When a restaurant's food cost runs five points above target, the instinct is to blame vendors and raise prices. The truth is usually closer: food cost doesn't leak in one place — it leaks in five, from purchase order to plate, and each leak has a specific, boring, effective control. Restaurants that run all five controls hold food cost within a point of target through every price cycle. Here are the leaks and the plumbing.
Leak 1: Purchasing — paying wrong before anything arrives
- The leaks: single-vendor comfort (rates creeping quarterly), spec-free ordering ('achha wala paneer' invites price games), and panic buying at retail rates
- The controls: rate contracts on your top-20 items reviewed quarterly against 2–3 quotes; purchase specs written (grade, size, brand where relevant); order calendars matching supplier cycles (mandi-day buying for produce)
- The metric: item-wise purchase price trends in your MIS — a ₹4/kg tomato drift across 300 kg/month is ₹14,400/year from one SKU
Leak 2: Receiving — where paid-for food doesn't arrive
The most-ignored leak: short weights (the 9.4 kg 'ten-kilo' crate), quality substitution (yesterday's fish at today's spec), and invoice games (rates differing from contracts). The controls cost almost nothing: a calibrated scale at the receiving door used on every delivery, spot quality checks against specs with a rejection habit (one documented rejection reforms a vendor for months), invoice-versus-rate-contract matching before goods-received entry, and receiving done by someone other than the person who orders (the classic collusion split). Ten minutes per delivery; 1–2 points of food cost recovered in most audits.
Leak 3: Storage — buying right, then rotting it
- The leaks: spoilage from temperature abuse, FIFO failures (new stock in front), unmeasured shrinkage, and the walkout leak (stock leaving in bags)
- The controls: temperature logs on every unit (twice daily, alarmed where possible), physical FIFO discipline with dated labels, weekly A-item counts against theoretical stock, store-room access limited with issue registers (kitchen draws against indents, not open doors), and CCTV positioned on stores/receiving — visible deterrence
- The metric: weekly variance (theoretical vs actual stock) per category — variance above 2% gets same-week investigation
Everything measurable runs through one weekly ritual: POS-theoretical consumption (recipes × items sold) versus actual depletion (opening + purchases − closing counts). The gap IS your leaks 3–5, quantified. Restaurants without recipe-level POS inventory are flying blind here — it's the strongest argument for the system.
Leak 4: Production — the kitchen's silent generosity
- The leaks: portion drift (the ladle that grew), prep wastage beyond standard yields (peeling losses, trim discarded instead of stocked), batch overproduction dying at day-end, and staff meals scaling informally
- The controls: portion tools at every station (ladles, scoops, weighing at the pass for proteins) with photo standards; yield audits quarterly (re-baseline your onion, chicken, paneer yields — new hands change them); production planning from sales forecasts (yesterday-same-day-last-week patterns) with batch sizes matched to day-parts; trim utilisation SOPs (bones to stock, trims to staff curry); staff-meal menus defined and costed (feed people properly and on-budget — informal grazing costs more than a good staff meal)
Leak 5: Service and billing — cooked, served, never charged
The final metre: unbilled items (verbal orders that skip the KOT), void/discount abuse (the cashier's brother eats free), delivery-platform errors (wrong-item remakes, unclaimed refunds), and complimentary culture without policy. Controls: no-KOT-no-food as an absolute rule (kitchen serves only against tickets), void/discount permissions restricted to managers with daily reason-logged reports, aggregator reconciliations weekly (their deductions and refunds audited — money leaks in those dashboards routinely), and a written comp policy (who may, for what, logged). The daily manager checklist ties all five leaks together: rate checks, receiving verification, temperature logs, variance flags, void report — fifteen minutes that guard 3–5 points of margin.
How Aidwish helps
Aidwish installs food-cost control systems — the five-leak audit on your current operation, recipe/POS configuration for variance reporting, receiving and store SOPs, and the manager rituals — typically recovering 2–5 points of food cost inside a quarter.