The plan
Today: Fort Worth bakes for itself, Saginaw and NRH. Grapevine bakes for itself and for Tex Donuts, our sister store.
Phase 1: all baking moves to Saginaw. The Grapevine baker ($5,900/month) comes off payroll. The other bakers (Fort Worth and Grapevine) move to Saginaw. Every store other than Saginaw runs with one cashier. One driver runs one route each morning in a new BrightDrop 600 van, and Tex Donuts pays $25/day for delivery. The stop costs us almost nothing because Tex is a block from NRH.
Route length comes from OSRM driving directions between the four shop addresses: 20.8 mi through the stops plus 23.0 mi back from Grapevine to Saginaw, about 44 mi a day in total. Tex is counted as the same point as NRH.
Monthly P&L impact
Loan payments are included in the monthly costs, so the net figure is cash in the bank after financing.
Break-even and payback
Can the van beat Grapevine's 4:30 open?
Drive times are OSRM free-flow estimates, which is about right before dawn. Each stop allows 10 minutes to unload. Store hours are from goldenglazedonuts.com.
What this doesn't count (upside)
- Cashier savings from going to one cashier per store. You didn't give me that number, so it's left out.
- Fort Worth and Grapevine space and utilities freed up once they stop running fryers.
- Less waste and more consistent product from one bake team. Unmeasured, so left out.
What has to be true (risk)
- Saginaw can bake four stores plus Tex by about 3:00 AM. This is the real gate, not the money. It needs fryer, proofer and hood capacity, bench space and walk-in storage. The $20k equipment figure is an estimate until someone walks the kitchen.
- Grapevine is the last stop and opens at 4:30. It's our second-best shop (+$36k Jan–Jul NOI), so a late van costs more there than anywhere else. The van has to leave Saginaw by about 3:05 AM. If the timing slips, swap the order to Saginaw → Grapevine → Tex → NRH → Fort Worth. Fort Worth opens at 4:00 but is only 6 minutes from Saginaw, so a second short run could cover it.
- Moved bakers cost the same as today. This model assumes their pay doesn't change. Extra drive time or overtime would come straight out of the savings.
- A single van is a single point of failure. The existing van is the backup. There's no plan yet for a breakdown or a driver who doesn't show.
- BrightDrop is a discontinued platform. GM ended production in Oct 2025, so resale is near zero and parts support is unknown. That's why the payback math doesn't count any resale value.
Why this is the right first step
It tests the central kitchen idea with no lease, no buildout and no raise. We learn our real bake capacity, whether the route timing holds and how product holds up in transit, all in a shop we already run. If Saginaw holds four stores, those numbers become the sizing input for the full kitchen on the main model in place of the research defaults.