Customer-funded: the quiet alternative to the seed round

Halyard’s first year was financed by 41 sailors who paid half up front for sails that did not exist. Tom Aldous and Ana Ferreira took $310,000 in deposits against a build queue they had drawn on a whiteboard, then spent the next eleven months converting it into cloth. Nobody called it a round. It was, in every respect that matters, a round: the money arrived before the product, it funded the equipment, and if the thing had failed, someone else would have lost it.

The difference is what they gave up, which was nothing except the obligation to actually build 41 sails. That obligation turned out to be considerably more demanding than a board seat, and considerably more useful.

Bigger than it looks

This isn’t a fringe. The Brackett Group put customer prepayment — deposits, pre-orders, annual-in-advance contracts and retainers — at $4.1B across companies in their first two years last year, against $2.8B of institutional seed into the same cohort. The customer-funded money is smaller per company, obviously: a median of $140,000 against a median seed of $1.9M. There are simply far more of them, and they never appear in a database because there is nothing to file.

The instruments are unglamorous and mostly predate the internet. A deposit is a deposit. What has changed is that founders now reach for it first rather than as the thing you do when the fund says no — Brackett found 44% of first-time founders in 2025 attempted a prepaid launch before approaching any investor, up from 19% in 2021.

It also filters. A pre-order is the only market research that costs the respondent money, which is why it is the only market research worth reading. Fieldnote’s Maya Okonjo priced her first routing contract at $18,000 annual-in-advance specifically to find out whether farm co-ops would pay for software at all. Nine said yes. She has described the other outcome — forty polite meetings and zero cheques — as the cheaper of the two, and she is not entirely joking.

It is debt, and it is denominated in work

The failure mode is specific and it is not the one founders expect. A deposit sits in the bank account looking exactly like revenue. It is not revenue. It is a liability payable in labour, at a price fixed months ago, and the moment a founder spends it on anything other than the thing it was collected for, the company is running an unsecured loan from people who did not agree to lend.

Brackett’s failure sample is instructive: among customer-funded companies that folded in year two, the most common precipitating event was not a lost customer. It was a queue the company had already been paid for and could no longer afford to build, because input costs had moved 20% and the price hadn’t. Deposits are a forward contract. Founders keep writing them without noticing they have taken a position.

The founders who run this cleanly share a small set of habits, none of which are complicated and all of which are annoying:

• The deposit lives in a separate account and is not treated as cash. Halyard has done this since month two and Ferreira describes it as the only accounting rule she would die for.
• It is recognised as it’s earned, not when it lands — which means the P&L shows the truth in month one rather than a triumph followed by a puzzle.
• The queue has a stated length and the price has an expiry. Both of these cost sales and both prevent the year-two event above.
• Refunds are offered before they’re demanded. Every founder who has done this reports the take-up is near zero and the goodwill is not.
• The build capacity is checked against the queue monthly, by someone who is not the person who sold it.

Whether this is better than a seed round is the wrong argument. It is a different instrument with a different cost, and the cost is that you have to be right about demand immediately, with no eighteen-month runway to discover you weren’t. What it buys is that the discovery happens in month one, for $140,000 rather than $1.9M, and the people who lose out are 41 sailors who get their money back rather than a fund that owns 20% of a company with a dead product.

Halyard delivered 39 of the 41 sails. The other two customers changed boats. Both got a refund, and both bought sails the following year, which Aldous notes is a return no term sheet was offering.