What Drives Car Wash Volume: A Demand-Driver Framework for Car Wash Feasibility Studies

Every car wash loan rests on one number, cars per day, and most feasibility studies derive it from a traffic count. The only published test of that method found that traffic explains about 6 percent of the variation in wash volume. The numbers that actually move a site are elsewhere: a customer base that has split into members who wash thirty times a year and retail drivers who wash three, a member book that turns over every 13 to 18 months, weather that removes half of retail volume on a rainy day, and a supply cycle that now sets demand per site corridor by corridor. This analysis ranks the six demand drivers behind car wash volume by what the evidence measures, separates measured effects from inherited rules of thumb, and shows how a lender-grade study should turn them into a volume and membership forecast that survives a credit committee.

An express tunnel sells the same three minutes to two very different customers. The member pays a flat monthly fee, washes when the car looks dirty, and keeps paying when it rains. The retail customer decides at the curb, pays once, and does not come back for months. Every forecast that blends them into a single cars-per-day figure is averaging two businesses that respond to different drivers, and it will be wrong in both directions: too high on retail in a wet year, too low on members in a corridor with no competitor.

The distinction matters more in 2026 than at any point since the express format took over the industry. At Mister Car Wash, the only operator that ever reported audited membership data, Unlimited Wash Club sales rose from 53 percent of wash sales in 2019 to 79 percent in the fourth quarter of 2025 (4). Across Rinsed’s panel of more than 3,000 sites, memberships were 71 percent of same-store revenue in the third quarter of 2026; membership revenue grew 9.8 percent and retail revenue fell 9.6 percent, the largest drop in five quarters (6). Monthly churn across that panel ran at 7.3 to 7.9 percent through 2026, which means a typical site replaces its entire member base every 12 to 14 months (6). New express openings peaked somewhere between 850 and 943 a year in 2022 and 2023 and have since fallen to roughly 550 (11, 12). Since 2023, at least a dozen municipalities have adopted moratoria, spacing rules or outright prohibitions on new washes (26).

That is the environment a feasibility study has to describe. A projection that shows a proposed tunnel reaching 3,000 members and 250 cars a day by month 24 needs to say which drivers deliver the members, which deliver the retail cars, how much of each is exposed to weather, and what happens to both when a third tunnel opens a mile away. This piece is the car wash counterpart of our hotel demand-driver framework, and the discipline is the same: name the drivers, separate the measured from the assumed, and show how each one enters the forecast.

Start with the mechanics, because they explain why the traditional model fails.

The frequency distribution has split. The last public point estimate of how often an American driver uses a professional wash is the International Carwash Association’s 13.6 washes a year per professional washer, from a survey fielded in October 2016 (1). It predates the subscription era and has never been updated in public. What replaced it is two numbers. ICA’s Industry Pulse puts the mean member at 2.6 washes a month, unchanged since May 2023, on a mean plan price of about $25 (2); DRB has cited 3.2, and newer Rinsed data 2.3 to 2.4 (8, 6). The retail customer washes two to five times a year on DRB’s data, about three on a 2023 industry presentation (8, 12). A member therefore washes roughly ten times as often as a retail customer and pays whether or not the car gets washed. Mister Car Wash’s final 10-K said the average member spends more than four times what a retail customer spends in one section and about seven times in another, an inconsistency the filing never resolved (4).

Member share rises for two reasons, and only one of them is growth. Rinsed reports same-store retail revenue down in 19 of the past 24 months (6). When the denominator shrinks, member share rises even if the member book is flat. A study that cites “membership now 75 percent of revenue” as evidence of demand strength has to show whether the member count grew or the retail count fell.

Conversion comes from retail, and retail is thinning. New members are made out of walk-up customers. Rinsed’s third-quarter 2026 report shows sites with more than 4,000 members converting 17.1 percent of retail customers, sites with 2,000 to 4,000 members converting 10.7 percent, and sites below 2,000 converting 3.3 percent (6). Scale compounds. The incumbent with the deep member book converts five times as efficiently as the new entrant across the street, out of the same shrinking retail pool. DRB describes the point around 70 to 80 percent member mix as a treadmill, where new sales mostly replace churn because the retail pool available for conversion has run dry (8).

