Food loss and waste (FLW) step by step: traditional method vs Masterestaurant method

What do you actually gain by measuring food loss and waste step by step?

You gain twenty-nine days of reaction time, and that is the whole matter. A waste log reconciled at month-end hands you a number after the same error has repeated four times;

a per-station count at the end of each shift hands you that number within twenty-four hours, early enough to fix Thursday's purchase order. FAO calculates that 14 % of food is lost between harvest and retail and another 17 % is wasted in consumption, with food services contributing 26 % of that final waste (FAO). On an annual food purchase of USD 40,000 —normal for a venue billing USD 120,000— cutting waste from 9 % to 5 % keeps USD 1,600 in the till. The MASTERESTAURANT method wins this criterion because it attacks latency first and everything else afterwards. LOSS is what falls out of the chain before reaching your prep table: the case of tomatoes that travelled twelve hours without refrigeration, the batch delivered already turning.

Loss is not waste, and mixing them up burns your whole budget

WASTE is what you throw out with the product fully under your control: the trim nobody used, the stock that went past on the night shift. An aggregate log blends both into one percentage, which is why an 8 % waste figure never tells you where to intervene. Step-by-step counting separates them from receiving onward: one is corrected with short supply chains and territorial feasibility of sourcing, the other with menu engineering and portioning discipline. Latin America loses and wastes some 220 million tonnes a year, valued by the IDB at close to USD 150 billion, enough to feed 300 million people. Separating them wins; aggregating them spends where the problem never was. An aggregate percentage is rigorously useless because waste almost never spreads evenly. In most kitchens assessed by SATE Institute, three to five references concentrate more than half of the waste measured in value, so the operator staring at «8 % overall» spreads attention across eighty inputs when four deserve it.

SKU-level granularity versus the aggregate percentage: where the money hides

The log gives you an average; the per-SKU record tells you the protein in your signature dish carries 31 % of the loss in money with barely 6 % of purchased volume. The practical difference is not statistical precision, it is where you will stand tomorrow at six in the morning. And a genuine concession belongs here: for years I recommended starting with the total count, and I was wrong, because the operator gives up before finding anything. Disaggregation by reference wins, no argument. The log costs nothing in licences and roughly nine monthly hours of administrative time, which at a loaded cost of USD 9 per hour adds up to USD 972 a year for work that produces a late number. The step-by-step protocol demands forty minutes of training per person at launch, two scales at USD 45 and about four minutes of recording per shift, meaning close to USD 430 annually between equipment and time.

What each system costs to install, with staff hours honestly counted?

On that basis, any reduction above 1.1 percentage points of waste already repays the investment in a venue billing USD 120,000. One detail matters, and Diego F.

Parra insists on checking it before signing any programme: if the recording does not fit inside the flow of the shift, it does not happen, and a system nobody executes costs more than the log. The protocol wins, under that condition. Two comparable branches, each with annual sales near USD 138,000 and food purchases of USD 44,000, started the same January under different methods. The branch that stayed with the monthly log closed the year at 8.7 % waste against purchases, practically where it began; the one that adopted per-station counting with SKU disaggregation finished at 4.9 %, meaning USD 1,672 recovered and, measured on the other axis, some 2.1 fewer tonnes of organic residue sent to landfill.

Mini-case: two venues in the same chain, the same menu, twelve months apart

That residue deserves its real scale: the EPA estimated 55 million tonnes of CO2 equivalent from food sent to United States landfills during 2020 (EPA, 2023). The venue that measured daily was not more disciplined by character; it saw its own number while it could still change it. It has a solvency problem, and that reading rarely reaches the table where these programmes get approved. That percentage swallows the entire operating margin of a typical gastronomic MSME in the region and pushes it toward informality or closure; when it closes, between four and eleven formal jobs disappear. The social bill does not stop there, because the sector is the doorway into the labour market: 51 % of adults had their first job in a restaurant (National Restaurant Association, 2026) and every dollar spent in the sector contributes USD 2.55 to the national economy (National Restaurant Association, 2024). That is why at Masterestaurant we treat food loss and waste as an SDG 8 indicator as much as an SDG 12 one.

A restaurant wasting 9 % of its purchases does not have an environmental problem

The log lets you sustain the sustainability speech; step-by-step counting lets you sustain payroll. If you bill above USD 120,000 a year, install step-by-step counting this month and stop deliberating: twenty-four-hour latency plus disaggregation by reference return between USD 1,400 and USD 1,900 annually on an investment near USD 430. If you sit below USD 60,000 with a menu under twenty dishes, the monthly log still serves you, provided you record your five highest-value references separately, because half the money lives there. Operating two or more venues leaves no middle case, since without comparable data per branch you cannot know which of the two kitchens pays for the other's inefficiencies. Start tomorrow with one scale, one sheet per station and three columns: what got thrown out, how much it weighed, at which point of the process it happened.

The four differences that actually move the needle

LATENCY governs everything else. A spoilage figure landing thirty days later describes a mistake you have already repeated four times; the same figure at twenty-four hours can still rewrite Thursday's purchase order. When SATE Institute evaluates FLW reduction programmes, the variable that best predicts the twelve-month outcome is neither budget nor site size, but how many days pass before the operator sees their own number. SKU-LEVEL GRANULARITY. The logbook yields an aggregate percentage —'we ran 8 % waste'— that is rigorously useless, because 8 % is rarely spread evenly: across the kitchens we have assessed, three to five references concentrate more than half the waste by value. Without SKU breakdown you spread effort uniformly over a concentrated problem, and that is the costliest waste of all, the waste of managerial attention. SEPARATING LOSS FROM WASTE. Here sits the tension I have watched resolve badly more than once: the operator assumes the kitchen is at fault and tightens portioning, when much of the volume disappeared upstream, in a long unrefrigerated supply chain.

The four differences that actually move the needle — in practice

If the diagnosis reads LOSS, the lever is short supply chains (SSC) plus a territorial prefeasibility study of suppliers; if it reads WASTE, the lever is menu engineering. Applying the second to the first problem burns six months of programme. VERIFIABILITY. A notebook is not evidence. For a programme officer at the IDB Group or IDB Lab, the line between funding and not funding a circular economy intervention in food service runs through whether the M&E produces a timestamped series a third party can audit. The four-point protocol produces it by design; the traditional method, by definition, never can, and that locks thousands of MSMEs already cutting their FLW out of multilateral finance.