| name | business-idea-evaluation |
| description | Evaluates business ideas using Lean Canvas, unit economics, moat analysis, and structured risk assessment to produce a Go/No-Go recommendation. Trigger phrases: "evaluate this business idea", "is this idea viable", "should I build this", "business model analysis", "startup idea feedback", "lean canvas for". Do NOT use for existing business operations, financial accounting, marketing execution, or competitive intelligence on established companies — prefer `competitor-teardown` or `market-research` for those.
|
| license | Apache-2.0 |
| compatibility | {"clients":["openai-codex","gemini-cli","opencode","github-copilot"]} |
| metadata | {"owner":"codex","domain":"business-idea-evaluation","maturity":"draft","risk":"low","tags":["lean-canvas","unit-economics","startup","business-model","moat","risk-assessment"]} |
Purpose
Systematically evaluate a business idea's viability using the Lean Canvas framework, unit economics estimation, competitive moat analysis, and multi-dimensional risk scoring to produce an evidence-based Go/No-Go recommendation with clear next steps.
When to use this skill
- User presents a business idea and wants structured feedback on viability
- User asks whether they should build or invest in a product/service concept
- User needs a Lean Canvas, business model analysis, or unit economics estimate
- User wants to compare two or more business ideas against each other
- A project is at the "should we build this?" stage and needs a decision framework
Do not use this skill when
- The business already exists and the question is about optimizing operations or growth — that is execution, not evaluation
- The user needs competitive intelligence on specific companies — prefer
competitor-teardown
- The request is about market sizing or industry trend analysis without a specific idea — prefer
market-research
- The user needs financial modeling, bookkeeping, or tax analysis — prefer
financial-tracker-ops
- The user wants to write a full business plan document (this skill produces an evaluation, not a plan)
Operating procedure
Phase 1 — Idea intake and clarification
- Extract or ask for the core elements of the idea:
- What is the product/service? One-sentence description.
- Who is the customer? Specific segment, not "everyone".
- What problem does it solve? Existing alternatives the customer currently uses.
- How does it make money? Revenue model (subscription, transaction, advertising, licensing, etc.).
- What stage is this at? Napkin idea, validated problem, prototype, or early revenue.
- If the user cannot articulate the customer or problem, flag this as a critical gap before proceeding.
Phase 2 — Lean Canvas construction
-
Build a complete Lean Canvas (Ash Maurya's adaptation of Business Model Canvas):
| Block | Question to answer |
|---|
| Problem | Top 3 problems for the target customer. List existing alternatives. |
| Customer Segments | Who specifically has this problem? Early adopters vs mainstream. |
| Unique Value Proposition | Single clear sentence: why is this different AND worth paying for? |
| Solution | Top 3 features that address the top 3 problems. |
| Channels | How will you reach customers? (Content, paid ads, partnerships, sales, virality) |
| Revenue Streams | Pricing model, price point, who pays. |
| Cost Structure | Fixed costs, variable costs, key cost drivers. |
| Key Metrics | The 3-5 numbers that determine if this business is healthy. |
| Unfair Advantage | What cannot be easily copied or bought? (Be honest — most ideas have none yet.) |
-
Flag any canvas block where the answer is vague, assumed, or "TBD" — these are validation priorities.
Phase 3 — Unit economics estimation
- Estimate the following metrics (use ranges if exact numbers are unavailable):
- Customer Acquisition Cost (CAC): Estimated marketing + sales cost to acquire one paying customer. Break down by channel if possible.
- Lifetime Value (LTV): Average revenue per customer × gross margin × average customer lifespan. For subscription: (ARPU × gross margin) / monthly churn rate.
- LTV:CAC ratio: Target is ≥3:1 for a sustainable business. Flag if <2:1.
- Payback period: Months to recover CAC from a single customer. Target <12 months for bootstrapped, <18 months for funded.
- Gross margin: Revenue minus direct costs of delivering the product/service. Software should target >70%; services typically 40-60%.
- State all assumptions explicitly. Mark each estimate with a confidence level: High (based on data or close comparables), Medium (reasonable inference), or Low (educated guess).
Phase 4 — Moat analysis
- Assess each potential moat type and rate as None / Weak / Moderate / Strong:
- Network effects: Does the product become more valuable as more people use it? (Direct: social networks. Indirect: marketplaces.)
- Switching costs: How painful is it for a customer to leave? (Data lock-in, workflow integration, retraining cost.)
- Scale economies: Does unit cost decrease meaningfully with volume? (Server costs, content amortization, supplier leverage.)
- Brand: Is there a trust or recognition advantage that takes years to build? (Typically irrelevant for new ideas.)
- Regulatory/legal barriers: Licenses, patents, regulatory approvals that block competitors. (Don't confuse regulation with moat — regulation can also block you.)
- Proprietary technology/data: Unique algorithms, datasets, or IP that competitors cannot replicate easily.
- If no moat exists today, assess whether one can be built over time and what it would take.
Phase 5 — Risk matrix
-
Score each risk dimension 1-5 (1 = low risk, 5 = critical risk):
| Risk type | What it means | Key questions |
|---|
| Market risk | Will anyone pay for this? | Is the problem validated? Is willingness to pay confirmed? |
| Execution risk | Can the team build and deliver it? | Does the team have the skills? What are the hard technical challenges? |
| Technology risk | Can it be built at all? | Are there unsolved technical problems? Dependency on immature tech? |
| Regulatory risk | Could regulations block or constrain this? | Industry-specific compliance, data privacy, licensing requirements? |
| Competitive risk | Can incumbents or well-funded startups crush this? | How fast can a competitor replicate this? Are there funded players? |
| Financial risk | Can this be funded to sustainability? | Runway requirements, capital intensity, path to profitability? |
-
For any dimension scored ≥4, write a specific mitigation strategy or acknowledge it as an unresolved blocker.
