The playbook

Four tests on one page. Print it. Run it on any proposal that starts with "we should give away".

Test 1 - The recapture path (Lesson 1)

PartQuestionFailing looks like
Lead itemWhat are we giving away below its own cost, and what does it cost per user?"It's basically free to run" - said about an AI product
Recapture pathBy what named mechanism does a taker become a payer?"And then later they buy the CRM"
Attach assumptionWhat fraction travel it, and what would prove us wrong?No number anywhere in the document

Strongest form of a path is necessity: the lead item makes the paid item required, not appealing. Second strongest is accumulated cost of leaving. Weakest, and most common, is goodwill.

Test 2 - The wedge (Lesson 3)

  1. Is the problem real and outside the incumbent's core?
  2. Can we win it with something simple enough to ship this quarter?
  3. Does winning it accumulate a bridge asset the expansion needs? Name it in one phrase.
  4. Is the expansion genuinely shorter from here than from a standing start?

Bridge assets: accumulated data, workflow position, distribution, trust and reference. Not signups. Second-order question: does the asset accumulate faster than a competitor can replicate it?

Beachhead check, separately: do they talk to each other, share one urgent reason to buy, can we deliver the whole product for them, is anyone entrenched?

For [people] who [urgent reason], we solve [narrow problem], which accumulates [bridge asset], which makes [expansion] materially easier than selling it cold.

Test 3 - The cap (Lesson 4)

Three jobs, in order: let the magic land, let the habit form, then bind. People forget the second.

DecisionGuidance
Metered unitTracks value received; tracks your cost; can be felt coming; grows with habit not luck. Never meter clients or properties - it punishes a good year.
ShapeUsage cap by default. Feature cap risks the magic never landing. Seat cap is inert against solo professionals.
Soft thresholdWhere pressure begins. Set so it arrives after the habit, not during it.
Hard thresholdWhere use stops. Wispr Flow: 2,000 soft / 5,000 hard per week on desktop, 1,000 / 1,500 on iOS.
The betCap capture and keep memory free (switching cost comes from history) or free capture and cap memory (history is the product). Both legitimate. Choose knowingly.
Failure modesToo generous: nobody converts. Too tight: the habit never forms. Expect to find the band by iteration.

Benchmarks: self-serve freemium 3-5 percent good, 6-8 percent great; later framing puts great at 8-12 and AI-native at 15-20. Always compute the signups implied by your revenue target before agreeing to a plan.

Test 4 - The moat (Lesson 5)

  1. Which of the three are you claiming - frequency, habit, or switching cost? Only the third is an asset.
  2. Which kind of switching cost: data, workflow, integration, social, contractual? Contractual is a delay, not a moat.
  3. Split the accumulated data. Re-derivable from sources the user still controls, versus captured live and stored nowhere else. Only the second half defends you.
  4. Is the claim a network effect or personal accumulation? Personal accumulation restarts at zero for every new user, including your competitor's.
  5. Time to value and time to lock. Both must be short, and the second only runs on the non-re-derivable half.

The three checks that catch the most errors

GlossaryReview checklist