Idea Validation — complete guide
How to use the validation system, what's in each of the 22 sections, how the verdict is calculated, and how findings flow into your project blueprint.
What Idea Validation does
Idea Validation gives you a 22-section, source-backed report on whether to build your SaaS idea — and what to do this week if the answer is yes. It runs across five phases (problem & customer, market, competition, economics, risk · legal · action) plus three cross-cutting outputs (verdict, roast, executive summary).
Two tiers run against the same intake:
- Quick Check — 5 sections + verdict, ~3 minutes, 100 credits, free-tier accessible.
- Full Report — all 22 sections + verdict + roast + executive summary, 12-15 minutes, 1200 credits.
Every new account gets 100 free credits on signup, which exactly covers a Quick Check. Use it to test the system on your real idea before deciding whether to upgrade to the Full Report.
How to run a validation
Paste your idea
On the dashboard, type or paste your idea into the validation textbox. Minimum 40 characters, plain language. No template, no Lean Canvas grid — just describe what you want to build and who it's for. The smart-intake analyser reads what you typed and figures out what context it still needs.
Answer 3-5 follow-ups
The intake asks only the questions whose answers it cannot infer from your idea. Common follow-ups: target user (industry, role, geography), founder advantage (what gives you an edge here), budget, timeline. You can skip any non-mandatory question. Answers are saved as you go — close the tab and resume later.
Click Run Quick Check
Quick Check charges 100 credits up front and starts streaming sections back. The first section completes in under 30 seconds; full Quick Check median is about 3 minutes. You can close the tab — generation continues in the background and an email lands in your inbox when it's done.
Review the verdict
The verdict card sits at the top of the report and updates live as more sections complete. It surfaces a one-of-four categorical decision (Build Now, Build With Pivot, Reconsider, Do Not Build), a 0-100 score, top three strengths, top three risks, and one concrete action for this week.
Decide: build, re-validate, or move on
If the verdict is Build Now or Build With Pivot, click Build this idea to enter the Phase-2 form (project name, platforms, tech stack, design direction). Promoting creates a Project with the validation snapshot frozen — every downstream generator reads it. If you want a fresh take, click Re-validate to spin up a new Idea version linked to the predecessor.
Phase 1 — Problem & Customer
Phase 1 answers: who feels the pain, how acute, and how they currently solve it. Five sections.
Problem validation
The foundational claim: the problem is real, severe, and frequent for the named segment. Includes severity score (0-10), frequency, evidence trail (URLs / interviews / industry knowledge), and a list of current workarounds the founder is competing against.
Customer personas
2-3 detailed personas, not template-stuffed marketing personas. Each has a snapshot, evidence-cited pain points, jobs-to-be-done in their voice, current workaround, buying triggers, channels they hang out on, and an anti-persona (who looks like them but isn't a fit).
Jobs to be done
The Bob Moesta / Christensen frame applied strictly. Each job uses the form "When [situation], I want to [motivation], so I can [outcome]" plus the four forces (push, pull, anxiety, habit) and the trigger event. Includes a Switch-interview script you can run as-is.
Founder–market fit
A 0-100 FMF score with evidence-cited strengths, honest gaps, and red flags. Compares you against the alternative founders this market would attract (a domain expert, a YC-backed team, an incumbent's PM). Lists what you'd have to change in the next 4-12 weeks to bump your score by 20 points.
Customer interview plan
Mom Test compliant interview script with 8-15 past-anchored questions. Banned phrasing ("would you", "do you think", "imagine if") is filtered out. Includes a recruiting brief, screener questions, decision rules (when to proceed / pivot / kill), and a note-taking template.
Phase 2 — Market
Phase 2 sizes the addressable market and pressure-tests the timing thesis. Four sections.
TAM / SAM / SOM
Bottom-up market sizing with a transparent calculation trail. Total addressable market, serviceable addressable market, and a realistic 3-year SOM the founder can credibly hit. CAGR estimate with source citation.
Why now
VC's #1 question. 2-5 specific, dated enabling shifts (technology cost curves, regulatory changes, behaviour shifts) that make this idea more attractive in 2026 than in 2020. Includes a "why not earlier" counterfactual and an "already happening risk" honest note.
Demand signals
Quantified evidence that real people are already trying to solve this — Google Trends, Reddit threads, GitHub stars, paid ads, job listings, funding rounds in adjacent categories. Composite read: pulling, stirring, or absent.
