2) Inventory unfair advantage + off-the-beaten-path signals
68- **Inputs:** Founder/team background; past work; lived experience; access.
69- **Actions:** List 5–15 unique signals: personal pain, workflows you’ve seen, niche communities, privileged distribution, proprietary data access, or operator insight.
70- **Outputs:** Unfair advantage + signals list.
71- **Checks:** Each signal is specific (who/where/when) and could plausibly lead to a differentiated idea.
72
73### 3) Run a shift scan (“Why now?” raw material)
74- **Inputs:** Domain + constraints; current trends the user cares about.
75- **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.”
76- **Outputs:** Shift scan + Why Now candidates.
77- **Checks:** At least 5 shifts are concrete and falsifiable (not vague hype).
78
79### 4) Generate opportunity theses (structured ideas)
80- **Inputs:** Signals + shifts.
81- **Actions:** Produce 15–30 opportunity theses using the template: *Customer → Job → Pain → Why now → Wedge → First test*.
82- **Outputs:** Opportunity theses table.
83- **Checks:** Every idea includes a Why Now statement and a proposed first validation test.
84
85### 5) Tarpit & differentiation check (prune)
86- **Inputs:** Opportunity theses table.
87- **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.
88- **Outputs:** Pruned list + notes on tarpits/differentiation.
89- **Checks:** The remaining shortlist has at least one concrete advantage (distribution, insight, data, speed, regulatory, workflow depth).
90
91### 6) Score + shortlist top 3–5
92- **Inputs:** Pruned list; [references/RUBRIC.md](references/RUBRIC.md).
93- **Actions:** Score each shortlisted idea with evidence and assumptions. Highlight the 1–2 criteria that dominate the outcome (sensitivity).
94- **Outputs:** Idea scorecard + top 3–5 recommendation.
95- **Checks:** Scores cite specific evidence or clearly labeled assumptions (no hand-wavy numbers).
96
97### 7) Draft the top idea 1‑pager + 2‑week validation plan
98- **Inputs:** Top idea; [references/TEMPLATES.md](references/TEMPLATES.md).
99- **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.
100- **Outputs:** Top idea brief + validation plan.
101- **Checks:** The plan includes: who to talk to, what to build (if anything), success criteria, and a stop/pivot rule.
102
103### 8) Quality gate + finalize pack
104- **Inputs:** Full draft pack.
105- **Actions:** Run [references/CHECKLISTS.md](references/CHECKLISTS.md) and score with [references/RUBRIC.md](references/RUBRIC.md). Add **Risks / Open questions / Next steps**.
106- **Outputs:** Final Startup Ideation Pack.
107- **Checks:** A stakeholder can review async and decide “validate / park / discard” without a meeting.
108
109## Quality gate (required)
110- Use [references/CHECKLISTS.md](references/CHECKLISTS.md) and [references/RUBRIC.md](references/RUBRIC.md).
111- Always include: **Risks**, **Open questions**, **Next steps**.
112
113## Examples
114
115**Example 1 (B2B):** “We’re ex‑operators in logistics. Generate and score startup ideas; pick 1 to validate in 2 weeks.”
116Expected: opportunity theses rooted in real workflows + a shortlist + a top idea 1‑pager with a concrete validation plan.
117
118**Example 2 (AI shift):** “We think new LLM capabilities enable something new in customer support; help us find a differentiated idea and Why Now.”
119Expected: shift scan → structured theses → tarpit check → top idea brief with a tight wedge and clear risks.
120
121**Boundary example:** “Give me 100 startup ideas with no context.”
122Response: 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.
123
majiayu000