How to measure gastronomic social impact measurement: the errors that void the data and the method that survives an audit

What separates an activity metric from an outcome metric?

An activity metric counts what you did; an outcome metric counts what changed in someone's life, and that difference decides whether a fund signs or shelves your portfolio.

Four thousand two hundred «trainees» is pure activity, while verified job placement at six months is an outcome, and that is precisely where most gastronomic portfolios in the region collapse. The background figure that makes the distinction urgent: according to the ILO, roughly 6 out of 10 employed young people in Latin America and the Caribbean work informally, so placing a graduate in a kitchen proves nothing if no contract exists. ECLAC measured 46,6 % labor informality in 2024, concentrated in micro and small firms, which is the universe where 90 % of independent restaurants operate. Count formalized jobs, never course attendees. Capture your baseline BEFORE the first participant touches a stove, because data gathered afterward is not a baseline but a memory reconstruction, and any serious evaluator spots it when input variables correlate suspiciously well with final results.

The baseline expires on opening day

Nine imperfect variables measured on day zero beat forty refined ones collected in month three. Inside a running restaurant those nine come free from systems you already own: monthly staff turnover, training hours per person, waste by station, share of purchases from local suppliers, food cost by product family. The National Restaurant Association places healthy food cost between 28 % and 35 %, so a venue starting at 38 % and landing at 31 % owns an efficiency story that is accountable and auditable. Without that opening photograph, the improvement belongs to nobody. The investment officer never asks whether your 61 % placement rate is true; the question is how much of that 61 % would have occurred regardless, and answering it demands an explicit counterfactual. A comparison group of waitlisted applicants works, a same-age same-neighborhood cohort without the intervention works, even a historical series from the same territory works. Gross numbers alone do not.

Attribution: how much of that 61 % would have happened anyway

With youth informality near 60 % across the region (ILO), some of your graduates were going to land precarious work with no program at all, and your value lives in the gap between that destiny and the formalized job you produced. At Masterestaurant, Diego F. Parra insists on printing the net figure on page one of the report, however much it stings: a gross 61 % that resolves into 23 net points is an excellent, defensible result. A 61 % without a counterfactual is propaganda. Pick ONE indicator for every outcome you declare and resist the twenty-metric battery, because a restaurant crosses at least three SDGs over the same cash register and measuring them separately yields incompatible reports. SDG 8 is measured in formal jobs sustained at twelve months. SDG 12 is measured in kilos of waste per thousand dollars of sales. SDG 9 is measured in local suppliers brought into the chain.

One indicator per outcome, or three reports that never talk to each other

Three numbers, three operational data sources your point of sale and digital payroll already generate. FAO estimates that roughly one third of food produced for human consumption is lost or wasted, and the UNEP Food Waste Index 2024 put household waste at 631 million tonnes in 2022, some 60 % of the total; food service contributes its share at the chain's final point. Your kilo of waste is environmental data and burned food cost at once. Measure waste first, because it is the only impact indicator that pays for itself within the same month you start tracking it. Food loss and waste account for 8-10 % of annual global greenhouse gas emissions according to UNFCCC and FAO 2024, carrying an associated cost around one trillion dollars a year. And a commercial kitchen shows a carbon footprint 2 to 5 times larger than other spaces, while food service concentrates 18 % of the food-linked footprint (Springer Nature, 2025).

Money and climate: why waste is the most profitable indicator to track

Translated into cash: every food cost point you recover on monthly sales of 120.000 dollars is 1.200 dollars returning to margin. That same point is CO₂ tonnage you can report. Environmental measurement stops being a compliance expense and starts funding its own data infrastructure. The regional problem is not missing data, it is that nobody designed the indicator before piling data up. A restaurant with point of sale, inventory control and digital payroll produces at zero extra cost the complete series an evaluator needs, with daily granularity and three years of traceability. An uncomfortable tension of the trade deserves an admission here: the same owner demanding scientific rigor from an external evaluator keeps inventory in a notebook and turnover in his head. Architecture solves it, not budget. Define the fields, freeze the definitions in writing —what counts as formal employment, what counts as avoidable waste— and leave them untouched through the measurement period.

Clean operational data or no evaluation at all, and that is where almost everyone fails

The WFP reported 84.000 million dollars of global school feeding financing in 2024, with 99 % coming from national budgets, and that money reaches whoever can prove what they do. Run the scenario to its end: if your program jumps from 400 to 4.000 participants using the measurement framework you have today, you do not get ten times the evidence, you get ten times the noise, because a poorly defined indicator scales its error alongside its coverage. Four thousand «workshop attendance» records still fail to say whether anyone's life changed. Worse, the cost of rebuilding the baseline retroactively grows proportionally until it can no longer be paid. With 140 million informal workers in the region according to the ILO, and food insecurity reaching 13,7 % of United States households in 2024 —47,9 million people, USDA ERS— the social problem is big enough that the urge to grow fast feels legitimate.

What would happen if your program grew tenfold on today's framework?

Even so, my recommendation is hard and I will not soften it: never scale a program whose indicator has not survived an external evaluation with a comparison group.

First: 28-35 % food cost (National Restaurant Association). Concrete action: measure your starting point this week by product family rather than in aggregate, and set the reduction target in percentage points over that dated baseline. Second: 8-10 % of global greenhouse gas emissions come from food loss and waste (UNFCCC/FAO 2024). Action: convert your monthly waste kilos into tonnes of CO₂ equivalent using a published factor and report it beside the dollar savings, in the same table. Third: 46,6 % labor informality across the region's MSMEs (ECLAC 2024). Action: report FORMALIZED jobs at twelve months with affiliation numbers, never «trainees». Start tomorrow with the waste baseline, the only one of the three that hands money back before the quarter closes.

Four differences an investment officer settles in ten minutes

First comes timing. A baseline captured after launch is not a baseline, it is a memory, and any seasoned evaluator spots it the moment entry variables correlate suspiciously well with outcomes. Sequence beats precision here: 9 imperfect variables captured beforehand outweigh 40 refined ones captured in month three. Attribution is the second. When a youth employability programme in hospitality reports that 61 % of graduates found work, the investment officer does not ask whether the figure is true; the question is how much of that 61 % would have happened anyway. Regional evidence from control-group evaluations places net attribution well below the gross number, frequently between a third and two thirds of the observed effect, and passing gross off as net is the fastest route to losing credibility with the fund. Third, formality. Counting «jobs» without opening the black box of formality, retention and wage leaves the indicator hollow against SDG 8, whose core is DECENT work, not mere occupation.

Four differences an investment officer settles in ten minutes — in practice

An informal three-week post adds to the report and subtracts from development. And the fourth moves the most money: the link between the social indicator and the cash register. Cutting FLW from 11.4 % to 6.8 % of purchases is not an environmental figure, it is recovered margin financing the formal payroll SDG 8 demands. Diego F. Parra returns to this point in every Masterestaurant diagnostic, and the mechanism is plain: impact that never touches the income statement gets cancelled in the first cash-flow crunch.