Analogies & post-mortems
2-3 success analogies + 2-3 failure post-mortems, all named companies with eras. Cross-cutting "patterns to copy" and "mistakes to avoid". Pattern read: favourable, mixed, or cautionary.
Phase 3 — Competition
Phase 3 maps where rivals win and where you can wedge in. Three sections.
Competitive landscape
Validated competitors with strengths, weaknesses, and the specific whitespace you can target. Cites evidence — public funding rounds, public reviews, product comparisons.
Switching costs
What it costs a buyer to leave their current solution for you (inbound friction), and what locks them in once they're with you (outbound moat). Each cost has a reduction lever you can pull.
Brand & naming check
Sanity-check across .com / .ai / .io domains, USPTO + EUIPO + IPO India trademark exposure, and X / Instagram / LinkedIn / GitHub handle availability. Every result includes a verifyAction so you can confirm in 30 seconds.
Phase 4 — Economics
Phase 4 answers: will the unit math actually leave you a business. Four sections.
Willingness to pay
Van Westendorp Price Sensitivity Meter with the four anchor questions per persona. Outputs an optimal price, an acceptable range, and a 3-tier pricing recommendation (starter / pro / enterprise). Marked as triangulated estimate — run the actual PSM survey before committing.
Unit economics
CAC, LTV, payback period, gross margin, monthly churn — with the formula trail, not just the numbers. Verdict: healthy, marginal, or broken. Top lever identifies the single biggest move to fix marginal/broken unit econ.
Distribution channels
Channel-by-channel viability across founder-led outbound, SEO content, paid ads, partnerships, community/PLG, integrations, and events. CAC estimate, time-to-first-result, and founder-effort per channel. Designates a primary channel to start with.
Content & SEO landscape
Target keyword opportunity map: 6-12 keywords across funnel intent, with estimated monthly search volume and SERP difficulty. Pillar topics + content gaps + a 90-day publishing plan. Marked as estimate — verify volumes with Ahrefs or Ubersuggest before committing to long-form content.
Phase 5 — Risk · Legal · Action
Phase 5 answers: what could kill it, and what to do this week. Six sections.
Legal & regulatory
Industry-aware compliance check. A separate intake call detects industry signals (fintech, health, edu, kids, AI, crypto) and loads only relevant rule sets. Government permissions, jurisdictional rules, costs, and timelines.
Founder-side legal
Incorporation jurisdiction recommendation, IP assignment, founder vesting, cap-table guidance, insurance needs by stage, tax setup. Always ends with the standard "directional, consult a startup attorney" disclaimer.
Tech-stack lock-in risk
Vendor / framework / model / infra lock-in assessment. Each dependency has an exit-pain description, an escape-hatch design, and timing on when to invest in the hatch.
Talent & hiring feasibility
Honest read on whether you can staff this product. Per-role comp ranges anchored to your geography, hiring difficulty, and concrete alternatives (contractor, agency, AI agent) when full-time isn't realistic.
Risk register
Synthesises the specific risks surfaced across all 21 prior sections into one register. Each risk has probability, impact, severity, mitigation, contingency, and the source section that flagged it (traceability).
Validation experiments
3-10 concrete, time-bounded, success-metric'd experiments to run BEFORE writing meaningful code. Each de-risks one or more registered risks. Becomes "Phase 0" in your project's PhasePlan when promoted.
Cross-cutting outputs
On top of the 22 phase sections, every Full Report also generates three synthesis outputs.
Verdict
A categorical decision aid: Build Now, Build With Pivot, Reconsider, or Do Not Build. Plus a 0-100 score, confidence level, top three strengths, top three risks, and one concrete action for this week. Calibrated against a target distribution (~30% Build Now, ~30% Build With Pivot, ~25% Reconsider, ~15% Do Not Build) so it pushes back on weak ideas.
Honest roast
A brutal-but-kind 4-7 sentence narrative naming the specific weakness you're underweighting. Not coach-speak, not catastrophising — the kind of feedback a senior peer gives at 11pm. Includes a "kind alternative" reframing for when you need to share it with co-founders.
Executive summary
A 4-paragraph institutional narrative (problem + market, product + edge, economics + GTM, risks + next) plus an elevator pitch and structured key metrics. Investor-/board-/cofounder-ready — paste it directly into a deck appendix.