Phase 6 — Comparable analysis
- Identify 3-5 comparable companies or products:
- Direct comparables: Companies that tried to solve the same problem for the same customer.
- Adjacent comparables: Companies in adjacent markets or solving analogous problems for different customers.
- Cautionary comparables: Companies that tried something similar and failed — analyze why.
- For each comparable, note: outcome (success/failure/pivot/acquired), business model, funding raised, key differentiator from the idea being evaluated.
Phase 7 — Go/No-Go scoring
-
Score the idea across 8 weighted dimensions:
| Dimension | Weight | Score (1-5) |
|---|
| Problem severity & frequency | 20% | |
| Market size & accessibility | 15% | |
| Unit economics viability | 20% | |
| Moat potential | 10% | |
| Execution feasibility | 15% | |
| Competitive landscape | 10% | |
| Risk profile | 5% | |
| Founder-market fit | 5% | |
-
Compute weighted score. Interpretation:
- 4.0-5.0: Strong Go — pursue aggressively, focus on execution.
- 3.0-3.9: Conditional Go — viable but has specific risks to mitigate first.
- 2.0-2.9: Weak — significant concerns. Recommend pivoting or shelving unless specific conditions change.
- 1.0-1.9: No-Go — fundamental viability problems.
-
A single dimension scored at 1 is an automatic flag regardless of composite score.
Decision rules
- Problem before solution: If the user leads with a solution and cannot articulate the problem, force problem articulation first. Solution-first ideas fail at higher rates.
- No assumed virality: If the growth model depends on virality, require evidence (comparable viral coefficients, inherent shareability mechanics). "People will share it because it's cool" is not a growth strategy.
- Customer validation beats logic: A logically sound idea with zero customer evidence scores lower than a messy idea with paying customers or strong intent signals.
- Honest moat assessment: Most early-stage ideas have no moat. Saying "no moat yet, but here's how to build one" is more useful than inventing a moat that doesn't exist.
- Revenue model required: If the user says "we'll figure out monetization later," flag this as a critical gap. Evaluate based on the most likely revenue model, but note the uncertainty.
- Beware of TAM fantasies: "If we capture just 1% of a $50B market" is not a market sizing strategy. Demand bottom-up sizing: specific customers × realistic price × achievable penetration.
Output requirements
Deliver all of the following as separate, clearly labeled sections:
- Idea Summary — Restated idea in one paragraph, confirming understanding.
- Lean Canvas — Complete 9-block canvas with flagged gaps.
- Unit Economics Estimate — CAC, LTV, LTV:CAC, payback period, gross margin with assumptions and confidence levels.
- Moat Assessment — Rating for each moat type with justification.
- Risk Matrix — Scored risk table with mitigation strategies for high-risk dimensions.
- Comparable Analysis — 3-5 comparables with outcomes and lessons.
- Go/No-Go Score — Weighted score table, composite score, interpretation, and automatic flags.
- Next Steps — 3-5 specific, actionable next steps ranked by impact and urgency. If Go: what to validate first. If No-Go: what would need to change.
Anti-patterns
- Solution-first thinking: Building a product then searching for a problem. The evaluation must start with the problem and customer, not the technology or feature set.
- Ignoring unit economics: "We'll make it up in volume" is a red flag, not a strategy. If the unit economics don't work at small scale, they rarely work at large scale either.
- Assuming virality: Baking viral growth into the base case without evidence. Viral coefficients >1.0 are extraordinarily rare. Treat virality as upside, not baseline.
- No customer validation: An idea that has never been discussed with a real potential customer is an untested hypothesis, not a business. Score it accordingly.
- TAM-down market sizing: Starting with a massive market and assuming a tiny slice. Bottom-up sizing (specific reachable customers × realistic conversion) is the only credible approach.
- Ignoring competitor dynamics: "No one is doing this" usually means either the market doesn't exist or you haven't looked hard enough. Investigate existing alternatives thoroughly.
- Conflating idea quality with execution ability: A great idea executed poorly fails. Assess the team's ability to execute, not just the idea's theoretical merit.
- Single-scenario planning: Presenting only the optimistic case. Always include a realistic case and a pessimistic case for key metrics.
Related skills
competitor-teardown — Deep-dive competitive analysis for companies identified in the comparable analysis
market-research — Broader market sizing and trend analysis to contextualize the opportunity
domain-scouting — If the idea passes evaluation, find a name and domain for it
financial-tracker-ops — Ongoing financial tracking once the idea moves to execution
Failure handling
- If the user provides only a one-line idea with no context, ask clarifying questions (customer, problem, revenue model) before attempting evaluation. Garbage in, garbage out.
- If unit economics cannot be estimated even roughly (no pricing model, no comparable data), mark the Unit Economics section as "Not estimable — insufficient data" and explain what information is needed.
- If the idea is in a regulated industry the agent has no knowledge of, flag regulatory risk as "Unassessed — specialist input required" rather than guessing.
- If the user asks to evaluate more than 3 ideas at once, recommend evaluating them sequentially with full rigor rather than doing shallow passes on all of them.