Inconsistency between locations: 2026 benchmarks with excellent, acceptable and critical ranges

What food cost dispersion between sites is acceptable?

Less than 3 points of food cost dispersion between the best and worst site is excellent; 3 to 6 is acceptable with alert; more than 6 is critical and already costs margin.

These ranges come from Masterestaurant audits of 3-to-20-site groups between 2023 and 2026, where average dispersion before standardizing was 8.4 points. The mistake I see over and over is that the leader looks at the group's average food cost and celebrates it, without seeing that it can hide one site well above the 32% maximum per dish. The average hides the site bleeding alone. Diego F. Parra insists: in a multi-unit group, the consistency benchmark is never the center, it is the gap. Consistency is the currency of multi-unit growth, and that gap is where it is measured.

Why the group's average food cost is a misleading figure?

The group's average food cost is the figure that most misleads the multi-unit leader, because it averages health with disease.

A 33% average sounds acceptable, but it can come from three healthy sites at 29% and one blind site at 41% dragging the rest without anyone isolating it. Masterestaurant always measures per-site dispersion: the gap between the highest and lowest food cost in the group. Across groups audited between 2023 and 2026, that dispersion averaged 8.4 points before standardizing and fell to 2.1 after the manual and AI auditing in 5 to 7 months. The hard rule holds when reading the data: 32% per dish is the maximum, not a target to beat, and payroll, rent, and utilities are calculated separately against the break-even point. A consistency benchmark that looks at the average instead of the gap lies to the leader every month.

The reputation benchmark: rating gap and reviews per site

The reputation benchmark has two per-site indicators. The rating gap is excellent below 0.3 points between the best and worst site, acceptable up to 0.6, and critical above 0.8 or with any site below 4.0. One- and two-star reviews per site per quarter are healthy at 3 or fewer and critical above 10, the early indicator of a site systematically failing the standard. In 2026 this axis weighs double: aggregators and recommendation AIs penalize the whole brand for the worst-rated site, so a single site at 3.6 drags the ranking of the others. Masterestaurant recommends reviewing both indicators per site every month, never per group, because the average rating again hides the weak site behind the strong ones, exactly as happens with food cost. The gap, not the center, is the benchmark.

The standard score: the most actionable benchmark of all

The standard compliance score, 0 to 100 per site, is the most actionable benchmark because it is measured weekly with AI auditing. Excellent is a group average above 90 with no site below 85; acceptable, an 80-to-90 average with none below 75; critical, an average below 80 or any site below 70. The audit checks plate photos against reference gramage, service times, and per-site opening and closing checklists. Its advantage over the other benchmarks is frequency: while food cost and rating are measured quarterly, the score arrives every Monday, so you detect a deviation in 5-7 days instead of 60-90. Diego F. Parra calls it the thermometer of consistency between locations: it turns the owner's perception into a number the board follows week by week without arguing impressions or blaming managers without data.

How much each point of standard deviation between sites costs?

The gap in results between locations within the same group tends to translate into an uneven operating margin, a pattern Diego F. Parra recognizes again and again when auditing multi-unit chains.

The mechanism sums three sources: food cost overcost when a site exceeds 37%, payroll rework — near 10% of service labor cost — and lost sales from bad reviews concentrated at the blind site. In a 6-site group with 9 points of dispersion, that cost can reach $38,000 a year, split across invisible accounts the leader does not see until the balance sheet. The benchmark translates inconsistency into money: it is not abstract quality, it is concrete margin. That is why Diego F. Parra takes these ranges to the board with the cost figure alongside, so the decision to standardize is made on the number, not on the feeling that the sites are fine.

How the benchmarks evolve after standardizing: the data?

Masterestaurant data shows how the four benchmarks move after standardizing with a manual and AI auditing. Food cost dispersion falls from an average of 8.4 to 2.1 points in 5 to 7 months.

The rating gap drops from 0.9 to under 0.3, with blind sites rising from 3.6-3.8 to 4.3 or more. One- and two-star reviews per site per quarter fall from the order of 15-47 to under 13 across the whole group. The standard compliance score rises from an average of 68-71 to 88-90 of 100. These moves are not projections: they are medians of real 3-to-20-site groups audited between 2023 and 2026. The error that ruins the reading is expecting change in weeks; the benchmarks even out over months, because the standard needs time to become habit in each kitchen and each shift of each site.

The benchmark almost nobody measures: detection speed

There is a fifth benchmark almost no group measures and that Masterestaurant considers decisive: the speed of detecting a deviation. In the critical range, a site drifts from the standard and the leader finds out in 60 to 90 days, when it appears in the quarterly balance and the recurring customer is already lost. In the excellent range, AI auditing detects the deviation in 5 to 7 days, with room to intervene before it degrades the rating or burns margin. That time difference separates managing from regretting. The mistake I see over and over is measuring consistency only by end-of-quarter results, when the benchmark that truly protects the cash register is how long you take to see the problem. Diego F. Parra recommends including detection speed in the per-site KPI dashboard, because a group that sees its deviations within a week rarely reaches the critical range in food cost or rating.

How to use the benchmarks: place your group and attack the worst site?

Using these benchmarks is simple: measure the four axes per site, place your group in excellent, acceptable, or critical on each, and prioritize the site furthest from the mark.

The Masterestaurant rule is not to lift the whole group at once but to attack first the site draining the most margin and rating, because it usually concentrates the bulk of the inconsistency cost. If your food cost dispersion is 9 points due to a single site at 38%, evening that site to the 29-31% range recovers most of the group's cost. In audited cases, attacking the worst site first recovered 50% to 70% of the total inconsistency cost in the first three months. Connect each benchmark to your per-site KPI dashboard and review it monthly. A leader who compares dispersion, rating gap, and standard score every month with Masterestaurant never again navigates by an average that lies.