Weather acts on one customer and barely touches the other. AMP’s data across more than 7,400 tunnels found rain cuts retail volume by more than half while member volume declines far less (7). Mister’s 2024 and 2025 quarterly comparable sales swung about five points on weather, and the swing came through the 21 to 26 percent of sales that were retail (4).

Supply acts on both, through capture and through conversion. A new tunnel takes retail cars from the incumbent and competes for the same walk-up pool that both need for member growth. Mister reported that 49 stores facing a competitor less than a year old comped down low single digits in the third quarter of 2025, while stores with no new competition comped up mid to high single digits (4).

The practical rule for underwriting follows directly. A car wash forecast is a stock and a flow. The stock is the member book: additions less churn, month by month, capped by the trade area’s households and the competitors’ existing books. The flow is retail cars: traffic, weather, seasonality and price. Everything that follows is about which drivers move which, and how well each has been measured.

Every car wash feasibility study lists demand drivers. Few distinguish between the ones whose effect on volume has been measured in audited or large-panel data and the ones inherited from a manufacturer’s site guide. The ranking below runs from best-documented to least. It is roughly the reverse of the order in which most studies present them.

Membership mechanics. This is the only driver with audited disclosure behind it, and the disclosure has just ended. Mister Car Wash reported member counts and member share every quarter from its 2021 IPO to its last 10-Q, for the first quarter of 2026: 2.468 million members, up 11 percent in a year, at 76 percent of wash sales (4). Leonard Green & Partners, already the holder of about 67 percent of the company, took the rest private at $7.00 a share and a $3.1 billion enterprise value, and the stock was delisted on May 19, 2026 (5). What remains is the vendor panels, and they agree on the shape. Industry monthly churn was 7.6 percent in the third and fourth quarters of 2025, 7.3 percent in the first quarter of 2026, a record 7.9 percent in the second and 7.8 percent in the third (6). Roughly half of it is involuntary: card failures ran at 3.0 to 3.6 percent a month across the Rinsed and DRB samples, which means close to half of member attrition is payment infrastructure, not customer intent (6, 8). Median member lifetime is 14.5 months, 13 months at sites with fewer than 1,000 members and 18 months at sites with more than 5,000 (6). The industry averaged about 2,600 members per site at the start of 2025 and peaked near 2,800 in the summer (8). Mister’s 2.271 million members across 548 sites at year-end 2025 imply about 4,140, a blended figure that includes immature stores (4). ZIPS reported about 2,250 per site at its bankruptcy filing, up from roughly 1,080 in 2020 (22).

Two findings in this data should govern every member forecast. First, survey intent is useless as an underwriting input: ICA finds 91 percent of members plan to keep their plan and fewer than 5 percent expect to cancel within 90 days, against measured monthly churn near 8 percent (2, 6). Second, promotional cohorts are not members. DRB’s cohort data show a constant $9.99 first-month offer producing 22 percent churn by the second month, and Rinsed found that nearly three in ten new signups are win-backs (8, 6). Grand-opening member counts annualized into a stabilized base are the car wash equivalent of a hotel study that extrapolates six tournament weekends into a year.

The price evidence cuts both ways. Mister raised its base tier by $3 to $22.99 in 2025, the first base increase since the program began, and its top tier to $39.99; member share held, revenue per member reached $29.56 by the third quarter, and management described a small period of elevated churn before it normalized (4). Rinsed, across its panel, found that locations with larger price increases generally saw higher churn (6). ICA’s fourth-quarter 2026 Pulse supplies the behavioral floor: members perceive break-even at about 2.2 washes a month, but 70 percent say they need at least three to justify the plan, and the members who fall short of their own threshold are the ones who cancel (2). A forecast at 2.6 washes a month and $25 to $30 a month is the base case. One at 3.0 or more is a throughput and chemical-cost stress, not a revenue upside.

