What a restaurant needs to receive outside investment: the mistakes that kill the round and the method that closes it

1. What does an investor look at before the concept?

They look at the cash return of the unit that ALREADY operates, and they want it paid back within 24-36 months, backed by twelve months of auditable numbers.

Concept, the line at the door and social recognition come later, once the file has proven the capital comes home. That the money exists is not in dispute: the U.S. Small Business Administration (2024) reports accommodation and food services as the most financed industry in the SBA 504 program, at 16,5% of fiscal-year loans. The cheque was never the scarce part; what is scarce is the operator able to hold an EBITDA figure month after month without restating it three times. And it helps to say what a funding round is NOT: it is no prize for good cooking, it is the purchase of a cash flow that somebody else will audit closely, with judgement that is not yours.

2. The file that carries weight: twelve months of normalized EBITDA, not gross sales

Valuation anchors to NORMALIZED EBITDA, and normalizing almost always lowers the number the owner carried in his head. Normalizing means putting the owner's market salary back into the P&L, taking out the personal expenses that slipped through the books and adjusting related-party rents to market price. A unit reporting an 18% margin usually lands at 11% or 12% once that exercise is done honestly, and a clean 11% raises capital far sooner than an 18% that collapses at the first question. Cost pressure explains why the bar moved up: the National Restaurant Association (2024) documents food costs up roughly 35% and labor up another 35% since 2019. With that structure, an inflated margin no longer survives a single week of serious due diligence. With annual revenue below 500.000 USD, the right call is secured debt or capital from close partners, never an institutional fund.

3. Under 500 thousand USD a year: this is not a round, it is debt

The reason is arithmetic: a unit in that band running 10% EBITDA generates 50.000 USD a year, and no professional investor builds a 15.000 to 25.000 USD legal structure to take a piece of that. Square (2024) puts opening a QSR or food truck under 150.000 USD, a figure that fits inside a bank line without diluting a single point of ownership. The threshold for this band is plain: twelve consecutive months of positive EBITDA above 8% and a debt-to-EBITDA ratio under 3x. Whoever falls short there does not have a capital problem, they have an operating problem, and pouring money on top only speeds up the loss. In this band outside capital starts to make sense, but only to open the second unit, never to plug the cash hole of the first. What the investor buys is proof that the model travels: same margin, same ticket, different address.

4. From 500 thousand to 1 million: the threshold is the second unit, not the first

Before seating a partner, the Masterestaurant method asks for 18 months of the original unit at normalized EBITDA of 12% or better, prime cost stabilized under 62% and monthly food cost variance below 2 points. On that base, operating savings become demonstrable: TimeForge (2025) documents labor cost reductions of 8% to 12% with AI-assisted scheduling and forecast accuracy above 90%. If the second unit is financed with 60-month debt and projected payback runs past 36, the number is telling you to wait. Past a million in revenue, what gets assessed is no longer margin but founder dependency, and that is the question that kills the most rounds: does this result hold if you step away for three months? Picture ninety days out. If food cost swings more than 3 points, if kitchen turnover spikes and the ticket drops 8%, the asset you are selling is you, not the restaurant, and nobody buys a job.

5. Above 1 million: the question changes, and it stings

Here the file demands operating manuals, a second-in-command with purchasing authority and KPIs signed by someone who does not share your surname. The National Restaurant Association with BLS (2024) counts 6,2 million Americans aged 16 to 19 in the workforce, 900.000 more than in 2019: hands are available, what is missing is the system that makes them interchangeable without the result wobbling. In this band the investor stops watching percentage margin and watches average unit volume, because it is the only thing that lets him project twenty units without drawing them. Market benchmarks are public: Wingstop reports an AUV of 2,13 million USD (FDD 2025) and Cava approaches 2,93 million USD per unit according to Technomic (2025). Here the high-end profile shows up —the media-chef restaurant, the large-format themed venue— and with it the trap: these are businesses whose AUV rides on one person or one novelty, with traffic curves that fall 30% or 40% from year two onward.

6. Above 5 million: AUV rules, and the celebrity case is the trap

A serious fund values them at a punished multiple, 3x to 4x EBITDA against the 6x-8x of a replicable brand. My position is firm: if your value thesis fits on one face, do not sign it. Above 10 million in consolidated revenue, the bottleneck moves from money to governance, and that is the paradox almost nobody anticipates: the more capital available, the heavier what money cannot buy. A group at that scale needs financial statements audited by a recognized firm, consolidation by business unit, a committee that meets monthly and an operations director accountable for prime cost across the network. The threshold for an institutional round in this band is 4 to 6 mature units with consolidated EBITDA between 12% and 15%, plus a new unit reaching break-even before month 9. Pricing discipline belongs in the file too: ACODRES (2025) reported in Colombia a 9,8% rise in dish prices since February to sustain 98.000 jobs, and a group that fails to pass cost through decapitalizes quietly.

7. What to do on Monday, without waiting for anyone

Close the normalized P&L of the last twelve months this week, month by month, with your market salary inside and personal expenses out. That document is worth more than any deck, because money does not punish low profitability: it punishes the VARIABILITY the owner cannot explain. A unit steady at 11% EBITDA raises capital sooner than one swinging between 4% and 19% with no identified cause, and that contrast is what Diego F. Parra has watched repeat across twenty years sitting in both chairs, the operator who needs the cheque and the board that signs it. With food and labor 35% higher than 2019 according to the National Restaurant Association (2024), decision architecture weighs more than promised growth. Start with your worst month and explain it with numbers. Operators sell potential; investors buy repeatability. One unit returning capital in 30 months, provable on paper, is worth more than three promised units with an internal rate of return drawn on a spreadsheet nobody audited.

8. Where the conversation with capital breaks

The file is not accounting. Accounting answers to the tax authority; operational due diligence answers a different question — does this result hold if the founder disappears for three months? Two documents, two purposes, and confusing them costs the round. Restaurant valuation anchors to normalized EBITDA, not gross sales. Normalizing means putting a market-rate owner salary back into the P&L, stripping personal expenses that crept in and adjusting related-party rents; the number almost always drops, and credibility almost always rises. Territory risk gets quantified or it gets discounted. When unit two opens eight blocks from unit one, cannibalization either appears in the model or the investor prices it themselves, with a penalty far harsher than reality. A celebrity-chef restaurant above 5M USD a year is not evaluated like a fast casual: image royalties, dependence on one personal calendar and the cost of maintaining the production are concentration risks that demand specific clauses, not a generic multiple.