Food loss and waste (FLW) trends: metrics, definition, and operational change in LAC restaurants

What are food loss and waste (FLW)?

We call it FLW: every kilo your house loses between what it buys and what it bills—raw stock, half-finished product, plated food that never left the pass—weighed by shift or converted into a share of food cost.

Hold on to one distinction almost nobody gets right, because responsibility for the money follows it: LOSS happens upstream, at harvest or in the distributor's warehouse, while WASTE happens behind your own doors. World Bank series put the region at 12 % to 18 %, whereas a house running a measurement system sits at 4–6 %. Six to twelve points of difference, and not one of them carries its own name on the income statement. Measured and sorted, that figure is what development banks eventually read as risk. A single percentage describes nothing, and that is where most dashboards I review go wrong. The loss spreads out—receiving 40 %, storage 15 %, kitchen 30 %, service 12 %, customer discard 3 %—and whoever writes down '15 % overall' and sleeps well is deciding blind for months.

Why FLW is not a single number: source and operational anatomy?

Put cash on it: USD 50,000 of monthly food cost at 15 % means USD 7,500 walking out unbilled.

Audit seven days, find that 45 % of it falls at receiving because the supplier delivers off-spec, and you will see that no twelve-month strategic plan is required—swap the supplier, inspect the packaging, sign a quality agreement. Thirty days later that control point is 5–8 points lighter. Across 8,400 accounts, Masterestaurant has seen each recovered point hand back 0.8 points of net margin. Until 2023, in informal markets the number came out of the manager's eye: some range, written down without method, and on to the next thing. That is why one three-unit chain would report variance from 5 % to 25 % while nobody could say whether the culprit was the receiving dock, the grill or the walk-in. With operational intelligence running, the picture stops offering opinions: weight in, weight accepted, weight down to the kitchen, weight on the plate, weight in the bin.

Before and after: how operation changes with measurement

The baseline gets written, and an alert fires the same day the figure drifts more than 2 points off the monthly average. What really changes is the clock. Without alerts the anomaly shows up at close, four weeks eaten already; with a system you hold it in 24 hours. The IDB, the World Bank and CAF all demand that discipline before releasing preferential credit to an MSME. Eight rate points separate two identical houses, and a notebook is the only thing standing between them. Hand a banker opaque margins and he has no option but to discount all of them as risk—16 % to 22 %, with the term capped at 24 or 36 months—and he is right to do it. Lay ninety days of clean records on that same desk and money reprices to 10–14 % with terms stretching to five years, because he is no longer underwriting your charm but your operational control.

Impact on margins, credit access, and labor formalization

Colombia confirmed it under the IDB and Banco Bogotá refinancing between 2023 and 2024: arrears in the participating group dropped from 18 % to 8 % across two years, traceable to documented improvement. Then comes the part that matters outside the register, since freed margin usually goes to payroll; ECLAC and the IDB measured in 2025 that those houses move from 2–3 informal hands to 4–5 employees with benefits, formalizing with 34 % higher probability inside eighteen months. Nobody ever tells the grower who supplies your berries how much of the crate comes back, so he harvests exactly the same way year after year. That silence carries a price paid in the field long before anything reaches your dock: same volume, same packaging, same share of bruised fruit, and a chain that never reaches scale. Break the silence and everything moves. Share the rejection rate by spec, by packaging and by transport damage, and you will watch the grower recalculate volumes, change the crate, negotiate the cutting date with you instead of against you.

Short supply chains: without FLW measurement, they do not exist

The IDB's #SinDesperdicio initiative documented exactly this across Colombia, Ecuador and Peru: fruit growers who receive feedback from their buyers cut 6 to 9 percentage points of on-farm loss in the first year. Less loss upstream and less waste downstream leaves margin for both sides. Start with what stays OUT, because that is where half the baselines I audit fall apart. Whatever disappears at harvest or on the truck belongs to the grower and the distributor, not to you. A badly costed recipe does not belong here either: that is a plating standard problem, and it gets fixed in a different notebook. Neither does a weak forecast, because ordering 20 kg of tomatoes and selling half is bad buying rather than operational waste, however much it stings at the register.

