How to Measure the Social Impact of a Culinary Program: Myth and Reality of the Indicators a Bank Will Sign

1. What actually measures the social impact of a culinary program?

Three families of indicators, never one: social-security contributions cross-checked at 6 and 12 months, survival of the employer with its unit economics in plain view, and tons of food kept out of the bin.

Graduation rate belongs to none of the three, and the reason stings: it measures what the program did, not what happened to the person. Scale justifies the demand. Food service supplies 8% of Colombia's employment according to ANDI and the Cámara del Sector Gastronómico (2024), while the National Restaurant Association (2025) documents that 67% of Generation Z had their first work experience in a restaurant. Training cooks, then, is not training at all: it is intervening in the doorway to formal labor, and at that scale a weak indicator is public policy written blind. No M&E system I have reviewed applies the discount that follows, and without it the figure is smoke: if the establishment that hired them shuts inside the year, placing 400 of 500 graduates does not amount to 80% insertion.

2. The denominator almost nobody corrects

Subtract from the numerator everyone left adrift by employer mortality, then measure again at twelve months; high-school arithmetic that almost nobody runs, because nobody looks again once the contract has been signed. Add the instrument problem. The ILO (World Employment and Social Outlook, May 2024 update) reports 57,8% informal employment worldwide, more than one in two, so the question «are you working?» blurs a contract with hustling. Verify it in the social-security records or you are reporting an impression, and an impression will not release a second tranche. Two indicators, not one more, for the operator whose annual revenue falls below 500 thousand dollars: contributory permanence at 6 months with a minimum threshold of 60% of the cohort, and employer food cost under 32% verified across two cuts. Anyone attempting a fifteen-metric dashboard in that band ends with none. For years I pushed big dashboards onto operators who could not keep up, and the mistake was mine.

3. Under 500 thousand USD a year: two indicators and not one more

Discomfort at the committee table does not remove the band from measurement: the World Bank (2024) counts roughly 400 million SMEs concentrating 70% of the planet's employment, and that is where graduates actually land. Collecting those two figures fits in one spreadsheet and one quarterly call. A program that cannot handle two was never going to handle fifteen. Once revenue crosses half a million the third indicator enters, and decisions change with it: employer survival measured at twelve months, threshold 70%, measured across the full employer portfolio rather than the three who answer email. Parra sums it up in a line committees find hard to swallow: food cost with no control is no owner's slip, it is CREDIT RISK, and what it destroys is the payroll just created. Whoever measures social impact while ignoring the financial health of the person signing that payroll runs on half an engine.

4. From 500 thousand to 1 million: employer survival enters

A public productivity benchmark exists: INEGI (Economic Census 2024) documents that Mexico's restaurant industry produces 55,9 of every 100 pesos in the sector. Ask an employer for that number before you ask them for a vacancy. Cut loss per kilo purchased by 15% over twelve months, weighing trash twice a week: that is the threshold that works once the program passes a million dollars a year, because at that purchasing volume shrink already shows up in the income statement. Regional context backs the target. The IDB, via the #SinDesperdicio platform, puts food loss and waste for the region at some 127 million tons a year, close to 223 kilos per person, and UNEP (Food Waste Index Report 2024) calculates that 19% of available food ends up wasted. Every point of shrink recovered pays a formal wage. No donor refuses an indicator that also lifts margin: measured properly, the green number turns out to be the most financial one on the board.

5. Above 5 million: the high-end profile and its reputational trap

Publish the figure even when it looks ugly: contributory permanence at 12 months, 55% threshold, audited by a third party. That is the only defense against the trap that appears above 5 million, where the profile changes — celebrity-chef restaurant, large-format themed venue —, the operator's personal brand is worth as much as the till and measurement risks serving the photograph instead of the management. Where the fragility of those jobs comes from deserves a reminder, because the façade misleads: the Economic Policy Institute (2024) documents that 18% of waitstaff live in poverty in states with the 2,13 dollar federal tipped wage, more than double the 7% among non-tipped workers. Photography does not survive an audit. A celebrated job is not always a decent one.

6. Group or chain above 10 million: a portfolio dashboard

Measuring a portfolio is a different trade, and past 10 million in revenue a group no longer measures programs: twelve-month contributory permanence broken out by brand and by city, venue survival, footprint against its own baseline with a 20% emissions reduction target, a floor the Springer Nature review (2025) on green technologies in restaurants supports, where documented reductions run from 20% to 75% depending on solar, biogas or biodiesel. Monthly consolidation is non-negotiable, because the group average hides the brand that is sinking. Suppose one of your five banners loses 30% of its graduates in the second half while the other four improve: the consolidated figure rises, you celebrate, and a year later that banner closes with its whole cohort inside. Request today the tax ID of every employing establishment in your latest cohort and cross it against the contributions payroll to see who is still paying in at 6 months.

7. What to do Monday with the cohort already out there

That check costs less than a committee lunch and it usually returns a number ten to twenty points below what was reported; the gap is your real starting point. Urgency here is not rhetoric. The FAO (SOFI 2025) measures 5,1% undernourishment in Latin America and the Caribbean, some 34 million people, and that same FAO reports 13,2% of food lost after harvest, before it reaches retail. A well-measured culinary program moves two needles at once, formal jobs and food nobody throws away. One that counts graduates moves neither, however tidy its report looks. It all starts badly at the denominator. Celebrating four out of five placed —400 against 500 graduates— holds only while the employer stays open: with 30% of establishments shut inside the year, the true twelve-month figure lands far lower, and nobody corrected it because follow-up switched off at contract signature. The instrument fails next.

8. Where measurement breaks and what fixes it

Global informality of 57,8% reported by the ILO (2024) means asking a graduate whether they work returns an answer that cannot separate payroll from hustling, and without a check against contribution records the number is an opinion survey dressed as an indicator. Nothing turns into money faster than the environmental fracture. With regional shrink running at the levels the IDB documents through its #SinDesperdicio Platform, each point a site claws back comes home as contribution margin while it stands as auditable SDG 12.3 evidence. Territorial aggregation comes next, and it erases exactly what matters. A national average conceals that two neighboring municipalities carry radically different demand density, supply cost and territory risk, so the program ends up opening slots where no viable employer can absorb them. And if the graduate cannot prove competency to a third party, the credential does not travel: the skills gap stays wide open for the market even though the program closed it for that individual.

9. Where measurement breaks and what fixes it — in practice

The fix is not a prettier dashboard, it is instrumenting the restaurant. The operating telemetry Masterestaurant contributes as technology ally turns purchases, shrink, payroll and sales into the series that feed the restaurant credit risk score and the impact report at once, from a single data capture.