Teaching example: subscription lunches
This is a classroom business: hot lunches delivered by subscription to offices. Every number below is a bet written before experience. None of these numbers is market evidence, a measured result, or a recommendation for a real kitchen.
The funnel in the example starts from 100 showings and a price of 4,500 in the local currency per week. Currency here is the one used in the original teaching deck.
L-shaped frame, not all 35 cells
The basic level does not fill the whole matrix. The Business row links 3–5 years to this week. The Week column collects the action: Customer, Product, and Founder.
Full matrix, as a picture of the bets
| Category | 3–5 years | 1 year | Quarter | Month | Week |
|---|---|---|---|---|---|
| Business | Company value of $1 billion | At least 1,000 active subscriptions; at least 18 million in four weeks | At least 24 first paid continuations | At least 20 requests; at least 8 paid continuations | At least 5 requests from 100 showings |
| Customer | At least 80% of customers renew | At least 60% remain after four weeks | 36 started a trial; 24 paid to continue | 20 requests, then 12 trials, then 8 payments | 100 showings, then at least 5 requests |
| Product | At least 98% of orders delivered on time | On time at least 95%; returns at most 3% | 36 trials of 7 days; at most 2 critical failures | At least 95% of deliveries on time | A page: menu, price 4,500, request form |
| Market | A working model in at least 10 regions | At least 3 stable segments in one region | 1 segment: 1,200 showings, then 60 requests | 400 showings to one segment | 100 showings to office managers |
| Economics | Gross margin at least 30%; cash for at least 12 months | Weekly cost at most 3,150 out of 4,500 | Cost to acquire a paying customer at most 10,000 | Four tests at most 80,000 | Test at most 20,000; one request at most 4,000 |
| Team | At least 10 regions run without the founder | 4 key roles filled | At least 95% of deliveries done to plan | 4 owners: menu, sales, kitchen, delivery | Each of 4 steps has an owner |
| Founder | At most 20 hours of work a week | At most 10 current decisions a week | At least 12 bets checked | Average cycle to a signal at most 7 days | A decision on the bet within 7 days |
One week, written as a contract
Bet. If we show a menu at 4,500 to 100 office managers, we get at least 5 requests and at least 3 agreed trials.
Actions. Business: 100 showings. Product: menu and bot. Sales: a call within 15 minutes.
Sample and threshold. 100 showings. Win line: at least 5 requests from the page and at least 3 trials.
Guardrail and deadline. Spend at most 20,000. Form failures = 0. Call within 15 minutes. Look on Friday at 18:00.
Why the signal is early in this example
A paid subscription is weeks away: a trial week, an invoice, a payment. Interest can show in the first 48–72 hours. If 100 showings produce 0 requests, the offer missed the pain, and the founder can change it on Friday instead of waiting for the month to end.
Four outcomes, with this week's lines
| Outcome | Fact in this example | Next decision |
|---|---|---|
| Bet held | At least 5 requests, guardrail intact | Scale: keep the result, or move toward payment |
| Bet failed | Fewer than 5 requests, guardrail intact | Change the offer, the audience, or the price |
| Guardrail broke | Form failures above 0, or spend over the limit | Stop and repair |
| Still collecting | For example, 30 of 100 showings done | Finish the showings. Do not change the offer yet |
A journal row from the teaching story
Week 1 starts with no customer behavior yet. The result is unknown. Arguments about architecture, before any customer has acted, are guesses. The decision is to run 100 showings by Thursday and read the scorecard on Friday.
A later teaching row: the bet was that a call within 15 minutes would yield at least 3 trials from 5 requests. The fact was 0 trials from 5 requests, because the calls went out after 4 hours. What was learned: the customer cools within 30 minutes, so speed of contact is the constraint. The decision: connect the bot within 2 hours and call within 5 minutes.
That second row is still a classroom story, not a measured study.
Scorecard shape
Month aim in the example: at least 8 first payments to continue, without making the service worse.
| Measure | Role | Week threshold | Fact at the start | Status |
|---|---|---|---|---|
| Requests from the page | Leading | At least 5 from 100 showings | 0 from 100 | Still collecting |
| Form and bot failures | Guardrail | Strictly 0 | 0 | On plan |
| Agreed trials | Leading | At least 3 from 5 requests | 0 from 5 | Waiting for requests |
| Continuation payments | Lagging | At least 8 by month end | 0 from 8 | Waiting for trials |
Read the facts before anyone explains them. Keep only the rows that change this week's actions.