This document is v0.1. Parts of it have been falsified. Read the Execution Plan first.
Since this was written we pulled Golden Glaze's actual purchase ledger, obtained the BakeMark cost sheet, and scraped Texas Bakery Supply's complete published price book. Two load-bearing theses in this document did not survive: price transparency is not a wedge (the incumbent already publishes 952 priced SKUs), and we do not beat them on price today — normalized per pound, Golden Glaze pays more than TBS shelf on 8 of 14 clean matches.
Everything on economics, cost to roll out, margin, routes, outreach and open action items now lives in the Operating & Execution Plan v1.0. Where the two disagree, v1.0 wins. This page is retained as the strategic frame and the market sizing.
The business in one page
Golden Star Supply sells the five commodity inputs every donut shop burns daily — mix, shortening, glaze sugar, boxes, bags — to independent shops across DFW, at a published price, delivered inside the overnight production window, ordered by text.
What we are
A narrow, high-frequency distributor. We do not intend to carry 4,000 SKUs and compete with broadline foodservice. We intend to own the handful of items that represent the largest share of an independent donut shop's cost of goods, win them on price and reliability, and expand the basket only once the delivery relationship is proven.
Why it works
Three assets already exist and are paid for by another business. Golden Glaze's purchasing volume gets us to distributor pricing tiers on day one. The centralized kitchen already planned for Golden Glaze is the same building, dock, and truck route a distributor needs. Golden Ops, the software already running the donut chain, is the ordering and fulfillment backend. The marginal cost of entering distribution is far below the standalone cost.
The honest version
Distribution is a working-capital business with thin margins and unforgiving delivery windows, and we would be selling to shops that compete with our own. Those are real constraints, addressed in Section 13. This plan is deliberately conservative on capture rate and deliberately loud about the cash cycle.
The ask
Approximately $300K–$400K of committed capital — the large majority of it working capital rather than fixed assets — to reach roughly 40 active accounts and a self-funding cash cycle within four quarters. Full build-up in Section 11.
Digital ordering is the engine, not the pitch
The founding instinct for this business was to remove the phone-and-text back-and-forth of wholesale ordering and replace it with a clean digital experience with integrated payment. That instinct is correct about the mechanism and wrong about the audience, and the distinction determines how we sell.
The ordering friction is painful to a buyer operating seven shops. It does not scale, it produces errors, and it consumes management attention. It is substantially less painful to a single-location owner placing one standing order a week with a representative he has known for a decade. For that owner, the text message is the convenience. He is 55, he has been awake since 2 a.m., and he is not going to adopt a vendor portal.
The reframe
Digital ordering is our cost-to-serve advantage, not our value proposition. It is what allows one operator to carry 300 accounts without hiring five order-takers. That cost structure is what funds a lower price. The customer-facing offer is price and reliability. The software is what makes that price possible. We sell the outcome, not the mechanism.
Where digital genuinely does sell: transparency, not convenience
Bakery distribution runs on opaque, representative-negotiated pricing. Two shops on the same street pay different prices for identical shortening, neither knows it, and price increases arrive without warning as a revised sheet. A flat, published, identical-for-everyone price list is genuinely disruptive in this category in a way that a slicker order form is not.
The message that lands is not "ordering is easier." It is "no haggling, no waiting on your rep to call back, no surprise price sheet in March, and the guy down the street pays exactly what you pay."
Design principle: meet them where they are
The winning implementation is not a portal. It is: the shop texts or messages the order exactly as it always has → it arrives structured on our side → the system auto-confirms with a line-item total → payment by link. To the customer nothing changed. To us it is clean data, no phone tag, and no transcription errors. We capture the scale benefit without asking the customer to change a single habit.
Any design that requires behavior change from a 2 a.m. baker is how this business fails quietly over eighteen months.
Sizing from the ground up
We built the market model from an actual census of the metroplex rather than a top-down industry estimate. Every figure below traces to a specific, verifiable count.