How the verdict is calculated
The four buckets
- Build Now — strong evidence across all sections (problem real, market sized, whitespace clear, regulatory manageable). Score typically 75-100.
- Build With Pivot — core opportunity exists but needs adjustment. The specific pivot is named (not 'narrow your focus' but 'focus on Indian fintech founders, not all SaaS founders'). Score typically 55-75.
- Reconsider — significant red flags. Specific research or angle suggested before committing. Score typically 30-55.
- Do Not Build — fundamental issue (no market, illegal, competing with giants with no edge, founder-market mismatch obvious). Suggests 2-3 better adjacent ideas. Score typically 0-35.
The distribution target
The verdict prompt is calibrated against a target distribution across reports: roughly 30% Build Now, 30% Build With Pivot, 25% Reconsider, 15% Do Not Build. If we drifted heavily toward Build Now on every idea, the report would be flattery. The roast section adds pressure in the same direction.
The admin dashboard at /admin/idea-validation tracks the live distribution against this target. Drift greater than ±15 points triggers a prompt-rot investigation.
What if you disagree with the verdict?
Re-validate. Click the Re-validate button to spin up a fresh Idea row in the version chain — your old report stays read-only as historical record. If your intake answers were thin, fix them and rerun. The Quick Check is 100 credits; a Full Report is 1200 credits.
Sources & citation system
Every numerical claim across the 22 sections links to a public URL when one exists, or is explicitly marked with one of three labels: [founder-supplied], [industry-knowledge], or [bottom-up-calculation]. No hand-waving.
Sources are stored separately from sections, so we can de-duplicate them across your report, recheck their liveness daily (a HEAD-check cron flags broken links), and surface that status in the UI. If a URL goes 404, you'll see it on the source citation in the report.
From validation to project blueprint
When you click Build this idea, the validation findings are frozen as a snapshot on your new Project (validationContextJson). Five downstream generators read that snapshot:
- Architecture — sizes components for your validated SOM, bakes compliance scaffolding from your validated industries, cites validation findings in node descriptions.
- Features — marks priority against the bar of your validated risks. 'Critical' = directly addresses primary problem or top-risk mitigation. Persona-aware actors.
- Phases — inserts a 'Phase 0: Validation Experiments' built from your validated experiments. Build phases ordered by RISK descent.
- Pricing (in product definition) — monetization anchored to your validated WTP optimal, not invented. Tiers bracket the WTP range.
- Outreach — primary channel surfaced from validated distribution analysis.
The snapshot is FROZEN — it doesn't update if you re-validate the source Idea. This is intentional: re-running Architecture later should produce consistent output, not silently change because validation moved. To bring fresh findings into a project, you can re-validate then re-promote (or manually refresh the snapshot in a future release).
Sharing and exporting your report
Public share link
Click Share on the report to mint a public URL at /r/[token]. Anyone with the link can view the full report. Optionally add a bcrypt-hashed password to gate access to specific people. Rotate the URL anytime to invalidate prior links.
PDF export
Click PDF on the report to download an A4-formatted, page-break-friendly PDF of all sections plus verdict, roast, and executive summary. Generated server-side via headless Chromium for paper-grade output. Takes 15-25 seconds — the button shows a spinner during generation.
Cost & refund policy
Two layers of cost protection on every run:
- Per-run cost ceiling — Quick Check capped at $1, Full Report capped at $6. If a runaway prompt hits the ceiling, the orchestrator stops and refunds your credits in full.
- Zero-success refund — if no sections completed (orchestrator hard fail), all charged credits are automatically refunded and the Idea status returns to intake_complete so you can retry without re-paying.
Partial-success runs (some sections complete, others fail) keep the charge — you got value from what completed. Failed sections show their error reason in the UI so you can decide whether to re-run individually.
When to use Quick Check vs Full Report
Use Quick Check when…
You have multiple ideas to triage; you're testing the system on your real idea before committing more credits; you want a fast read on whether to proceed to deeper research; or you're a free-tier user.
Upgrade to Full Report when…
Quick Check returned Build Now or Build With Pivot and you want grounded depth before writing code; you need to share with co-founders / advisors / investors; you want the validation experiments and risk register to seed your phase plan; or you want the executive summary and PDF for fundraising.