Subway operated roughly 980 German units at its mid-2010s peak. Today the number is approximately 660. Eight years, a third of the network gone, no exogenous shock to absorb the blame. The Roark Capital acquisition that closed in 2024 has placed a private-equity overlay on a franchise architecture that was already shedding units before any sponsor entered the picture.
That sequencing matters. The German Subway contraction is not the story of a PE buyer breaking a healthy chain. It is the story of a franchise model that had been hollowing out for the better part of a decade — and which now sits inside a sponsor whose return mathematics presupposes that the model still works. The German case is the legible precedent for what PE will press on harder, in every market that still carries unit economics like the German ones.
What we see
A third of a national network removed in eight years from a chain that never faced a category-specific demand shock. German labour costs compounded with food-cost inflation against a per-unit revenue base too low to absorb either. Royalty and marketing-fund mechanics — fixed off the top of revenue, indifferent to franchisee profitability — stayed where they were. The franchisee P&L closed first; the network total followed on a multi-year lag.
What it tells us
The high-density, low-investment, single-operator franchise model is a fair-weather architecture. It works when input costs are stable and per-unit revenue can absorb a 12–13% top-of-revenue royalty stack. It fails — visibly, mechanically — when either side of that equation moves. Germany moved both sides between 2015 and 2024. The Roark acquisition does not change the equation. It changes who collects on it.
Why it matters now
Roark's institutional thesis on Subway presupposes that the cash flows the system has historically generated can be defended and tightened. The German experience says the system was already cannibalising its own franchisee base to produce those cash flows long before any sponsor arrived. Every market in the global portfolio that operates on similar unit economics is on the trajectory the German market completed between 2016 and 2024. PE accelerates the visible part. It cannot reverse the underlying mathematics.
The franchise architecture — and what breaks it
Subway's growth model is the cleanest expression of a specific design choice: maximise the count of franchises sold, minimise the capital per unit, and extract the system's return through entry fees plus a top-of-revenue royalty stack rather than through unit-level profitability. A German entry fee around USD 15,000 in the build phase, build-out costs for a 60–90 square-metre location well below what McDonald's or Burger King required, and a model that opened the franchisee universe to operators without restaurant capital — together those choices scaled the German network from one Kurfürstendamm location in 1999 to nearly a thousand units within fifteen years.
The royalty stack is where the architecture's exposure lives. A German Subway franchisee paid roughly 8% of revenue in royalty and 4.5% to the marketing fund — 12.5% off the top, before rent, payroll, or goods. At an average per-unit revenue base of approximately EUR 520,000 (the implied 2024 figure from a roughly EUR 350 million national revenue total across about 670 sites), that 12.5% equals around EUR 65,000 to the system per store, per year, before the franchisee earns a single euro. For comparison, a Peter Pane unit operates at roughly EUR 2.5 million in revenue. The same 12.5% extracted from that base does not threaten unit viability. Extracted from a EUR 520,000 base, it does.
The architecture works for the system irrespective of whether it works for the operator. The system collects the entry fee at the moment of franchise sale — structurally rewarded for opening units, not for keeping them solvent. Royalty income scales with gross revenue, not with margin. Once the German minimum wage compounded with post-2022 food-cost inflation, input-cost exposure exceeded what the per-unit revenue base could absorb — and the franchisee layer closed the books on units that the system still wanted to keep open.
A second design choice compounded the problem. Many German Subway contracts contained no territorial exclusivity. The system retained the right to authorise a second Subway inside the same catchment — generating a second entry fee and a second royalty stream for the system, cannibalising guests and reinvestment incentive for the original franchisee. Stores that opened in 2005 still looked like 2005 in 2022. Not because franchisees lacked the will — because the incentive structure punished reinvestment. The equilibrium was predictable from the contract terms.
The German numbers
The peak is the harder number to pin down precisely. The chain's own communications and German trade reporting cluster the mid-2010s figure around 950–980 units, with reasonable sourcing for both ends of the range. What is not ambiguous is the direction. Subway Deutschland had already entered contraction by 2010 — the first year in which closures exceeded openings, with the network down to roughly 700 units by year-end. The chain stabilised through the early 2010s, then a sustained downward leg compounded through the second half of the decade. By 2021 the network had reached its documented trough at approximately 620 units. The post-Roark recovery has been measurable but conservative: roughly 657 units in January 2024, 671 by mid-2024, and approximately 660–666 through the first half of 2026.