The census
We mapped every donut shop, kolache shop, panaderia, and independent bakery across the DFW metroplex and outer counties — 1,669 businesses, each one geographically verified inside a Dallas–Fort Worth bounding box. Methodology in Appendix A.
| Segment | Count | Assessment |
|---|---|---|
| Tier A — donut & kolache shops | 1,061 | Primary target |
| of which: franchise units | 116 | Mandated / preferred supply — largely unwinnable |
| of which: true independents | ~945 | Serviceable target market |
| Tier B — bakeries, panaderias, Asian bakeries | 276 | High flour/sugar/shortening volume; under-courted |
| Tier C — requires classification review | 187 | Expansion pool |
| Tier X — national chains, grocery counters, non-bakery | 145 | Corporate supply chains — excluded |
The franchise finding
116 of the 1,061 tier-A shops are franchise units — 44 Shipley Do-Nuts alone, plus Daylight Donuts, Donut Palace and similar. Franchise agreements typically mandate or strongly prefer approved suppliers. Treating the tier-A count as the addressable market would overstate the opportunity by roughly 11%. The real target is ~945.
Spend per shop
An independent donut shop's cost of goods runs roughly 25–30% of revenue. Core baking inputs and packaging — mix and flour, frying shortening, glaze and icing sugar, boxes, bags, cups — represent the majority of that line, with the remainder in coffee, dairy, and kolache proteins we do not initially intend to carry.
| Market layer | Units | Annual spend | Definition |
|---|---|---|---|
| Tier-A independents | 945 | ~$47M | Core serviceable market |
| Tier-B bakeries | 276 | ~$14M | Adjacent, higher flour volume |
| Total DFW serviceable | 1,221 | ~$61M | Excludes franchise + chain |
| Year-3 target share | 250 | ~$12.5M | ~20% of serviceable units |
A $61M serviceable market inside a single metro, served predominantly by two or three incumbents with opaque pricing and no digital layer, is a defensible place to start. Expansion geography — Houston, Austin, San Antonio, Oklahoma City — carries similar independent-shop density and is deliberately out of scope until DFW route density is proven.
Who actually buys, and what they care about
The independent donut shop is overwhelmingly an owner-operated family business. The owner is on-site during production, personally places the orders, personally signs for delivery, and personally feels every dollar of cost of goods. Decision-making is fast and undelegated — which cuts both ways: there is no procurement committee to navigate, and there is also no one to overrule a bad first impression.
What they actually care about, in order
1. Price
Ingredients and packaging are the largest controllable line in the P&L. A 6–10% reduction on that line is meaningful, immediate, and easy for an owner to verify against last month's invoice.
2. Delivery timing
Production runs roughly midnight to 5 a.m. A delivery that misses the window is not an inconvenience — it is a lost sales day. Reliability is worth more than a marginal discount, and it is where incumbents most often fail.
3. Order minimums
Broadline minimums force small shops to over-order and tie up cash in inventory they do not have the storage for. Low or no minimums are a genuine differentiator for a single-location operator.
4. Terms
Cash flow is tight and seasonal. Net-14 or net-30 against COD is a real decision factor — and, from our side, the primary driver of working capital requirements.
5. Accuracy and credit handling
Short shipments and wrong items are routine in this category, and the resulting credit fight is a recurring irritation. Getting the order right, and resolving errors without argument, builds durable loyalty.
6. Price predictability
Commodity volatility in flour and oil gets passed through without warning. Advance notice of price movement is a low-cost, high-trust differentiator.
Volume segmentation
Absent published revenue data, Google review count is the best available free proxy for customer traffic and therefore ingredient burn. Across the 1,061 tier-A shops the median is approximately 111 reviews.
| Segment | Shops | Coverage approach |
|---|---|---|
| Whale — 500+ reviews | 29 | Personal visit from the founder. Disproportionate volume; worth bespoke pricing and terms. |
| Core — 110–500 reviews | ~430 | Route-based field sales. The engine of the business. |
| Long tail — under 110 reviews | ~490 | Inbound and self-serve only. Do not spend field time until route density exists. |
Sequencing consequence
The 29 whale accounts represent the fastest path to route density and credibility. They should be worked personally and first — not assigned to a representative, and not approached with the same script as a two-employee shop.
