Cross-Location Consistency: The Data That Dismantles the Second-Location Myth

What does consistency across locations actually measure?

Consistency across locations is the DISPERSION of three or four metrics around the network average, expressed in percentage points between your best and worst site.

A six-unit group averaging 29.8% food cost with one store at 34.1% carries a 4.3-point gap, which on monthly revenue of USD 120,000 per site means roughly USD 5,160 vanishing every month without anyone signing a check. That number exists whether you measure it or not. The International Franchise Association projected more than 850,000 franchised units by the close of 2025, and most of those networks govern quality with written manuals rather than data series. A manual describes what should happen; dispersion describes what happened last night during closing shift. Start with the gap, not with the document. Instrumenting one site costs less than a single month of food cost drift, and that is the argument that wins board meetings.

Measurement versus declaration: instruments cost less than drift

A calibrated digital scale runs USD 150 to 400, a forty-point audit template is a spreadsheet, and management time comes to four hours per location each month. Against that, the 4.3-point gap in the Bogotá case drains around USD 62,000 a year from one store alone. Put the two figures side by side and the debate ends. The group that only wrote the manual believes it is consistent because it documented how things ought to run; the group that measures knows its worst location sits 2.1 points high and acts on Monday morning. What separates them is not culture or discipline: it is instruments. Twelve monthly forty-point audits catch a deviation within thirty days; the annual two-hundred-point audit catches it after eleven months of quiet margin loss. The arithmetic is brutal and allows no nuance: a 2-point food cost deviation on USD 120,000 monthly costs USD 2,400 a month, so the annual model pays USD 26,400 for the privilege of a thick report.

Frequency versus depth: twelve short audits beat one long one

Deloitte has documented across multi-unit operations studies that chains running monthly operational reviews hold margin better than those on an annual cadence, and the mechanism is plain: they do not audit better, they audit sooner. I got this wrong for years recommending exhaustive checklists. Fewer variables, more often, with the same person signing off. That is the formula that holds quality while you replicate. Five variables explain most of the dispersion between sites: food cost per plate, closing waste in kilos, ticket time in minutes, kitchen staff turnover and average public rating. Everything else is noise or a consequence of those five. Chick-fil-A added 179 net locations in 2025 to reach 2,863 units, according to QSR Magazine, and no network grows at that pace while watching two hundred indicators per store: it grows because the watched set is short and gets measured daily. Wingstop opened 278 net restaurants between 2024 and 2025 on the same narrow-dashboard logic.

The five variables that explain almost all the variance

Pick your five, set tolerance in points —I work with ±1.5 points on food cost and ±90 seconds on ticket time— and publish the by-site table where the whole team can read it. These ranges read differently by size, and applying them without translation is the fastest way to lose a year. With two or three locations, track only food cost and waste, weekly, on a shared sheet: your target gap between best and worst site is 2 points, and the owner can still walk all three kitchens on the same Saturday. At four to eight sites you need a dedicated operations supervisor and the full five variables with a signed monthly cut, tolerance ±1.5 points. Past nine locations the problem stops being measurement and becomes data governance: point-of-sale integrated with inventory, automatic alerts when a site breaches range two weeks running, and a thirty-minute meeting that discusses only the worst store.

How to read these numbers in YOUR operation?

Never the average. The figures above come from public industry sources —International Franchise Association, QSR Magazine, Deloitte, the Small Business Administration, CANIRAC in Mexico, Tormo Franquicias in Spain— crossed with the reading we do at Masterestaurant on multi-unit operations.

Their limits deserve saying out loud. Almost every expansion aggregate describes limited-service chains in the United States, whose average check and payroll structure look nothing like a casual dining group in Bogotá or Mexico City. Regional revenue data —EUR 7.23 billion in Spanish franchised restaurants during 2024, BRL 495 billion in Brazilian food service for 2025 per ABRASEL— helps you size a market, not set your internal tolerance. A benchmark is worth the distance between it and what you measured yesterday. Without your own measurement, it is presentation decoration. Assume you open the seventh site with today's dispersion uncorrected, then follow the consequence all the way out, because that is where the trap sits.

The scenario almost nobody runs before signing site seven

If your gap today runs 4.3 points between best and worst store, site seven inherits the average of a miscalibrated system rather than the standard of your strongest kitchen, and by month four it will breach range exactly like site six. Multiply it: two locations out of tolerance on USD 120,000 monthly each come to roughly USD 124,000 a year, plus the management cost of firefighting instead of opening the eighth. The Bureau of Labor Statistics puts first-year restaurant failure near 14%, and in multi-unit groups failure rarely arrives as a loud bankruptcy: it is margin evaporating store by store. Close the gap before you sign the lease. The thicker the manual, the less consistent the network tends to be, and the explanation is operational rather than ideological: nobody consults two hundred pages during a Friday shift with a full kitchen. Diego F. Parra frames it in Masterestaurant audits as a choice between describing and verifying, because a head chef's attention budget is finite.

The thick-manual paradox and how it resolves

The way out of the paradox is a short manual covering critical process —portioning, temperatures, closing sequence— plus a verification template that produces a NUMBER every week. The manual says how it is done; the template says whether it was done. In the Bogotá case, two deviations were enough to lose 4.3 points: a scale off by 40 grams and a shift portioning fifteen minutes ahead of process. No manual would have caught them. A scale verified every Monday would. MEASUREMENT VERSUS DECLARATION. A group with a manual and no measurement believes it is consistent because it wrote down how it should be; the group that measures knows its worst site runs 2.1 points higher on food cost and acts on Monday. What separates them is instrumentation, not culture, and it costs less than most owners assume: one calibrated scale, one audit template and four hours of management time per site each month.

The four differences money decides, not opinion

FREQUENCY VERSUS DEPTH. The annual 200-point audit from a large consultancy catches deviations that have been draining margin for eleven months; twelve 40-point audits catch them within thirty days. Deloitte reports that chains running monthly operational reviews hold 2.3 points more operating margin than annual reviewers, and that figure survives any methodological argument because the mechanism is obvious: an error corrected in week two never reaches week twenty. DECLARED TOLERANCE VERSUS IMAGINED PERFECTION. A process without a declared tolerance band is not demanding, it is unauditable: if the spec says 'a generous portion of protein', every cook defines generous and you now run five recipes under one brand. Declaring ±3 grams lets you state precisely who is inside the band and who is outside, and that precision turns an awkward conversation into a five-minute technical correction. PREFEASIBILITY VERSUS ENTHUSIASM. Half the inconsistencies blamed on operations are born before opening, in a location decision made on instinct.

The four differences money decides, not opinion — in practice

Applied location intelligence — pedestrian traffic, catchment spend, cannibalization with a sister site 1.2 kilometers away — explains much of the average-check gap the manager is later asked to compensate for. I got this wrong for years: I blamed teams for a check decline the map had already announced.