Cancellation friction is also now a regulated variable. The FTC’s click-to-cancel rule was vacated in full by the Eighth Circuit on July 8, 2025, and the agency restarted rulemaking in 2026, but California’s amended automatic-renewal law took effect July 1, 2025 and similar statutes apply in Minnesota, Oregon and Colorado (32). Easier online cancellation raises voluntary churn for operators that still depend on in-person or phone cancellation, and a diligence request should ask which the sponsor uses.

What the capital markets did with this data is instructive. The Mister take-private cleared at roughly 9.0x trailing adjusted EBITDA by our arithmetic, inside the 8.0x to 10.5x precedent range in the bankers’ materials and in line with the oil-change and auto-service comparables they used (5). Net-lease car wash real estate listed at an average asking cap rate of 6.29 percent in September 2026, with tenant spreads from 5.36 percent for Quick Quack to 7.15 percent for ZIPS (25). The most member-dense operator in the country was valued like an auto-service chain, and net-lease spreads track tenant credit. There is no evidence of a separately priced recurring-revenue premium. Memberships are month-to-month and cancellable at will; legally and economically they are revenue, not a receivable.

Weather. Weather is the largest short-run swing in car wash volume, and its effect has been measured on panels large enough to trust the direction if not the decimal. AMP’s third-quarter 2025 analysis of more than 7,400 tunnel sites found that rainfall cuts retail volume by more than 50 percent, that member volume declines far less, and that when showers pass early, volume rebounds within four to six hours and can exceed normal later the same day (7). AMP does not disclose its rain threshold or the member figure. Rinsed found that the typical site had 10 percent more days with precipitation in the second quarter of 2025 than a year earlier; over the same quarter retail revenue fell 11.9 percent despite a 3 percent increase in retail price, while membership revenue rose 15.2 percent (6). That is a correlation across one quarter, not an elasticity, and Rinsed says so.

The operator record points the same way. Mister called weather a tailwind in the fourth quarter of 2024 and the first quarter of 2025, when comparable sales rose 6.0 percent in each, and a slight headwind in the second quarter of 2025, when they rose 1.2 percent and retail comps fell low double digits (4). Driven Brands attributed its car wash segment’s 4.1 percent same-store decline in the second quarter of 2024 primarily to lower retail volume from unfavorable weather, partially offset by membership revenue (21). Mister’s IPO prospectus had made the point four years earlier: members pay on a monthly basis, irrespective of the weather (4).

Snow is a lag effect, not a same-day one, and timing matters more than totals. Rinsed’s first-quarter 2025 report ties more snowfall days to more road salt and more retail washes (6). Operator reports of two to three times normal volume for weeks after a storm are common and anecdotal (35). The reverse case is documented too. Mister’s management explained a 1 percent comp in the first quarter of 2024 by noting that its northern stores got one or two big snowfalls that were quickly followed by rain that washed the salt away (4). A snow event followed by a thaw rain produces almost no wash demand. Pollen in the Southeast and lovebugs on the Gulf Coast are real and seasonal: Tidal Wave reported volume sometimes doubling in South Carolina’s pollen season, and the two lovebug flights last about four weeks each in April to May and August to September (35, 29). Pollen seasons across North America have lengthened by about 20 days since 1990 (30). None of this comes with a volume series a study can borrow.

Drought cuts the other way, and the rules are firmer than the demand data. California requires in-bay and conveyor washes built after January 1, 2014 to recycle at least 60 percent of wash and rinse water; metro Atlanta’s water district requires recycling at new conveyor washes, with Hapeville and Gainesville setting 50 percent reuse (27). At higher drought stages Los Angeles bans home vehicle washing outright, Davis, California limits washing to commercial recirculating washes, San Antonio restricts residential washing to once a week while certified commercial washes operate daily, and Raleigh allows only reclaimed commercial washes (27). Those rules favor a compliant tunnel in principle. In practice, no one has measured a demand shift from them, and Mister’s 10-K treats water restrictions as a supply risk, not a demand tailwind (4). For an arid-West site we treat reclaim compliance as a precondition and drought as roughly neutral.