Three assets a distributor cannot buy
The case for this business is not that distribution is attractive. It is that we can enter it at a fraction of the normal cost and with credibility no incumbent can replicate.
We operate the customer's business
Nobody at the incumbent distributors has fried a donut at 3 a.m. We run seven shops going on ten. We can say "this is the shortening we run in our own stores" and mean it literally.
That produces a second-order advantage: we know real burn rates by shop volume, which makes par-level automatic replenishment genuinely accurate rather than a sales gimmick.
Anchor volume from day one
Distribution economics are volume-tier economics. Golden Glaze's own seven shops — ten by year end — are guaranteed baseload consumption that lifts us into better purchasing tiers immediately.
This inverts the usual cold-start problem: we can be price-competitive to a third-party customer before winning a single external account.
Shared infrastructure
The centralized kitchen already planned for Golden Glaze is the same asset a distributor needs: receiving dock, dry and cold storage, racking, and overnight truck routes that already pass the same neighborhoods.
And Golden Ops, the software already running the chain, is the ordering and fulfillment backend — an extension of a working codebase, not a build from zero.
The capital-efficiency argument
Read together, these mean the marginal capital to enter distribution is materially below the standalone cost, and the payback is faster. The commissary is being built regardless. The trucks are running regardless. The software exists. Golden Star Supply is a second revenue stream layered onto infrastructure already being purchased for another purpose.
What we put in front of a shop owner
Do not ask for the whole order
Displacing an entrenched distributor by asking a shop to switch its entire supply relationship fails — it is too much risk in one decision, against a vendor the owner has trusted for years. The reliable method is a beachhead: win a small number of high-volume, easily comparable items, prove delivery, and expand the basket over the following two quarters.
| Beachhead SKU | Why this one |
|---|---|
| Donut mix / flour | Largest single volume line. Directly comparable unit-for-unit. Immediately verifiable saving. |
| Frying shortening | High spend, high frequency, commodity-priced. Where opaque pricing hurts shops most. |
| Glaze & icing sugar | Consumed daily, simple spec, low switching risk. |
| Boxes & bags | Zero product risk — packaging cannot ruin a batch. The easiest first "yes" in the entire catalog. |
| Cups & lids | Rounds out the packaging basket and consolidates a second vendor. |
The door-opener
"Send me your last invoice. I'll price it line for line and show you the delta."
This is the highest-converting opening in wholesale distribution. It asks for no commitment, it makes the value concrete and specific rather than promissory, and it hands us competitive pricing intelligence on every single attempt — whether or not the account closes.
Risk reversal
- Published price list. Same price for every shop. No negotiation, no rep discretion, no volume games below the posted tiers.
- Delivery window guarantee. A defined overnight window with a credit if we miss it. This is the promise incumbents will not make, and the one owners care most about.
- No long-term commitment. No contract, no exclusivity on the first engagement. The switching cost of trying us must be near zero.
- Advance notice on price movement. Written notice before commodity pass-through takes effect.
- Low or no minimum on the beachhead items. Directly attacks the over-ordering problem broadline minimums create.
How the thing actually runs
The ordering layer
Built as an extension of Golden Ops. The customer-facing surface is deliberately unremarkable:
- Shop sends its order by text or messaging app, in whatever informal format it already uses.
- The message is parsed into structured line items against that shop's catalog and pricing.
- An automatic confirmation returns the interpreted order with a line-item total for approval.
- Payment by link, or on terms for approved accounts.
- The order drops into the pick list and route plan without human transcription.
A web ordering surface and standing par-level auto-replenishment sit on top of the same system for the accounts that want them — expected to be the younger and multi-unit operators — but no customer is ever required to use them.
Fulfillment
- Warehouse: co-located with the Golden Glaze commissary. Dry storage dominant; limited cold requirement in the beachhead catalog, which materially simplifies phase one.
- Routes: overnight delivery aligned to the production window. Built by geographic cluster — density per route is the primary determinant of unit economics.
- Fleet: one box truck at launch. A second added at roughly 60–70 accounts depending on cluster spread.
- Inventory: narrow, deep, fast-turning. The discipline of the beachhead catalog is what keeps working capital survivable.