What FLW is not: common interpretation errors?

What does count you can point at with a finger:

the kilo turned away at the dock because it arrived bruised, the one that rots in the walk-in for want of FIFO rotation, the ounce of protein a clumsy cut sends to the bin, the plate that comes back because it misses standard. SATE Institute measures those five control points and nothing else. Blur the categories and your baseline is born false, which sends every improvement dollar to the wrong place. One error repeats itself with almost comic fidelity: the house builds the system to win the loan and lets the log die the moment funds land. The whole play gets thrown away there, because what they just abandoned was never a banker's requirement but the one dataset that organizes tomorrow's decision. Whoever keeps it alive knows which point holds 80 % of the loss, and that knowledge fixes the return on every dollar—receiving at 45 % means audit the supplier, kitchen at 35 % means train the chef or rebuild the plating standard.

Measurement as asset: from opacity to competitive edge

Masterestaurant has watched houses that sustain daily measurement for 24 months reach 3–4 %, when the international benchmark for a formalized operation barely touches 8–10 %. Suppose your neighbour copies your entire menu tomorrow: he still buys with 12 % shrinkage. That is the moat, and in markets where 90 % never weigh anything, no menu reproduces it. Development agencies read this indicator with two eyes at once. One watches SDG target 12.3, which asks for less loss and waste along the food chain, and that is the environmental axis everyone recognizes. The other watches decent work and poverty—SDG 8 and SDG 1—because recovered margin ends up paying formal wages, buying stability, keeping people who would otherwise leave. ECLAC and the IDB attached a number to it in 2025: a house entering an M&E program with MSME refinancing formalizes within eighteen months at 34 % higher probability than its control group.

FLW and development targets: SDG 12.3 with credit impact

Here the false tension between green and profitable resolves itself, which is why the IDB treats the metric as an economic CONDITION rather than an ecological whim. For you the translation is concrete and gets signed in an office: 10–12 % instead of 16–20 %, sixty months instead of thirty-six. **Loss source visibility:** opaque waste is all you own until the system goes in. Once it does, every kilo gets an owner—the dock carries 40 %, the kitchen 30 %, the walk-in 15 %, the pass 12 % and the guest barely 3 %—and the improvement budget aims at one place instead of scattering into gestures. Put a competent hand on the receiving dock and 5–8 points come off; fix stock rotation and another 3–4 follow. **Short supply chain:** your grower harvests and packs blind as long as nobody sends a number back, and that silence manufactures waste in the field before the truck even loads.

The 5 most critical operational differences

Quality agreements—what rejection rate is tolerated, what packaging spec holds—come out of receiving data, and with those in hand the farm moves volumes. Colombia has it measured: fruit growers inside an SSC with restaurant feedback cut 6–9 points of on-farm loss over twelve months. **Margin liberation and reinvestment:** shrinkage that never shows up as a cost line still takes 2–8 margin points on its way out. Clean it up and those points surface, and then the house chooses: lift wages 25 %, go formal, or open another door. The multiplier is real, since a payroll that adds up carries 4–5 employees with benefits where it once carried 2–3 off the books. **MSME credit scoring:** development banks treat this M&E as a risk lever, not an environmental courtesy. Their Central American refinancing program cut arrears among participating restaurants from 18 % to 8 % over two years (2023–2024), and documented improvement is what did it.

The 5 most critical operational differences — in practice

Skip the measurement and credit is either absent or predatory. **Remediation speed:** four weeks of margin vanish when the anomaly surfaces at month-end close, which is exactly when it surfaces if no alert exists. A live system flags it inside 24 hours, and that same afternoon somebody checks the supplier, the receiving protocol or the kitchen standard. Remediation costs 70 % less and the financial hit stays under a single day.