Two things are missing from the public record. There is no quantified temperature threshold from any operator or vendor, and there is no published weather-normalization standard from the Appraisal Institute, ICA, SBA or any lender: nothing that defines washable days or applies NOAA normals to a car wash forecast. The data to build one exists. NOAA’s 1991 to 2020 Climate Normals publish, by station, the monthly number of days with precipitation at or above 0.01, 0.10, 0.25 and 0.50 inch and the number of snowfall days (28). The method we use is described in the forecast section below.

The underwriting consequence is simple and frequently missed. Weather risk in a car wash is retail risk. A site with 75 percent member revenue in Phoenix, which sees roughly 15 rainy business-hour days a year on one vendor’s count, has almost no weather exposure (40). A retail-heavy in-bay in a South Florida market whose wet season carries 60 to 70 percent of annual rainfall has a great deal (29). The haircut belongs on the retail line, sized to the local rain-day count, not on total revenue.

Supply and saturation. This driver has the best counts and the worst inferences. Census County Business Patterns counted 19,807 car wash employer establishments in 2023, up from 16,388 in 2017, a gain of 20.9 percent in six years; the Bureau has not released 2024 and has no planned date for it (20). Those counts exclude non-employer self-serve bays and washes attached to fuel stations, and they cannot tell an express tunnel from a full-service wash. ICA’s format census counted about 62,750 wash points in 2020, roughly 17,500 conveyors, 29,000 in-bay automatics and 16,250 self-serves, more than three times the Census figure (38). No federal dataset identifies the conveyor roster that actually competes with a proposed tunnel. Every per-tunnel ratio in every study is hand-built, and the study should say so.

The build cycle has turned. Sonny’s data presented in October 2025 show new tunnel builds rising from about 220 in 2015 to more than 900 in 2023; Sonny’s CEO Curt Hutchins says annual builds, after topping out at 850 to 900, now hover closer to 550, plus or minus 100, and well over 90 percent of them are express (11, 12). A widely repeated figure of 943 openings in 2022, which we have cited ourselves, cannot be traced to any primary document, and the sources disagree on whether the peak year was 2022 or 2023 (11). Credit tightened before the distress headlines: SBA 7(a) approvals to car washes fell from $676 million across 391 loans in fiscal 2021 to $272 million across 243 loans in fiscal 2025 (31). Mister’s CFO told investors on the third-quarter 2025 call that new competitor openings within three miles of its stores were running an estimated 40 percent below the prior year (4).

What the evidence says about a specific competitor is the finding that matters. WashIndex, tracking 6,559 openings across 2024 and 2025 with Google review velocity as a proxy for volume, finds that a single new competitor produces almost no measurable damage; incumbents within a mile of a new express tunnel lose about 14 percent of review velocity, and the third tunnel roughly doubles the per-incumbent damage (14). Review velocity is not a car count, but the shape matches the operator disclosure: Mister’s stores facing a competitor less than a year old comped down low single digits while the rest comped up mid to high single digits (4). An incumbent absorbs the first competitor, weathers the second and takes a material step down on the third. A study that models cannibalization as a flat percentage per competitor is modeling the wrong curve.

Saturation is a submarket condition, not a national one. Steve Gaudreau of Brink Results, who writes the Top 50 commentary for Professional Carwashing & Detailing, calls the U.S. market overall not overbuilt and names the overbuilt suburban submarket as the problem: within a population of 50,000, four washes within two miles of one another, one closed, two barely alive and one doing well (11). His threshold for an overbuilt submarket is fewer than 25,000 people per conveyor wash, down from an older rule of 50,000 with no competitor within three miles (11). Colorado Springs, with 73 conveyor washes in a metro of about 750,000, works out to roughly 10,300 people per conveyor and is the most cited case (11). Warren, Michigan counted 25 car washes in 35 square miles; Cape Coral, Florida counted 16 standalone washes, nine attached to fuel and 12 more in permitting when it adopted its ordinance; Hemet, California counted 26 existing, entitled or pending on one corridor (35, 26). The field signal, in Gaudreau’s words, is price wars and stagnant revenue. In ours, it is the appearance of a $3 to $7 wash on the corridor.