Why the narrow catalog is a strategy, not a limitation
Every additional SKU consumes cash, storage, and forecasting attention while diluting purchasing leverage. A distributor that carries five items at genuine scale beats one that carries four hundred at none. Basket expansion should be earned by proven route density — not attempted at launch.
A field business, not an email business
We built and geo-verified a 1,669-record lead list for DFW. Working that list revealed something that should shape the entire sales plan: this segment is unreachable by email. Only 185 of the 1,669 shops have a website at all, and exhaustively crawling every one of them produced just 51 valid email addresses across roughly 45 distinct businesses.
Channel conclusion
Email covers under 5% of the market and will not improve materially. This is a phone, drive-by, and referral business. Any go-to-market plan built on email sequences is planning against reality. The lead list's value is its verified addresses and route clustering, not its contact emails.
The motion
Phase 1 — Cluster pilot
Select one ZIP cluster with high shop density. Work 25–30 shops in a single geography rather than spraying across the metroplex. Route density is what makes delivery economics work, and a tight pilot tests the offer, the pricing, and the delivery promise simultaneously.
Target: 8–12 accounts from the pilot cluster.
Phase 2 — Whale accounts
The 29 shops above 500 reviews, worked personally by the founder. Bespoke pricing where volume justifies it. These accounts anchor route density and generate the referrals that make phase three cheap.
Phase 3 — Cluster expansion
Replicate the pilot playbook cluster by cluster across the metroplex. Add route capacity only when density supports it — never ahead of it.
Phase 4 — Referral engine
This is a tight, connected community with dense informal networks. A satisfied owner is a more effective sales channel than any campaign we could run. Referral incentives should be introduced only once delivery reliability is genuinely proven.
Call timing
Owners are on-site and reachable during production and early morning; they are asleep or unavailable in the afternoon. Field contact should be scheduled accordingly — a detail that materially affects contact rate and is routinely ignored.
Who we are taking share from
| Competitor type | Strengths | Where they are vulnerable |
|---|---|---|
| Texas Bakery Supply and regional bakery specialists |
Entrenched relationships, broad catalog, established routes, decades of trust | Opaque per-customer pricing; no digital layer; relationship quality varies by rep; slow to respond to service failures |
| Broadline foodservice Sysco, US Foods, PFG |
Enormous scale, full catalog, deep logistics capability | High minimums; small independents are unattractive accounts; generalist reps with no donut-specific knowledge; poor fit for a single-location shop |
| Cash-and-carry restaurant depots, wholesale clubs |
No minimum, immediate availability, no delivery dependency | Owner's time cost is real and unpriced; inconsistent stock; no terms; no delivery into the production window |
| Direct from manufacturer | Best unit pricing at volume | Pallet-scale minimums put it out of reach for nearly every independent |
Our position
Between the broadline distributors that do not really want these accounts and the cash-and-carry option that costs the owner their own labor, there is room for a narrow, donut-native, transparently priced, reliably delivered supplier. That gap is the business.
Expect retaliation
Incumbents will defend accounts on price when they see a pattern of losses. We should assume our early wins provoke targeted discounting, and should not plan to win a price war — our defensibility has to come from delivery reliability, pricing transparency, and operator credibility, all of which are harder to copy than a discount.
The model
Per-account economics
| Metric | Assumption | Basis |
|---|---|---|
| Addressable spend per shop / year | $50,000 | Midpoint of modeled $45–60K band |
| Beachhead capture (year 1) | ~60% | Five core SKUs, not full basket |
| Revenue per account / year | $30,000 | Ramping toward full basket over time |
| Gross margin | 18–22% | Food distribution norm; anchor volume supports upper half |
| Gross profit per account / year | $5,400–6,600 | |
| Cost to acquire an account | $400–800 | Field time, samples, first-delivery incentive |
| Payback period | ~6–8 weeks | On gross profit |
Payback measured in weeks rather than years is the attractive feature of this model. The constraint is not customer acquisition cost — it is the cash tied up in inventory and receivables while accounts scale.
