Restaurant business model: before vs after with Masterestaurant

Step 1 · Cost the dish before touching a single piece of marketing

Redesigning the model starts with the recipe sheets for your ten best-selling dishes, not with the editorial calendar. Take each dish, weigh the ingredients in real grams —waste included, because that is where 3 to 6 points of food cost hide— and divide that cost by the pre-tax selling price. Anything above 32% gets flagged red; that 32% is the CEILING in the Masterestaurant costing contract, never the target. Payroll, rent and utilities do not belong on the plate: they live in the break-even calculation, and mixing them is why so many owners believe they have margin when they do not. DELIVERABLE: a sheet with ten rows, a cost column, a price column, a percentage column and a contribution margin column in currency. Verify it by adding the margin weighted by last month's units sold; if that sum fails to cover your fixed costs, the model is broken and no Reel will fix it.

Step 2 · Write the value proposition as one sentence answering three questions

A useful value proposition fits in a single sentence and answers for whom, at what moment and for how much. «Working lunch for downtown office staff, served in under 18 minutes, between 28,000 and 34,000 pesos» is a proposition; «home cooking made with love» is decoration. Here is the acid test I apply: if the competitor around the corner can copy your sentence word for word and still be telling the truth, the sentence is useless. That moment of consumption matters more than it looks, because customers already split their week across channels: 47% of US adults order takeout weekly (Escoffier, 2025) and over 40% order delivery or takeout three to five times a month (UpMenu, 2024). DELIVERABLE: the sentence written down, posted in the kitchen and at the point of sale. Verify it by asking three employees separately who the restaurant is for; three different answers mean you rewrite it.

Step 3 · Pick the channel by acquisition cost, not by vanity

The right channel is the one that brings you a cheaper table, and that comparison runs on numbers rather than preferences. Delivery platforms in Latin America charge between 18% and 30% of order value, a toll that on a dish carrying 30% food cost swallows nearly the entire contribution margin; and the volume they move is real —iFood processes roughly 60 million orders a month, per Sacra 2025— so the channel was never the problem, the price you pay for it was. Your own video content works differently: fixed cost. Shoot four pieces a month with a phone and a tripod, and every extra table arriving through it drops the unit cost. Commit to TWO channels, not five. DELIVERABLE: a table listing channel, monthly cost, attributed tables and cost per table. Verify it when you can point a finger at which of the two channels runs more expensive. Attribution gets installed with one question at the register and one checkbox in your software, and that is the entire infrastructure you need to stop funding a channel blind.

Step 4 · Install attribution before publishing anything

Train the cashier to ask «did you see us on social?» at every check close, log the answer as yes, no or no reply, and within three weeks you will hold a sample big enough to split your tables by origin. A restaurant serving 400 covers a week accumulates 1,200 records in that window, plenty to read a trend. I got this wrong for years, recommending coupons and promo codes nobody redeemed, when the direct question performs better and costs nothing. DELIVERABLE: a weekly percentage of tables attributed to content, plotted against production spend. Verify it by crossing that percentage with total period sales; if content claims 20% of tables while sales have been flat for six months, somebody is logging badly. Your editorial calendar derives from the value proposition, never the other way round: each piece must prove one of the three answers —for whom, at what moment, for how much— through a concrete scene inside the restaurant.

Step 5 · Turn the value proposition into a calendar of four pieces a month

Four monthly pieces done properly beat twenty improvised ones, because sustained output matters more than the opening volume almost nobody keeps up past month two. Reserve one of the four for the item with the highest contribution margin in currency, not the highest percentage: a dish at 28% food cost leaving 9,000 pesos earns you less than one at 33% leaving 22,000. And if you sell alcohol, give it room: 46% of operators surveyed by Technomic place the category among the menu's highest-margin ones. DELIVERABLE: four short scripts with date, dish, message and call to action. Verify it by checking that every script names a price or a service time. The mistake I run into again and again is driving more traffic to a badly costed dish, and it is the most expensive of the lot: at 41% food cost, every new table widens the loss instead of shrinking it.

Step 6 · The four mistakes that wreck the redesign, and how to dodge them

Mistake two is measuring reach rather than tables —340,000 views do not cover payroll— and the checkbox from step 4 fixes it. Third comes swapping the value proposition every six weeks because one video flopped; the market needs three to six months to tie your name to a moment of consumption. Fourth is loading rent and payroll onto plate cost, raising prices accordingly and losing traffic for being expensive when the real trouble sat in the fixed structure. Fix them in that order. DELIVERABLE: a one-page note listing the mistakes found in your operation and the date you closed each. Verify it by re-costing the ten dishes thirty days later. The model is right when you can answer four questions without opening a file: what margin in currency your star dish leaves, what sentence defines who you cook for, what percentage of tables arrives through content, and what each of those tables costs you.

Step 7 · How to know the model came out right

Miss one and the redesign is half done. Put numbers beside them: weighted food cost under 32%, attribution measured four weeks running, two active channels with cost per table calculated, and four pieces published last month. The industry is moving —the National Restaurant Association projects US$1.55 trillion in US sales for 2026, and weekly visits climbed to 2.19 per person according to Revenue Management Solutions— yet only the operator with the equation written down catches that tailwind. Diego F. Parra and the Masterestaurant team work this sequence with owners across 43 countries, and the order never changes. Start today with the costing sheet. The first difference is ORDER. The inertia model starts with marketing because marketing is visible; the written model starts with dish costing because that is what decides whether the marketing makes any sense at all. Driving more guests to a dish at 41% food cost means accelerating toward the cliff with better music playing.

The five differences that actually move cash

Second comes ATTRIBUTION. A restaurant that cannot say what share of its tables arrived through content is funding an entire department blind. One question at the point of sale — did you see us online? — plus a checkbox in the software gives you, within three weeks, a number you can hold on to. Third, the VALUE PROPOSITION. It is not a slogan, it answers three questions: for whom, at what moment of the day or of life, and at what price. Once that sentence exists, the Reel script writes itself and servers stop improvising their recommendation. The fourth difference is structural: REVENUE LINES. The same kitchen, the same crew and the same dead hours can carry a virtual delivery brand or a corporate catering service. This is where the conversation with a restaurant investor changes register, because you stop describing a venue and start describing a platform with reusable assets. And the fifth, hardest to accept: READING SPEED.

The five differences that actually move cash — in practice

An owner who learns the numbers 75 days late does not lead, he reacts. Closing food cost and revenue by channel every Monday turns 90-day mistakes into one-week mistakes, and that single change is worth more than any campaign you could run.