Masterestaurant Analysis of Restaurant Cost Stress Scenario Simulation 2026: why only 34 of every 100 businesses reach year five

Finding 1 — The figure that frames everything else: 34 out of 100

Thirty-four. That is how many of every 100 companies created in Colombia are still open at year five, and Confecámaras documents it in the report Bloomberg Línea picked up. Almost none of the two thirds that vanish died from a spectacular blunder: a run of small increases nobody added up in time killed them, one invoice at a time. I read that number as the floor under any serious cost conversation, never as a mortality statistic. The industry employs 10% of the U.S. workforce, according to the National Restaurant Association (2024), so real payroll sits behind every closure. Stress scenarios put a date and a number on the increases already heading your way, weeks before the invoice lands. A cost shock arrives chained because the three lines that move margin (inputs, payroll and channel commission) answer to the same cycle and tend to shift inside the same quarter.

Finding 2 — Why does a cost shock almost never arrive alone?

The ingredient climbs first. Then the year's wage adjustment lands, and on top of it the platform moves its commission two points over off-premise operations that already carry roughly 75% of traffic, according to Circana.

Separately you absorb any of the three; together they take the whole contribution margin without anyone committing a visible fault. The owner reviews the P&L, finds no culprit, and blames sales. Wrong culprit. What this synthesis does is put a number on that chain before you live it: how far your signature dish bends if the ingredient jumps 18%, what survives if payroll adds 9% in the same quarter. Six or eight weeks of head start. That, and nothing beyond it, is what a scenario simulation buys, and in this trade the head start is the entire room to maneuver. Skip it and you meet the shock when the invoice arrives, with the supplier aware you have no alternative and the menu already printed.

Finding 3 — The difference is not in the spreadsheet, it is in the calendar

Run it and you got there first: volume renegotiated, two dishes redesigned, recipe card corrected, portion weight moved without touching the guest-facing price. Let me commit on the order of the levers, because the majority reflex is to raise price first and price is the last resort. Fewer than 4% of Latin American and Caribbean firms use artificial intelligence against more than 20% in Europe, according to ECLAC (2024): getting ahead stays cheap and stays unclaimed. Counterintuitive as it sounds, the single-venue operator is the one who most needs this exercise, and the arithmetic proves it without poetry: the ten-unit group absorbs one location's blow with the margin of the other nine, while the single venue carries 100% of it against its own till. The industry sold simulation as a multi-unit luxury with a staff analyst, so the small operator writes it off. That asymmetry explains a good share of why only some 34 out of 100 Colombian companies reach year five, a figure from Confecámaras via Bloomberg Línea.

Finding 4 — The single location needs simulation MORE than the ten-unit group

Nobody has the time, I know. Even so, four scenarios on one sheet, three critical inputs and a range of increases capture 80% of the value in an afternoon. Look for margin in the bin before you look for it on the menu. ReFED (2025) attributes 70% of foodservice waste to food the guest never ate, and ReFED (2024) puts more than 43% of U.S. surplus on full-service restaurants. Translated into the operation: the problem lives less in the kitchen than in portion size and plate design. When the scenario shows protein climbing 15% with no way around it, the rational response is to check portion weight against what tables actually consume, not the selling price. Portion weight is a design decision, never an accident of the line: fix it on the recipe card and it holds across every shift. Those margin points cost you no customer at all, and few points that free are left anywhere on the menu.

Finding 5 — The social cost of a closure: a broken ladder

Here is an argument that never reaches the P&L: every restaurant closing under cost stress breaks an entire ladder of mobility. Eight in ten owners in the sector, and nine in ten managers, started in an entry-level job, according to the National Restaurant Association (2026), which makes this trade one of the few advancement routes still working without a university credential. Add the ILO estimate that women make up 60% to 70% of workers in hotels, catering and tourism, and Spanish hospitality closing 2024 with 54.3% women, according to the Anuario de la Hostelería. A venue that absorbs the shock keeps the ladder standing for the next cohort of dishwashers, hosts and line cooks who would otherwise never get the first rung. So we read simulation through SDG 8: holding the till holds the ladder. No figure in this analysis comes from a sample of ours, and that is worth saying up front.

Finding 6 — Where these figures come from and what this synthesis adds

Confecámaras, ReFED, ECLAC, the ILO, the FAO and the National Restaurant Association publish the numbers with their year and their verifiable link; we supply the reading. What the synthesis adds is the ORDER in which those sources get read and the threshold at which each one forces a specific lever: when waste rules, when the recipe card rules, when nothing is left but price. And the scarcity context weighs: 181.9 million people in Latin America and the Caribbean cannot pay for a healthy diet, according to the FAO, while between 638 and 720 million went hungry in 2024, according to the SOFI 2025 report. Three inputs, three ranges, eight dishes: that is the whole start. Move the families that concentrate most of your raw material by 10%, 20% and 35%, then cross them against the real costing of your eight highest-turnover dishes. The exercise fits in one afternoon and tells you which dish blows first, and it is rarely the one you expect.

Finding 7 — How to build your first simulation this week

Next comes the year's foreseeable payroll adjustment and the extra commission point from platforms, with roughly 75% of traffic already off-premise, according to Circana. What emerges is no forecast; it is a map of thresholds. With fewer than 4% of the region's firms using AI, according to ECLAC (2024), one well-built sheet already puts you ahead of the shop next door. Build it before the next price list. The change does not happen in the spreadsheet, it happens on the CALENDAR. Whoever skips simulation meets the shock when the invoice lands; whoever runs it saw the shock six or eight weeks out and used that time to renegotiate with the supplier, redesign two dishes and correct the recipe card without moving menu price. That head start is all the room this business gives you. A real tension deserves a straight answer. Stress simulation looks like a luxury for multi-unit groups with an analyst, so the single-venue operator drops it; the truth runs the other way.

Finding 8 — What actually changes between not simulating and simulating

A ten-venue group absorbs one location's shock with the margin of the other nine, and the single venue has no such net. Simulation matters MORE the smaller the business, however the industry has sold it. Against the widespread assumption that price is the lever: 181.9 million people cannot afford a healthy diet in Latin America and the Caribbean, according to FAO (SOFI 2024), and that ceiling is hard across much of the region. A scenario resolved by raising price and nothing else is badly built. Order matters, and here I commit: input shock first, labor second, channel last. Whoever opens with delivery commission ends up arguing with a platform that will not cut its take, leaving food cost untouched when food cost is the thing actually under control. Food cost per dish should never pass 32%, the ceiling rather than the target; payroll, rent and utilities stay off the dish and belong to break-even.

Finding 9 — What actually changes between not simulating and simulating — in practice

A shock handled well preserves payroll, and that is where micro-operations touch the macro indicator. Nine in ten managers reached the job from an entry-level position, and eight in ten owners did the same, according to the National Restaurant Association (2026), so every venue that holds sustains a ladder a closure erases at one stroke. Parra keeps pressing a point owners resist: simulation does not predict the future, it bounds the damage. Nailing the exact percentage of the hike is not the goal; knowing which of the three levers breaks first is. Masterestaurant built its framework on that idea, and the Restaurant Model Canvas is where the scenario notebook gets organized.