Three-year projection
| Year 1 | Year 2 | Year 3 | |
|---|---|---|---|
| Active accounts (exit) | 40 | 120 | 250 |
| Share of ~945 independents | 4.2% | 12.7% | 26.5% |
| Avg revenue / account | $30K | $42K | $50K |
| Revenue | $1.2M | $5.0M | $12.5M |
| Gross margin % | 18% | 20% | 22% |
| Gross profit | $216K | $1.0M | $2.75M |
| Operating expense | $340K | $700K | $1.5M |
| EBITDA | ($124K) | $300K | $1.25M |
Year one is intentionally modeled at a loss. A distributor that reaches profitability in its first year has almost certainly under-invested in route density, and route density is the entire long-term moat. Note also that the year-three figure represents roughly a quarter of the independent shops in the metroplex — an ambitious but not implausible position for a specialist in a category served by generalists.
Operating expense build
| Line | Year 1 | Note |
|---|---|---|
| Warehouse (incremental) | $36K | Marginal cost within the commissary footprint, not standalone lease |
| Driver | $52K | One overnight route |
| Inside sales / operations | $48K | Order desk, account service |
| Field sales | $60K | Base plus commission; founder-led in early quarters |
| Vehicle — operating, fuel, insurance | $34K | |
| Software & payment processing | $18K | Golden Ops extension; low marginal build cost |
| Insurance, licensing, compliance | $22K | Food handling, commercial auto, general liability |
| Bad debt provision | $24K | 2% of revenue — realistic for terms-based independents |
| Contingency | $46K | |
| Total | $340K |
What it takes to start, and where it goes
| Use of funds | Amount | Note |
|---|---|---|
| Opening inventory | $90K | Narrow catalog, deep stock, fast turns |
| Accounts receivable float | $120K | The real constraint — see below |
| Delivery vehicle | $55K | Used box truck; lease alternative preserves cash |
| Warehouse setup — racking, handling equipment | $25K | Within commissary buildout |
| Software build | $20K | Golden Ops extension, not greenfield |
| Operating runway to breakeven | $60K | |
| Contingency | $30K | |
| Total | $400K | Lean case ~$300K with vehicle leased and slower ramp |
Working capital is the actual risk in this business
More than half the capital requirement is inventory and receivables, not equipment. We pay suppliers before customers pay us, and that gap widens with every new account we win. This is the mechanism by which growing distributors run out of cash while showing a profit on paper.
Mitigations: negotiate supplier terms at least as long as customer terms; open new accounts on COD or card and earn terms over time; keep the catalog narrow so inventory turns fast; treat the receivables line as a hard operating metric reviewed weekly, not a quarterly finance exercise.
Relationship to the Golden Glaze raise
Golden Glaze is separately planning a ~$1M raise for the centralized kitchen. These are related but should be presented as distinct capital decisions with a shared asset base. The commissary is justified on Golden Glaze economics alone; Golden Star Supply improves the return on that asset by loading additional throughput onto it. Investors should be able to evaluate each on its own merits, and the supply business should not be used to justify the kitchen or vice versa.
Milestones and decision gates
| Phase | Objective | Gate to proceed |
|---|---|---|
| 0 — Validate Weeks 1–4 |
Replace modeled assumptions with real numbers: Golden Glaze per-shop purchase data, live supplier quotes at anchor volume, true landed cost on the five beachhead SKUs. | Verified gross margin ≥15% at realistic pricing. If not, stop here. |
| 1 — Pilot Months 2–4 |
One ZIP cluster. 25–30 shops worked. Deliver on Golden Glaze's own shops first to prove the operation before exposing it to a paying customer. | 8+ external accounts, >95% on-time delivery |
| 2 — Prove the route Months 4–8 |
Expand to adjacent clusters. Work the 29 whale accounts personally. Digital ordering layer live. | 25+ accounts, positive contribution margin per route |
| 3 — Scale DFW Months 8–18 |
Second truck and route. Basket expansion beyond beachhead. Referral program. | 100+ accounts, EBITDA positive |
| 4 — Expand Month 18+ |
Second metro, or deepen DFW share and widen the catalog. | Decision point — do not pre-commit |
Phase 0 is the whole plan right now
Everything downstream depends on two numbers we have modeled but not measured: real spend per shop and real gross margin at our purchasing volume. We own seven donut shops. Both numbers are sitting in our own invoice history and in quotes we could request this week. Nothing else should be committed until they are confirmed.
