Most F&B operators track revenue daily and everything else monthly. By the time the monthly numbers land, four to six weeks have passed since the period being measured. Margin issues that were visible in week two of the month aren't surfaced until week six — which is week two of the next month, which has its own issues already in motion.
The four numbers below are the ones that decide whether a Singapore F&B business runs at viable margin. Each one has a natural cadence — some daily, some weekly, some monthly. Track at the right cadence and you catch drift while it's small. Track only monthly and you catch it at the same time as your accountant — too late.
1. COGS — daily-feed, monthly-true
Cost of goods sold is the food and beverage cost of what you actually sold. Two layers:
The estimate (daily)
Each menu item should have a theoretical food cost based on its recipe. POS multiplies sold-quantity by theoretical-cost-per-item to estimate daily COGS. This isn't accounting-accurate — it ignores waste, portion drift, theft — but it's directionally right and available immediately.
The actual (monthly)
Month-end stock-take reconciles theoretical to actual. The variance — actual COGS minus theoretical COGS as a percentage of revenue — is the early warning system for operational drift. Variance under 1% of revenue is tight; 1-2% is normal; above 3% means something is structurally off.
For a Singapore café, food cost typically runs in the mid-20s to mid-30s as a percentage of food revenue, depending on concept and pricing. Beverage cost (non-alcoholic) typically 15-25%. Alcoholic 20-30%. Your own concept's historical norm matters more than industry benchmarks.
2. Labour — weekly
Labour is the most actionable variable cost on a weekly cycle. Track three views:
- Total labour cost as % of revenue, weekly, per outlet.The headline ratio. When it drifts more than 2 percentage points above your outlet's target for two weeks running, the rostering needs review.
- Labour cost by role per outlet. Kitchen vs FOH vs management. Different concepts have different targets; track each one independently.
- Cost per labour hour, blended. Total labour cost divided by hours worked. Tracks wage inflation over time.
For Singapore F&B, labour typically runs in the high 20s to mid 30s as a percentage of revenue. Quick-service models trend lower; full-service and premium casual trend higher. Track your own concept's trend rather than an industry benchmark.
Source: payroll system with tracking categories, integrated into Xero. Weekly views can be estimated from rostering data before payroll runs.
3. Rent and fixed costs — monthly, as % of revenue
Rent doesn't change between months. What changes is rent as a percentage of revenue— which tells you whether your top-line is actually supporting the location's fixed cost burden.
For Singapore F&B, rent typically runs 8-15% of revenue. Above 15% is usually a structural margin compressor that needs offsetting somewhere else (higher gross margin, lower labour ratio, premium pricing). Above 20% rarely works long-term unless the location is genuinely premium-positioned.
Track monthly, by outlet. Add other outlet-fixed costs (utilities, insurance, cleaning) into the same view to see total fixed-cost burden as a percentage of revenue.
The conversation this drives: which outlets are carrying which. The consolidated P&L hides this; outlet-level visibility makes it obvious.
4. Cash flow — daily-live, 14-day forward
Bank balance is a lagging indicator. The decisional view is cash position calibrated against the next 14 days of inflows and outflows.
Components:
- Current cash across operating accounts
- Expected card / platform / receivables inflows next 14 days
- Known outflows next 14 days (payroll, CPF, rent, supplier batches)
- Net position at end of 14-day window
- Days of runway at trailing 30-day burn rate
See our deeper piece on seeing cash flow clearly without waiting for month-end for the structure in detail.
The summary view: prime cost
Combine the first two — COGS plus labour — and you have the most-watched F&B ratio: prime cost as a percentage of revenue.
For viable Singapore F&B operations, prime cost typically targets high-50s to low-60s as a percentage of revenue. Above 65% sustained, the operation is structurally tight. Above 70%, something needs to change soon. Track monthly per outlet; trend quarter-over-quarter.
Cadence summary
- Daily: cash, revenue, theoretical COGS
- Weekly: labour ratio, exception feed
- Monthly: actual COGS, prime cost, rent %, outlet P&L
- Quarterly: trend review, structural drift analysis
How AI agents help with these metrics
The data ingestion is the work, not the math. Our named agents handle the routine layer:
- The Receipt Catcher keeps food and beverage purchase data flowing in promptly, so COGS actuals are clean at month-end
- The Outlier flags abnormal cost patterns — a spike in produce cost, a labour ratio spike for the week, an unusually large supplier payment
- The Compassdrafts the weekly summary that surfaces all four metrics for the owner's 90-second read
Common questions about F&B metric tracking
How do I track COGS daily if I only stock-take monthly?
Daily COGS is theoretical (recipe-based, from POS sales). Monthly COGS is actual (stock-take based). The difference is the variance, which becomes the operational signal. Daily theoretical isn't accounting-grade — but it's directionally right and lets you spot drift before the month-end stock-take confirms it.
What if my labour cost is consistently 40% of revenue?
For most concepts, sustained labour-cost ratios above the high 30s indicate either over-staffing, under-pricing, or a structural service model mismatch. Some premium concepts run higher deliberately. The diagnostic is whether the labour ratio is delivering — review productivity per hour and revenue per labour dollar, not just the ratio in isolation.
How do I separate food cost from beverage cost when they share suppliers?
At supplier invoice coding time, split the invoice across the two accounts. For multi-category suppliers (a beverage distributor that also sells snacks), the AI Categoriser can draft the split based on historical patterns; a human approves. Get the chart of accounts right (separate food and beverage COGS accounts) and the split discipline becomes routine.
Should I be tracking these metrics weekly or monthly for a single outlet?
Cash and revenue daily, labour ratio weekly, COGS and prime cost monthly. Single-outlet operators can simplify the by-outlet cuts but the cadence per metric is the same. The most consequential change at scale is adding the by-outlet dimension; the metrics themselves don't change.
How accurate do the daily theoretical COGS numbers need to be?
Directionally right within a few percentage points. The point is to see drift early, not to replace the month-end actual. If theoretical and actual reconcile within 2% of revenue at month-end, the daily figures are accurate enough to act on.
Four numbers, four cadences. Operators who run them well make their decisions on current information. Operators who don't make them on whatever the monthly report shows, four weeks late. The first group compounds; the second works harder than they should to stand still.
— The Acuitas team