The ask · R0 pilot round
We are raising $95–125k to run ten or twelve real Saturdays out of one farmers market, and $450–650k after that to turn a working pilot into product v1. The wedge is that we never build a warehouse: one public origin, optimized routes out of it, and farms that earn ownership before they sign anything. Everything below is a target from the plan, not a result.
What it is
A weekly, chef-designed box of food shopped at the Boulder Farmers Market on Saturday morning by a crew member who bags it there and drives it out on a solved route the same day.
Who owns it
A Colorado cooperative with four member classes: eaters, producers, crew and chefs. Patronage accrues by participation from the first Saturday, including for stalls that have not claimed their page yet; one member, one vote within each class.
Why it compounds
No facility to outgrow, and a producer base that onboards by being shopped. The data layer is open — DFC catalogs, ATProto identity, published lexicons — so a second co-op strengthens the commons instead of competing for it.
Show the split
These are lines in a settlement run, not sentiments in a values statement. The rate card is published in the co-op’s charter, the solver that applies it is deterministic, and every member can read the week’s books.
One box dollar · target
100¢
For comparison — the U.S. food dollar
USDA ERS farm share, most recent series
About 15¢ of a retail food dollar reaches the farm. The other 85 goes to marketing, distribution and margin. A grower at the market keeps the whole price they set, because there is nobody standing in between to take a share of it.
68¢
Stalls
Spent at the market tables on Saturday morning, at the price chalked on the sign. No wholesale discount, no invoice, no waiting.
7¢
Crew
The member who shops the run, bags it and drives it. A $120 floor per run, posted on the shift card before anyone claims it.
10¢
Chef
Ten percent of the box, to the chef whose menu you chose. Written into the rate card, not negotiated week by week.
12¢
Co-op
Insulated totes and gel packs, insurance, the software, and the steward who runs the week. The charter caps this at 15¢.
3¢
Reserve
Shorts, credits and redeliveries, plus the retained capital that eventually opens a second market day.
Targets, not results. The chef fee (10%) and the reserve (3%) are the published rate card to the basis point. The spend target is 68% of the box price with an ±8% band, and the band is the co-op’s problem rather than yours — your price is fixed when you order. The charter caps the co-op’s take at 15¢; the 12¢ and 7¢ above are how this pilot budgets inside that cap. Every Monday the real numbers replace these in the open books.
The scoreboard
The same numbers go to members and to investors, on the same day, in the same open books. Stage 0 is over when the first four hold for a month.
80/wk
Boxes delivered
On three Saturday runs out of one market, ~20 boxes to a run.
12/hr
Stops per crew-hour
Below this the route does not pay for the person driving it.
85%
Boxes shopped in band
Within ±8% of the 68% spend target. The gap is margin, either way.
12
Stalls with attributed spend
Three weeks running by week eight, each with patronage accruing.
≥55%
Eight-week retention
Households still ordering two months in. NPS ≥50 alongside it.
<2%
Error and credit rate
Shorts and substitutions that end in a credit rather than a fix.
≥8
Stalls holding for us
Stage 1: claimed stalls setting produce aside before the crowd.
≤$8
Logistics cost per box
Shop, bag and drive, all in, at twelve stops an hour.
Alongside them, every week: patronage by role, the rung mix (how much of our spend went to stalls at L0, L1, L2 and L3), and what is owed to producers who have not claimed it yet. That last number going down is the whole strategy working.
Where this goes
We are not building a company that runs food co-ops. We are proving a pattern one market at a time — and the pilot’s surplus is earmarked to seed the next one.
A co-op of co-ops
Surplus buys patronage shares in the next town’s foodshed — working capital for their first season, redeemed out of their own patronage engine, and then we step out. No equity held, no perpetual cut, no vote in their governance. Growth without extraction, enforced by the same ledger you can read every Monday.
The standard is a public good
The data layer under this co-op — a farm’s catalog, what is in season, what is offered where — is an open standard we intend to hand to a nonprofit, so any market, food hub, or researcher can read it and no platform can fence it. A farm’s inventory should outlive every app that ever displayed it, ours included.
The Open Food vision →Recipes that travel
With university protocol researchers we are scoping an open recipe standard: every ingredient a real product type, so a dish knows what the Front Range can grow in August. The week’s recipes ride printed in the box and land in the member’s own data store — their cooking history, portable to any app.
R0 — pilot
$95–125k
Sept–Dec 2026
One market · one chef · ≥12 stalls · 50–150 households · 10–12 Saturdays
R1 — product v1
$450–650k
Jan–Jun 2027
Second market day · native apps · ledger v1 · governance · DFC read/write
Capital-light by construction
One origin, and it is a public street on a Saturday. No lease, no cooler room, no truck, no inventory risk carried overnight. The first dollar of this round buys software and a steward, not square footage, and the same is true at ten times the volume: we scale by adding market days, not buildings.
Producer acquisition, solved
A stall does not have to sign anything for us to start buying from it, attribute the spend, and accrue ownership in its name. Twelve stalls with money waiting is a much easier first conversation than twelve wholesale contracts, and every rung a farm climbs makes the week cheaper to run.
Founding members
A founding-member share plus pre-paid boxes for the first twelve Saturdays. You get fed, the co-op gets working capital before week one, and your patronage starts accruing from the first delivery, before incorporation is finished.
Individuals · Boulder County households
Community round
An investor-member class inside a Colorado limited cooperative association: a capped, patient return, no control over the member classes, no pressure to sell the network. Structure under review with co-op counsel now.
Reg CF community round or patient philanthropic capital
R0 use of funds
$95–125k
What we spend at the stalls and what we pay the crew come out of revenue, not out of this. Founding-member pre-purchases sit on top of it. There is no facility line because there is no facility.
What it buys
This is a gate rather than a runway. Eighty boxes a week on three Saturday routes, twelve stops per crew-hour, 85% of boxes shopped inside the ±8% band, twelve stalls with three weeks of our money against their name, and a patronage ledger that reconciles to the cent. If those numbers don’t come in, we say so and stop.