The contraction was not evenly distributed. Exposure to high-traffic urban locations — Bahnhof and city-centre sites, university districts, shopping-centre tenancies — meant that the units most exposed to commercial rent inflation came off first. Suburban and Autobahn-adjacent locations held up better. That distribution explains why the post-2022 modernisation programme has emphasised highway sites: the system is retrenching into the geography where the math still closes.
The master-franchisee story is less linear. Subway's DACH coordination runs through the Cologne-based Subway Vermietungs- und Service GmbH, with the Swiss operation now under a separate master-franchise agreement with Convenience House Schweiz AG signed in 2024 — targeting an additional 70 stores over seven years, effectively doubling the Swiss network from its current base of around 54 units. The German announcement for 2025 sits at a different scale: the Swiss target presupposes that the model still scales; the German programme presupposes only that the existing footprint can be defended and partially upgraded. Two different reads of the same architecture in adjacent markets with similar cost stacks.
The Roark overlay
Roark Capital closed the Subway acquisition for approximately USD 9.6 billion. The transaction placed Subway alongside Arby's, Buffalo Wild Wings, Sonic, and Jimmy John's inside a portfolio whose sponsor has operated through multiple restaurant cycles. The institutional read at announcement was that Roark would do for Subway what Inspire Brands has done for the rest of the portfolio: tighten the royalty structure, consolidate vendor relationships, recalibrate the marketing fund, modernise the technology stack.
Each move has a predictable second-order effect on franchisee economics, and each is now visible inside the documented experience. Royalty enforcement on gross revenue appears more rigorous than under the previous founder ownership. Corporate-mandated discounts and coupon programmes — designed to drive comparable-sales numbers that support the sponsor's return narrative — compress franchisee margins because the discount cost is absorbed at the unit. Vendor consolidation — the "Sysco-fication" pattern documented across other Roark and Inspire-adjacent portfolios — concentrates supply economics in contracts the franchisee cannot opt out of.
The visible-portion signals belong in the same file. The metal covers installed over meat preparation counters in 2023–24 — framed as temperature management — were read by guests and franchisees alike as portion management. The chain's historical "Eat Fresh" positioning was always anchored on the moment of customisation: protein, vegetables, sauces, assembled in front of the guest. Anything that obscures that moment erodes the brand promise. The system gains a procurement margin. The brand pays the cost.
None of this is unique to Roark. It is the standard sequence a PE sponsor runs on a franchise portfolio with stable royalty mechanics and a franchisee base large enough to absorb the compression. The German eight-year contraction means there is materially less base left to compress — and the compression lands harder, per unit, than it did in the previous decade.
Comparator briefs
McDonald's franchise discipline. McDonald's operates roughly 1,370 German units at per-unit revenues well above EUR 4 million and reinvests systematically. The structural difference is not the royalty stack — McDonald's royalty plus rent frequently exceeds Subway's 12.5% — it is the revenue base from which that stack is extracted. McDonald's also owns or controls the underlying real estate at most German locations, which stabilises franchisee cost structure and gives the system a separate income stream that does not depend on franchisee margin compression. Subway's choice to remain pure royalty without a property base is the variable that matters.
Burger King master-franchisee collapse. The Yi-Ko Holding insolvency in 2014 — 89 German Burger King units, simultaneous closures, criminal investigation around hygiene and labour — sits in the comparator file for a different reason. Yi-Ko was a single master-franchisee operating a brand it did not control, under cost pressures it could not pass through. The structural lesson is identical: when the layer holding the operating risk is separated from the layer collecting the royalty, network quality is what gives way. Burger King Germany has since reconstituted under McWin Capital ownership. The reconstitution took roughly a decade.
Domino's DACH consolidation. Domino's acquired the Joey's Pizza network in 2016 and converted approximately 200 German units to the Domino's brand by 2018. The integration succeeded because the underlying unit economics worked: per-unit revenues in the EUR 1.0–1.4 million range, a delivery-first cost structure, an app-first ordering model that captured digital margin the franchisee could not be priced out of. Domino's is the closest analogue to a healthy high-density franchise model in Germany. The difference from Subway is the revenue model, not the franchise structure.
Five Guys's contained footprint. Five Guys entered Germany in 2017 with deliberately slow expansion, premium positioning, and unit economics targeting per-store revenue well above EUR 2.5 million. The German footprint remains under 30 units a decade in. That is not a failure — it is the architecture working as designed. The Five Guys profile is the precise inverse of the Subway one. Both can be built. The Subway one cannot survive German cost inflation. The Five Guys one is built to absorb it.