What could kill this
| Risk | Severity | Mitigation |
|---|---|---|
| We compete with our own customers Golden Glaze is a donut chain expanding 7→10 shops |
High | Separate brand and separate entity. A written commitment not to open a Golden Glaze location within a defined radius of an active supply customer. Direct, rehearsed handling of the objection rather than avoidance — "I'm a donut operator, I'm buying this volume anyway, I'd rather share the curve than not." This objection will arise in nearly every first conversation and must be answered before the first sales call, not during it. |
| Working capital exhaustion | High | Supplier terms ≥ customer terms. New accounts start COD. Narrow catalog for fast turns. Weekly receivables review as an operating metric. |
| Delivery failure A missed overnight window costs the shop a sales day |
High | Pilot on our own shops before external customers. Conservative route density. Backup driver arrangement. Treat the first service failure at any account as a founder-level escalation. |
| Incumbent price retaliation | Medium | Do not compete on price alone. Defensibility from reliability, transparency, and operator credibility. Anchor volume gives real cost headroom. |
| Commodity volatility flour and oil move sharply |
Medium | Explicit pass-through terms with advance notice. Do not hold fixed prices without a supply hedge behind them. |
| Founder capacity two operating businesses already |
High | Honestly the most underrated risk in this plan. Phase 1 needs a dedicated operator, not spare evenings. Gate phase 2 on that hire. |
| Customer concentration | Medium | Whale accounts accelerate density but create dependency. Cap any single account's share of revenue as the book grows. |
| Franchise segment inaccessible | Low | Already excluded from the model. 116 units treated as unaddressable rather than assumed winnable. |
| Assumptions prove wrong at Phase 0 | Medium | This is a feature of the plan. Phase 0 is cheap and fast, and it is explicitly designed as a kill gate before meaningful capital is committed. |
Data & methodology
A. How the lead list was built
- Source 1 — OpenStreetMap / Overpass. Bounding box covering the full metroplex and outer counties. ~605 points of interest; 260 independent donut shops plus 173 bakeries and panaderias. Gaps in address data backfilled by reverse geocoding.
- Source 2 — Google Maps. 144 DFW-area city queries with scroll-to-end result extraction, yielding 1,319 unique places.
- Merge and dedupe. Normalized name plus rounded coordinate pair as the dedupe key. 1,669 records after merge.
- Geographic verification. Latitude and longitude extracted from the coordinate fragment embedded in every Maps place URL, then filtered against a DFW bounding box. 64 records removed as out-of-area — including an Austin shop that matched on a duplicate suburb name and one record geocoded into the Pacific Ocean. 100% of retained records are coordinate-verified.
- Classification. Tiering driven by the platform's own category field where present, with name-pattern fallback. National chains, grocery bakery counters, and non-bakery businesses routed to tier X.
- Exclusion. Nine Golden Glaze locations identified and removed from the outreach population.
B. Known data gaps
- Phone coverage: 325 of 1,669. Per-place enrichment was interrupted by platform rate limiting. Resolvable in a single clean pass with a paid Places API key (~$40 one-time for the metro) — decision outstanding.
- Email coverage: 51 addresses across ~45 businesses. Only 185 shops have any website; all were crawled exhaustively. This is a ceiling, not a gap — see Section 8.
- Owner names: not captured. Requires field work or a paid data source.
- Current supplier: unknown. Best captured through the "send me your last invoice" motion.
C. Assumptions requiring validation before external use
- Addressable spend per shop per year ($45–60K modeled) — replace with Golden Glaze actuals
- Gross margin at our purchasing volume (18–22% modeled) — replace with live supplier quotes
- Account ramp rate and year-one capture (40 accounts modeled) — unvalidated until the pilot cluster runs
- Incremental warehouse cost within the commissary — depends on final kitchen scope and layout
- Franchise supply restrictions — assumed prohibitive; not individually verified per brand