The first five prepayments

At the end of a pre-sale, every founder wants the result to be a win. Oscar admits it: if only two shops had paid, he'd have found a way to call that a win too. Everyone would. That's exactly why the rule went on paper before the first ask.

GarageDesk's rule: build if at least 3 of the 12 shops prepay by the end of January. The result: 5 prepaid, 4 said "call us when it's ready", 3 said no. Five passes three, so Nina builds from 1 February. The decision was made before anyone felt anything about it.

The rule only counts prepayments. It's tempting, when the count falls short, to add other answers: "two paid, and four more are interested, so that's six". But "call us when it's ready" costs nothing and is a "not now". Adding it to the count is moving the line after seeing the result.

If two had paid, the honest reading would have been: the rule says don't build yet. Go back to the problem or the offer, not to the code.

After a pre-sale, the first number to know is the prepaid cash: everything shops paid, across every plan.

prepaid cash = shops on each plan × that plan's price, added up

GarageDesk's five prepayments: Greg and two others took three months at 1,800 UAH; Dennis and one other shop took the year at 6,000 UAH.

• Three-month plans: 3 × 1,800 = 5,400 UAH.
• Yearly plans: 2 × 6,000 = 12,000 UAH.
• Total: 5,400 + 12,000 = 17,400 UAH.

Until 1 April that money is a promise, not income. If GarageDesk isn't live by then, every hryvnia goes back. So the founders hold the whole sum aside and spend nothing from it that they couldn't take back.

Two slips are common: multiplying every shop by the same price (5 × 6,000 = 30,000 or 5 × 1,800 = 9,000), and forgetting that the cash and the monthly revenue are different numbers. The 17,400 arrived at once; the monthly revenue it represents comes later, spread over the months of service.

The refund promise is part of what the shops bought. "Full refund if we're not live by 1 April" is what made it reasonable to pay for a product that didn't exist yet. So when a shop asks for its money back, even before the date and even if the launch is on track, give it back, quickly and politely.

Why not argue?

• The promise was the point of the offer. A refund given grudgingly tells every other shop that the guarantee was only words.
• Word travels between shops. Repair shop owners talk to each other; a story about a startup that wouldn't refund spreads further than the refund itself.
• A reluctant customer is a poor early adopter. Keeping money from someone who wants out doesn't make them a customer; it makes them a critic.

It's fine to ask, once and without pressure, why they changed their mind. The answer is useful information. But the refund doesn't depend on the answer.

This is also why prepaid cash is held aside: a refund should never depend on whether the money has already been spent.

After launch, the question changes from "how much cash came in?" to "how much does GarageDesk earn each month?" That's MRR, monthly recurring revenue, and each plan counts by what it brings per month:

• a three-month shop: 1,800 ÷ 3 = 600 UAH a month;
• a yearly shop: 6,000 ÷ 12 = 500 UAH a month.

For GarageDesk's five: 3 × 600 = 1,800, plus 2 × 500 = 1,000, so MRR = 2,800 UAH a month.

Now compare it with the ladder. Hosting and tools cost 2,000 UAH a month, so 2,800 covers the tools milestone: the product stops costing the founders money. The first payment rung was reached the day Dennis paid. The next rung, Nina's 30,000 UAH, is still a long way off: 2,800 is under a tenth of it.

Keep the two numbers apart. 17,400 UAH is cash that arrived once; 2,800 UAH is what the business earns in a month. Planning spending on the first instead of the second is how a promising start runs out of money.

A good pre-sale invites overstatement. It helps to sort what's really true from what only feels true.

True at the end of January:

• the pre-sale passed the rule written in advance: 5 prepaid, 3 were needed;
• the first money milestone is reached: a shop has paid real money because the product addresses a real problem. Dennis paid first.

Not true, however good it feels:

• the 17,400 UAH can't be spent freely: it's refundable until 1 April;
• Nina's living costs aren't covered: her milestone is 50 shops at 600 UAH, and GarageDesk has 5 prepaid shops and 2,800 UAH of monthly revenue to come;
• the four "call us when it's ready" shops aren't future customers: they said "not now". One of them has called back to ask about moving his notebook into the system, which is the first sign a "not now" may become a yes, but he hasn't paid.

Reading the result precisely isn't pessimism. It tells the founders exactly where they stand: a test passed, a first rung climbed, a long ladder ahead.

Twelve answers to a pre-sale aren't twelve units of "interest". Each kind tells the founders something different and calls for a different action:

• Prepaid (three months or a year): commitment. These are the early adopters. Keep them informed and involve them before launch.
• "Call us when it's ready": timing, not commitment. Record it as "not now" and share the launch date without pressure. Note who calls back first.
• "Our regulars come back anyway": fit. Small places where everyone knows the owner may simply not be the audience. Only worry if bigger shops start saying it too.
• "I won't put customers' phone numbers online": a risk. You'll hear it again. Log it and prepare a clear answer before launch about whose data it is and who can see it; "we'll figure it out" is not an answer.

Recording the reason behind every no is what turns a pre-sale into learning. Lumping the answers together as "interest" would hide all four of these lessons.

The five shops that prepaid are GarageDesk's early adopters: they paid for what the product will become, not for what exists. That makes them generous and fragile at the same time.

Generous: they'll forgive rough edges, missing features and a plain design, because they expected a first version.

Fragile: they paid on trust, so silence hurts more than bugs. Five weeks with no news before launch makes a prepayment feel like a mistake.

What they need between payment and launch:

• the date, stated plainly: live on 3 March;
• regular, short updates, even when there's little to show;
• concrete requests that involve them: the client list, service dates, who will enter the mileage. Dennis has already asked when he can hand over his niece's notebook;
• the refund promise repeated, so the guarantee stays visible.

What they don't need: new features that weren't part of the offer, promised to keep them excited. Those create obligations the founders didn't plan for.

Treated well, early adopters become the first stories other shops hear.