Forward read
The Roark acquisition's first full operating cycle now runs through 2026 and into 2027. The German market provides the clearest read on what to expect at the portfolio level.
The franchisee base will continue to compress. The 2025 plan — 25 new openings against 40 modernisations — is a defence signature, not a growth one. New openings concentrate in highway-adjacent and selected suburban geographies where the unit economics still close. Modernisations target high-traffic units the system needs to retain for visibility reasons but where reinvestment has been deferred for a decade. Both moves stabilise the network at roughly the 2024 footprint.
The royalty mechanics will tighten further. Enforcement of gross-revenue royalty calculation, corporate-mandated promotional programmes, vendor consolidation — the pattern documented across other Roark and Inspire-adjacent portfolios is in early implementation at Subway. Marginal-revenue franchisees will exit through non-renewal or closure. The pace is slower than 2016–2021 because the marginal-cost cohort has already left, but the compression mechanics keep applying to what remains.
The international read-across is what matters for capital. The global Subway portfolio operated at approximately 37,000 units in 2023, down from a 2015–16 peak above 44,000. The contraction has been markedly more aggressive in cost-stack-heavy markets — Germany, the UK, parts of Western Europe — than in markets with lower labour and rent bases. Roark's return hurdle on a USD 9.6 billion acquisition presupposes that the global portfolio can be stabilised at scale and the exit multiple expanded. The German evidence says the architecture has structural shrinkage built in wherever the cost stack matches German conditions.
When does a franchise system pass the recovery point?
The hardest question on a contracting franchise network is whether the contraction is mean-reverting or absorbing. A mean-reverting contraction stabilises at a smaller equilibrium and reopens at the margin once cost conditions improve. An absorbing contraction continues — slowly, then suddenly — because the variable driving it is structural, not cyclical.
Subway Deutschland sits closer to the absorbing case. The variables that drove the contraction — royalty stack on a low per-unit revenue base, absence of territorial protection, separation between system and franchisee economics — have not changed under Roark. They have intensified. The 2024 stabilisation at around 660 units reflects franchisees who can survive the current cost stack at the current royalty mechanics. It is not evidence that the architecture has recovered.
The recovery point on a franchise system is the moment when franchisee reinvestment exceeds the system's extraction rate. Subway Deutschland has not reached that point in fifteen years. The 2025 modernisation programme is system-side reinvestment, and it covers 40 of approximately 660 units. At that pace, the chain modernises its German footprint roughly once every sixteen years — a cadence the brand experience cannot defend against operators reinvesting on a five-to-seven-year cycle.
For an institutional investor evaluating the global Subway portfolio under Roark, the German case answers the underwriting question with unusual cleanliness. The architecture is fair-weather. The cost stack in any developed market will eventually press it. The sponsor's return mathematics requires that the architecture hold. Two of those three statements are observationally true. The third is the bet.
We read the franchisee P&L before the system P&L. The German Subway record is fifteen years of evidence on what happens when those layers diverge — and what the divergence costs at the moment of sponsor exit. The same divergence is now visible, in earlier form, across most of the global footprint. The German case is the preview.
Related research
- The Parent-Company Problem: KFC Germany, Taco Bell's USD 60 Million Miss, and the Franchise-DNA Variable
- The PE Playbook for Restaurant Chains: What Operators Should Read Before the Term Sheet Lands
- Five Guys Germany: market-entry lessons from a high-profile miss
- The Hundred-Unit Wall in DACH chained foodservice
Sources
- Subway Deutschland — Handelsregister filings, Frankfurt am Main
- Roark Capital — Subway acquisition announcement and follow-on press, 2024
- food-service.de — Subway DACH unit-count reporting, 2015–2024
- handelsdaten.de — Ranking Umsatz Systemgastronomie Deutschland 2024 (Subway revenue and unit base)
- HOGAPAGE — 2025 expansion plan for Subway Deutschland (25 new openings, 40 modernisations)
- Foodaktuell.ch / presseportal-schweiz.ch — Swiss master-franchise agreement with Convenience House Schweiz AG, 2024
- Reuters / Wall Street Journal — Subway franchise disputes coverage, 2020–2024
- Subway-Franchise-Deutschland — "20 Jahre Subway Deutschland" historical reporting (2019)