GGB The Operating Index · Q4 2026 · band set 250601e4
The bands disciplined restaurants run against. UAE restaurant cost benchmarks · modelled bands
Food, labour, rent, delivery and prime cost as shares of revenue — the published GGB operating bands, with the structural red-lines that mark where tweaking stops working. One registry, every GGB instrument, honestly labelled.
The operating bands
The one number we hold the line on
Prime cost — food + labour — belongs between 55% and 62%.
Hold prime cost inside the band and the rest of the P&L has room to breathe. Lose it, and revenue rarely rescues the month — which is why every GGB diagnosis starts here, whatever door the operator came through.
Food cost
band 28–32% · red-line >38%
The first line read in every GGB diagnosis. Past the ceiling, margin leaks faster than volume can refill it; past the red-line, tweaks stop working and the structure has to be reset.
Labour cost
band 25–30% · red-line >35%
Read against covers by daypart, not headcount. Past the ceiling the wage line is usually running the operator; the red-line marks where the roster shape itself is the problem.
Rent
band 6–12%
Fixed at signing, so it amplifies every other line. The one major cost that cannot be operated down — which is why it is filtered before the lease, not negotiated after.
Delivery commission
band 3–6%
Delivery commission as a share of TOTAL revenue. (The separate per-order aggregator rate carries its own structural red-line above 25% — read in the Delivery Margin Recovery tool.)
Scale shown 0–40% of revenue. Red-lines are the owner-signed structural thresholds GGB's grading runs on — deliberately ABOVE the published maxima: a line is urgent when structurally past it, not merely at it. The per-order aggregator commission carries its own >25% red-line, read in the Delivery Margin Recovery tool.
Methodology, honestly
These are modelled ranges from GGB's operating benchmarks — the single published registry every GGB diagnostic, article and report runs against, maintained from 28+ years of founder-led operating record across the GCC, India and Singapore. They are not a market survey: food cost, rent and labour vary hugely by locality, lease and concept, so a flat "the market averages X%" claim would be unsupportable — and we don't make it. Where an operator should sit inside a band depends on their concept, not their city. Band set 250601e4; a new edition exists only when the registry itself changes.
UAE restaurant cost benchmarks, row by row
What are healthy restaurant costs in the UAE? In GGB’s published bands, food cost sits at 28–32% of food revenue, labour at 25–30% and prime cost at 55–62% of revenue, with rent at 6–12%. Each row below states how the figure is measured and where it comes from.
| Benchmark | Band (share of revenue) | Red-line | How it is measured | Basis |
|---|---|---|---|---|
| Food cost | 28–32% | above 38% | Cost of food consumed (opening stock + purchases − closing stock) ÷ food revenue, for the same month. | GGB benchmark registry; the published maximum every GGB diagnostic reads. |
| Labour cost | 25–30% | above 35% | Total labour cost ÷ revenue, where labour includes salaries, benefits, visas, accommodation and end-of-service accruals. | GGB benchmark registry. |
| Prime cost | 55–62% | — | (Cost of food and beverage consumed + total labour cost) ÷ revenue. | GGB benchmark registry; food plus labour, held inside the band. |
| Rent | 6–12% | — | Annual rent under the lease ÷ annual revenue. | GGB benchmark registry; read before the lease is signed. |
| Delivery commission | 3–6% of total revenue | per-order rate above 25% | Commission paid to delivery platforms ÷ total revenue across all channels. The per-order platform rate is a different measure (GGB’s glossary gives 15–30% as the typical GCC range). | GGB benchmark registry. |
| Rent share by area | Not published | — | Rent for a specific address comes from its lease, read against the rent band above. | No verified area-by-area data on this site yet; a figure appears here only from a sourced, dated dataset. |
| Fit-out cost per m² | Not published | — | Area × your contractor’s rate per m² for your shell condition, as the opening-cost guide sets out. | No verified market figure exists to publish; the CAPEX Range Builder refuses to invent one too. |
Cite this page: GGB Consulting, “UAE restaurant cost benchmarks: the GGB Restaurant Operating Index”, band set 250601e4, https://ggb.consulting/restaurant-operating-index. Free to quote with a link; the band set changes only when the registry does.
A line past its red-line is structural, not a tweak — that is the point where a restaurant turnaround consultant in Dubai resets the menu, purchasing and roster together. A line inside the band still leaks if the rest of the P&L does not hold; the free Profit Leak Audit reads all four on your device.
By P. Dayaparan, founder of GGB Consulting: 28+ years in hospitality leadership, PMP. Updated .
Read your own lines against these bands — free, on your device, in minutes.
The measured pool
Where opted-in operators actually sit.
Every tool result on this site carries an unticked box: include my anonymised figures in the GGB Index. Operators who tick it contribute two percentages — food cost and labour as shares of revenue — with a market and concept label. No name, no venue, no contact, no revenue figure. Each quarter the pool is aggregated into an edition; a group publishes only when it holds at least 8 operators, and below that it is reported as exactly what it is: collecting.
Edition 2026-Q3 (built 2026-09-04): 0 opted-in runs — the pool publishes its first group at 8 opted-in operators. Until then this section says exactly that — a threshold, not a number.
Contribute to the pool
Two percentages and five labels. No name, no venue, no contact — the same seven fields the tools offer, here on the page that publishes them. The full form and its terms: the GGB Index.
Editions (1)
- 2026-Q3 · built 2026-09-04 · 0 opted-in runs · 0 published groups
The Index, answered straight
- What is a healthy food cost percentage for a restaurant?
- In GGB's published operating bands, food cost belongs at or under 32% of revenue (typical range 28–32%). Past 38% it is structurally past tweaking in our method — the menu, purchasing and yield need a reset, not a nudge. Where inside the band a specific operation should sit depends on concept and market, which is why these are diagnostic bands, not targets to chase blindly.
- What should prime cost be in a restaurant?
- Prime cost — food plus labour — belongs between 55% and 62% of revenue in GGB's bands. It is the one number we hold the line on: keep prime cost inside the band and the rest of the P&L has room to breathe; lose it and revenue rarely rescues the month.
- Where do these numbers come from?
- From the GGB benchmark registry — the single published band-set every GGB diagnostic tool, article and report runs against, maintained from the founder's 28+ years of operating record across the GCC, India and Singapore. They are modelled diagnostic ranges, honestly labelled — not a survey of other consultancies or a claim about any specific market's average.
- How often does the Index change?
- Only when the underlying registry changes — each band-set carries its identity code, and a new edition exists only when a band genuinely moves. The bands are deliberately stable: they mark operating discipline, not market fashion.
- How do I compare my restaurant against the Index?
- Run the free Profit Leak Audit with last month's figures — it reads your food, labour, rent and delivery lines against these exact bands and ranks where margin is leaking, on your device, in minutes.