Franchise Unit Benchmarking: Why the League Table Misses Your Weakest Unit

A franchise unit ranked sixth of twelve in your network can be the one that most needs help, while the unit ranked last may be doing fine. Franchise unit benchmarking is the practice of scoring each location against peers of the same age band and format instead of against the network average. Score units that way and the weakest one often sits mid-table, above the average, hidden by a league table that puts year-one units and ten-year veterans on the same list.

What is franchise unit benchmarking, and why does a league table fail?

Franchise unit benchmarking compares a location's result with a group of similar locations, usually by dividing the unit's figure by the group's median. A league table fails because it ranks units of different ages and formats on one list, so the order mostly reflects how long each unit has been open and not how well anyone runs it.

Age is the first difference worth removing. The U.S. Bureau of Labor Statistics follows cohorts of new private-sector establishments across all industries (these are not franchise outlets specifically). Of the establishments that opened in the year ending March 2021, 79.1 percent were still operating a year later and 47.2 percent five years later, according to the agency's Business Employment Dynamics survival table. Those figures measure survival, not sales, but they show that a unit in its second year and a unit in its tenth sit at different points on a path that just over half of new establishments leave within five years.

A worked example: twelve units, three age bands

The numbers below are invented for illustration. They do not come from any real brand or from Revscale client data. Twelve units report annual sales in thousands of dollars. Five are mature (60 or more months open): 1,420, 1,310, 1,260, 1,180 and 1,020. Four are in ramp (24 to 59 months): 1,050, 990, 940 and 880. Three are new (under 24 months): 520, 480 and 470.

The twelve sales figures add up to 11,520, so the network average is 11,520 divided by 12, or 960. A league table puts the 1,020 unit (call it Unit E) sixth of twelve and above that average. The 470 unit (Unit K) ranks last, and the 520 unit (Unit L) ranks tenth.

Now divide each unit by the median of its own band. The mature median is 1,260, the ramp median is 965 and the new median is 480. Unit E scores 1,020 divided by 1,260, which is 0.81, or 19 percent below its peers, the lowest score in the network (the next lowest is Unit I at 880 divided by 965, or 0.91). Unit K scores 470 divided by 480, or 0.98, within about two percent of its peers. Unit L scores 520 divided by 480, or 1.08, first in its band.

The league table sends field support to Unit K. The peer view sends it to Unit E, which sells 240 thousand dollars less per year than the typical mature unit (1,260 less 1,020) and which nobody would flag because it clears the network average.

How many peers does a peer group need?

In my working rule, a peer group needs at least five units before its median means much. That is a judgment call, not a statistical standard. With three units the median is just the middle unit, so one unit scores exactly 1.00 by definition and the other two are measured against a single neighbor.

The example's new band has three units, so Unit J scores 1.00 automatically. Merging bands to reach five units brings the original problem back. Fold the new band into the ramp band and you get seven units with a median of 880, which makes Unit L (520 divided by 880, or 0.59) look like a failing unit when it is the best new unit in the network. A better option is to keep the age bands, score the small band, and label it provisional until the band fills. A second option is to compare a new unit with the brand's own sales history for units at the same month of operation.

Sample size matters for audits too, and the arithmetic for how many locations an audit needs to cover shows how quickly small networks run out of statistical room.

What does the FTC Franchise Rule say about comparing subsets of outlets?

Under the FTC's Franchise Rule, an Item 19 financial performance representation must state whether it covers all outlets or a subset, how many outlets supplied data, and the number and percent of them that reached or beat the stated result. This is a U.S. federal rule, and the text was current on the eCFR as of October 22, 2026.

Your internal benchmark runs into the same question. In the example, the 960 average was met or beaten by seven of the twelve units, which is 7 divided by 12, or 58.3 percent. A franchisor that stated that average would also have to state that seven of twelve outlets, or 58.3 percent, reached it, along with the characteristics that define the group, under 16 CFR 436.5(s). The rule also requires a reasonable basis and written substantiation for any such representation. If you cannot say how many of your units reached a number, you are not yet benchmarking, and Item 19 disclosures that stop at revenue leave the same gap for candidates.

A four-cut scorecard for fair peer groups

Make four cuts before you score anything. The first cut is age band, measured in months open, with the band edges set by where your own ramp curve flattens. The second is format, so end-cap, inline and freestanding units, or kitchens and storefronts, are never scored against each other. The third is market tier, which groups units by local rent and wage level so a high-cost metro is not scored against a low-cost town. The fourth is group size, using the minimum above.

Then score each unit as its figure divided by its peer median, and flag anything at or below a threshold you choose from your own history. A line at 0.85 would have flagged only Unit E in the example. Pair that score with a four-number unit economics scorecard so a unit that sells enough but keeps too little still shows up.

Which unit gets the next field visit?

Visit the unit with the lowest peer score before the unit with the lowest rank. In the example that is Unit E, where the gap to its peers is 240 thousand dollars a year and the network average would never have revealed it. Before the visit, check which cut moved: a unit that fell from 0.95 to 0.81 in two quarters needs a different conversation than one that has sat at 0.81 for three years.

The division is simple enough for a spreadsheet. Revscale's network intelligence platform runs the same peer-group scoring from unit-level data across a whole system, which matters once the cuts multiply past what anyone updates by hand each month.

Franchise unit benchmarking pays off when it changes who gets help first. Build the age bands this month, score every unit against its own peers, and compare the new list of weakest units with the list your league table produced. Where the two lists differ is where your support hours are going to the wrong place. This is general information, not legal, tax or financial advice; confirm specifics with a qualified advisor.