Franchisee selection in 2026: what actually changed and what is noise

Which trend rules franchisee selection right now?

The operational filter now outranks the financial one, and the measurable signal behind it is loan performance: franchise SBA loans averaged a 9.9% default rate between 2010 and 2021 (U.S.

Small Business Administration data), while restaurant and food service sits between 12% and 15% under normal conditions according to Crestmont Capital, three points above the all-franchise average. Credit access does not explain that gap, since the candidate already qualified the day the bank released the money; what explains it is everything that happens afterward, on the floor, with payroll and waste. For a group under ten units the fix fits inside ninety days: add a full OBSERVATION day at a live unit to the application file, eight hours of real service, and score what the candidate actually does once the kitchen backs up.

Payback windows got longer, and that changes who you should approve

A franchisee who cannot survive five tight years of cash is no longer a viable candidate, and the published figures say so plainly: a fast food unit recovers its investment in 18 to 36 months according to BusinessDojo, yet a Domino's franchise takes 3 to 5 years on investments of 156,000 to 682,000 dollars, McDonald's runs 5 to 7 years on tickets of 525,000 to 2.7 million, and Chick-fil-A lands between 4 and 6 years (Restaurant Velocity, 2025). We are talking about horizons that double or triple the quick-and-cheap fast food benchmark. If your brand lives in the long range, the admission criterion is not how much capital the candidate brings, but how much capital is LEFT OVER after signing. Demand a twelve-month cushion of fixed expenses, documented and kept apart from CapEx, and turn down anyone who barely clears it.

The financial file answers «can they open?», never «can they hold on?»

I got this wrong for years, and I signed off on cash-rich candidates I would not sign off on today. Checking net worth, credit score and bank references answers exactly one question —whether the doors open— and none of those three checks measures whether the franchisee can read a food cost report before the month closes on its own. At Masterestaurant, Diego F. Parra keeps pushing the other question, the one that shows up in year three: can this operator hold margin when protein climbs 14%? The file stays silent. A seven-unit group in Guadalajara closed three franchises in eighteen months, all three with spotless paperwork and zero hours of restaurant operation, and burned roughly 640,000 dollars of sunk CapEx. The correction is cheap: add a P&L reading test to your scoring, using real numbers from a unit inside the network. Drawing the polygon first and letting the franchisee choose inside it is the trend with the most direct effect on break-even, because it pulls the territorial call away from investor enthusiasm and hands it back to the brand.

Territory gets decided before the candidate, not after

Chipotle holds a net unit growth target of 8% to 10% per year (CRE Daily, 2025) while running a model where site selection stays corporate, and Yum China reached 12,640 KFC stores by September 2025 with that same planned-expansion discipline. Set that against the traditional process, where the candidate shows up with a location already in mind and the franchisor merely approves or rejects it. For an operation of three to fifteen units the switch costs no software: map the polygon by density, traffic and competition, publish it in the commercial pack, and refuse any file that arrives with its own address outside the boundary. Scoring floor behavior beats scoring solvency today, and building it takes two weeks of internal work. Write a weighted six-criteria rubric: food cost variance against standard in a simulated exercise, reaction time to a complaint at the table, literal manual compliance on a trivial task, running a shift one person short, inventory count discipline, and reading the sales report by daypart.

Observed-performance scoring: how to measure it without buying anything

Each criterion gets a 1 to 5 grade across eight hours of real service. A candidate below 21 points out of 30 does not get in, however clean the paperwork looks. Flip it around: if the 0.9% first-year restaurant failure rate recorded in 2025 (Datassential) is the lowest since 2018, mortality no longer lives at launch but in years two and three, which is precisely where missing operational judgment hits hardest. Drop the competency interview as your central filter, even though the whole industry sells it as proof the process went professional. The methodology is not the problem, the raw material is: it measures what the candidate NARRATES about past situations, and an investor with a commercial background tells a beautiful story about a payroll crisis he never lived through. In the Guadalajara case all three interviews were flawless and all three units closed. It does help screen for honesty and return expectations, which is why I keep it at stage three of the funnel, after the floor observation and never before it.

The overrated trend: the structured competency interview

The same goes for entrepreneurial psychometric tests, which correlate with risk appetite rather than cost discipline. If your committee will spend three hours on a candidate, spend them watching him work instead of hearing him describe how he would. Adopt two things immediately and watch a third without committing budget. Adopt now: the floor observation day as a mandatory funnel stage, and the twelve-month cushion requirement kept separate from CapEx, both doable without hiring anyone and consistent with recovery windows of 3 to 7 years depending on the brand (Restaurant Velocity, 2025). Keep watching: predictive scoring models fed with your own network data, useful once you hold at least thirty units with two years of comparable history —below that the sample does not speak, it opines—. One warning about pace: growing 8% to 10% net per year the way Chipotle does (CRE Daily, 2025) demands a funnel with a high rejection rate; if your candidate approval rate runs above 35%, you are not selecting operators, you are selling contracts.

What a badly chosen franchisee really costs, in cash terms?

A misawarded unit does not cost you the CapEx, it costs the CapEx plus the hole of a territory locked up for the length of the contract.

The Mexican case burned around 640,000 dollars across three units, and on top of that sit eighteen months in which that polygon produced neither royalties nor brand density. If normal recovery for that brand runs 3 to 5 years (Restaurant Velocity, 2025) and you waste a year and a half on the wrong operator, the return clock restarts from zero with the second one, it does not resume where the first left it. What would have happened if that same group had rejected all three and waited six more months? It would have given up six months of royalties on three units and kept 640,000 dollars. That subtraction favors waiting in every scenario I have run. First difference: WHAT gets measured.

Four differences that move the cash

The traditional route measures ability to pay, which answers «can they open?»; the Masterestaurant route measures observed floor performance, which answers the only question that matters at year three: «can they hold margin when protein cost jumps 14%?». Different questions, and a financial file never answers the second one. Second, WHEN territory gets decided. In the traditional process the candidate arrives with a site in mind and the franchisor validates or rejects it, which hands the territorial decision to investor enthusiasm; in the Masterestaurant process prefeasibility comes first, the brand draws the polygon and the franchisee picks inside it, exactly the reverse of standard practice. Third, the COST of being wrong. A traditional process runs 1,800 to 3,200 dollars per candidate and produces, at industry average, one closed unit in five; the operating pilot lifts process cost to roughly 6,000 dollars, and those extra 3,000 sit against 180,000 to 260,000 dollars of sunk CapEx per failed unit.

Four differences that move the cash — in practice

The arithmetic leaves no room for debate. Fourth, and nobody enjoys hearing it: SPEED. The Masterestaurant method takes twice as long to award a franchise. A franchisor under initial-fee pressure sees a cash-flow problem there, and within the quarter that view holds; across thirty-six months the slow-selection system keeps more units alive, earns better recurring royalty and owns a brand that can still be sold.