Key Principle
A venture plan is built on assumptions derived from logic and research, not proven facts. Until tested through observed customer actions -- not words -- they remain leaps of faith. Steps 20-23 form the framework's culminating feedback loop: identify key assumptions, test each empirically, bundle validated assumptions into a Minimum Viable Business Product, then prove with quantitative data that customers adopt, pay for, and advocate for it. This sequence prevents the single most expensive category of startup failure: mistaking logical sufficiency for market acceptance.
Why This Matters
The sequencing is deliberate and cannot be collapsed. Step 20 (identify assumptions) ensures that Step 21 (test assumptions) targets the beliefs that actually determine venture viability, not random hypotheses. Step 22 (MVBP) integrates individually validated assumptions into a sellable artifact that tests the meta-assumption that validated parts work as a system. Step 23 (dogs eat the dog food) demands quantitative proof of both adoption and monetization. Without Step 20, testing is unfocused. Without Step 21, the MVBP bundles untested beliefs. Without Step 22, there is no systems-level test. Without Step 23, the entrepreneur mistakes component-level confidence for market readiness. Each step depends on the prior; skipping any one defeats the learning mechanism the entire framework is built to serve.
Good Examples
Home Team Therapy -- MVBP feature stripping (Step 22). Before MVBP discipline, the product included Kinect-based gesture feedback -- the founder's signature idea. After applying the MVBP test, the team stripped it down to simple online video plus therapist connection. Removing the Kinect eliminated technological, logistics, compatibility, and user-comfort risks while preserving the ability to test the only assumptions that mattered: patient signup, usage, doctor signup, and willingness to pay.
StyleUp -- manual-first MVBP and dogs-eat-the-dog-food proof (Steps 22-23). Kendall Herbst tested personalized weather-appropriate fashion advice by hand-crafting daily emails to friends before writing a line of code. Organic growth to approximately 40 users, then strangers added with open rates holding. Validation metrics: 70% email open rate versus 14% industry average, click-throughs to purchase pages validating the monetization model, referral-link tracking showing organic sharing. Growth from 0 to 8,000 users with minimal spend demonstrated organic pull -- categorically different from push-driven acquisition.
ThriveHive -- three-pillar scalability proof (Step 23). Tested market access (hundreds of qualified leads per month via organic content and AdWords), sales process (single inside sales rep achieving viable unit economics after approximately 6 months), and value delivery (monthly churn at low end of industry comparables, organic referrals exceeding 15% of base with no incentives, over 50% of customers expanding their businesses). Beta-to-paid conversion rate: 74%. All three pillars had to converge -- market access without sales process means leads die; sales without value delivery means churn kills unit economics; value delivery without access means the venture cannot grow.
Counterpoints
Self-referential bias -- smartphone shopping app (Step 21). The team assumed Whole Foods shoppers used smartphones while shopping because the team members did. Direct store observation revealed virtually zero shoppers in the target demographic used a phone. Interviews confirmed ownership but no desire to change existing habits. The corrective is direct field observation in the actual usage environment -- not surveys, not interviews, not self-reflection. Building for a user who does not exist is discovered only after product investment when observation is skipped.
Logic does not guarantee adoption -- IBM electronic medical records (Step 23). Technology was sufficient, logic was compelling, but doctors refused to adopt for over two decades. Hundreds of startups died on this assumption gap. A product can satisfy every rational criterion -- superior performance, lower cost, lab validation, stated customer intent -- and still fail at adoption because humans are not always rational.
Free usage without payment is not validation (Step 23). Free usage does not prove a business exists. The conversion rate at the moment customers must pay is the true test. Payment without real usage is equally dangerous -- churn will kill the venture. Either dimension alone creates a false positive. Both adoption and monetization must be demonstrated simultaneously.
Key Quotes
"Actions speak louder than words -- verbal validation is insufficient because customers are polite, aspirational, and unreliable reporters of their own future behavior." (Step 20)
"Everything should be made as simple as possible, but not simpler." (Einstein, cited Step 22)
"You do not have a business until someone pays; you do not have a product until a buyer extracts value." (Step 22)
"Beta-test their wallet." (Dharmesh Shah / HubSpot, cited Step 23)
"Customer preferences surfaced this way are 'the gold that will make you rich.'" (Step 23)
"Be intellectually honest when interpreting whether customers are truly using and paying." (Step 23)
Rules of Thumb
- Decomposition rule (Step 20): Compound assumptions must be decoupled into independent, single-variable statements, each testable by one experiment. Do not skip assumptions because testing seems difficult.
- 5-10 key assumptions (Step 20): Review all prior steps and extract the untested logical conclusions. Priority domains: Persona priorities, value proposition attractiveness, workflow integration, cost targets, Next 10 Customer lighthouse vs. linchpin distinction, DMU validation.
- Customer commitment ladder (Step 21): Signal strength descending: prepay, deposit, letter of intent, agree to pilot, express strong interest with conditions. Verbal enthusiasm without commitment is noise.
- Cheapest, quickest, easiest test (Step 21): Most tests require no code and no physical product. Expensive tests create sunk-cost pressure to interpret results favorably.
- Distinguish hypothesis failure from experiment failure (Step 21): When a test fails, the first question is whether the test design was flawed, not whether the hypothesis is wrong.
- Revealed preference over stated preference (Step 21): What people do is stronger evidence than what people say they would do. Test in the real environment; artificial settings produce artificial behavior.
- MVBP, not MVP (Step 22): Three simultaneous conditions -- customer gets value from use, customer pays, product starts a feedback loop. MVP permits technology demos that never face the market's ultimate question: will someone pay?
- Manual processes are legitimate MVBP components (Step 22): StyleUp and ThriveHive both used manual fulfillment behind automated facades. Manual execution isolates demand-signal from technology risk.
- Two-dimension validation (Step 23): Must prove both usage (the dogs eat) and monetization (someone buys the dog food). Either alone creates a false positive.
- Virality coefficient (Step 23): Measures whether customers tell others. High virality drives organic growth, directly reducing COCA and improving the LTV >= 3x COCA equation. Track via referral links and Net Promoter Score.
- Three-pillar scalability test (Step 23): Market access (repeatable lead generation), sales process (viable unit economics), and value delivery (surplus value creates retention and referrals). Only convergence of all three proves the business is real.
Related References
- unit-economics.md -- LTV and COCA projections are the assumptions most critically tested by MVBP deployment and dogs-eat-the-dog-food metrics.
- competitive-positioning.md -- Core definition, DMU mapping, and acquisition process assumptions from Steps 10-13 feed directly into the assumption inventory of Step 20.
- business-model-pricing.md -- Business model and pricing choices become testable assumptions validated through MVBP payment data and beta-to-paid conversion rates.