EVIDENCE → DECISION

Validated enough means the next bet is earned

Certainty is unavailable and endless research is hiding in a sensible cardigan. The practical question is whether the evidence justifies the next specific, reversible commitment.

THE VERDICT

Proceed when the right customer repeatedly shows the problem, a realistic offer earns meaningful commitments, delivery works at small scale, and basic economics can support the next expense. Otherwise test the weakest condition again.

Use a six-part evidence gate

  • Customer: interviewees match the people you can actually reach and serve.
  • Problem: they describe recent examples and current workarounds without prompting.
  • Offer: several qualified prospects accept the same clearly priced offer.
  • Delivery: you can fulfill the promise legally, safely, and on time.
  • Economics: price covers direct costs and leaves room for customer acquisition and overhead.
  • Learning: objections and failures now repeat instead of producing entirely new surprises.

Demand stronger proof for a larger or less reversible bet

Next commitmentReasonable evidence
$0–$100 testClear problem interviews and access to the target audience
Paid pilot or small batchQualified buyers attempt or complete purchase at a realistic price
Recurring software or equipmentRepeated demand plus demonstrated time, capacity, or quality constraint
Lease, payroll, large inventory, or debtDocumented sales pattern, margin, cash-flow stress test, and downside plan

These are decision levels, not universal dollar laws. A regulated product may require more work before the first sale. A consulting pilot may require almost none. Evidence should rise with irreversibility and downside.

Prove the sale is worth making

Write a one-sale model: price minus materials, packaging, shipping subsidy, marketplace and payment fees, direct labor, refunds or spoilage allowance, and other variable costs. The remainder is contribution margin—the money available for fixed costs, tax obligations, owner compensation, and profit.

A sale that loses money can still be a deliberate experiment. It cannot validate a sustainable price. Record what the future price and future cost must become, then test those rather than celebrating revenue that injures you.

Choose proceed, test, pivot, or stop

Proceed

All six gates have direct evidence and the next commitment is proportionate.

Test

One important assumption remains weak and a cheap experiment can answer it.

Pivot

The problem is real, but the audience, offer, channel, price, or delivery model repeatedly fails.

Stop

Qualified prospects show no urgent behavior, no credible buying path, or economics that cannot work.

Set a decision date. Research that never reaches a date becomes emotional insurance. Reddit is full of founders who built for months because coding felt more comfortable than asking for a purchase. The opposite trap also exists: repeating interviews forever because setup feels frightening.

Buy only what the next transaction requires

Formation, permits, insurance, a business bank account, contracts, payment tools, and bookkeeping may be necessary depending on the business and location. Research the actual requirement. Then ask, “What sale or delivery does this expense enable or protect?” Premium branding and stacked subscriptions rarely pass that test.

SETUP BUDGET RULE

List each proposed expense beside the evidence that earned it. If the reason is “to look legitimate,” ask which customer, law, bank, insurer, or workflow specifically requires it.

Your finish line

Write a one-page decision memo: evidence collected, score for each gate, one-sale economics, next commitment, maximum loss, and review date. Validation ends when the memo makes the next action obvious—not when anxiety disappears.