The municipalities have noticed, and their response has shifted from temporary to permanent. Hemet adopted a 45-day urgency moratorium in August 2023, extended it a year, and in August 2024 ended new full- and self-service washes in all commercial and manufacturing zones (26). Perrysburg, Ohio ran a moratorium from January 2023 and extended it to the end of 2025; Hesperia, California adopted one in February 2024; Birmingham, Alabama froze automated wash licenses in March 2024 and extended the freeze twice, citing traffic and pedestrian safety on its corridors (26). Cape Coral’s April 2024 moratorium expired in January 2025 and was replaced by permanent location and enclosure standards (26). McAllen, Texas set a half-mile minimum spacing; Mission, Texas set one mile in January 2026, adopted a 180-day moratorium in March, repealed it in April after a federal lawsuit alleging retaliation and violation of vested rights, and then widened spacing to two miles, which in a city that size is a ban (26). Warren, Michigan adopted up to eight months on new gas stations and car washes in May 2026, and Avon, Ohio extended a moratorium on car washes and self-storage in June 2026 while it studies caps (26). Texas has moved the other way at the state level, restricting municipal development moratoria under HB 2559 (26).

The pattern across the list is consistent. Most temporary moratoria lapsed and were replaced by spacing rules, conditional-use requirements or zone prohibitions. Those tools lower the entitlement probability for a new entrant and raise the defensibility of the incumbent across the street, and a study should price both effects.

Consumer frequency and penetration. This is the driver the industry quotes most and has measured least recently. ICA’s long-run series of drivers who wash “most often” at a professional facility rose almost without interruption from 47.6 percent in 1996 to 71.6 percent in 2014, dipped to 69.2 percent in 2016, was reported at 77 percent in 2019 and 79 percent in 2023, and fell to 73 percent in ICA’s fourth-quarter 2024 Pulse (1, 3, 2). ICA named a decline in price satisfaction among transactional customers as the key correlating factor; only 30 percent of non-members were very satisfied with price, the lowest reading since early 2023 (2). The 2016 survey had already found cost to be the reason 90 percent of home washers gave for washing at home, up from 81 percent two years earlier, and that 40 percent of all washers used both a professional wash and their driveway (1). Two decades of conversion from home washing have stalled. The next decade of volume growth has to come from frequency and share shift, not from recruiting people who have never used a car wash.

Three cautions apply to any study that uses these figures. The 1996 to 2016 series sampled vehicle owners; the Pulse waves sample people who used a car wash in the past year, so part of the drop from 79 to 73 percent may be survey design (2). ICA’s “83 percent of Americans washed at a professional car wash” measures past-year incidence, not the main washing method, and the two get mixed in marketing material (3). And a figure that circulates on consultant websites, 15 to 30 wash transactions per registered vehicle per year as “the standard institutional benchmark,” has no primary source and sits above ICA’s 13.6 washes per professional washer, which is a narrower base (1). We have never found it in a published dataset and do not use it.

On who washes more, the public evidence supports two variables. In ICA’s 2016 survey, 32 percent of owners of vehicles zero to three years old washed professionally a few times a month or more, against 18 percent of owners of vehicles over ten years old, and ICA attributed the rise in frequency between 2014 and 2016 largely to new-car owners (1). On income, the observable comes from the operator side: Mister’s third-quarter 2025 retail decline was concentrated, by management’s account, in lower-income demographic locations (4). Claims about commute length, vehicle value, generation and urban versus suburban residence have no public frequency data behind them, and a study that uses them is reporting the analyst’s judgment.

Seasonality has no audited national curve either. The split most often repeated, winter 32 percent, spring 25, summer 25 and fall 18, traces to a 2014 trade-press citation whose method and sample cannot be verified, and it conflicts with operator surveys in which 40 percent name spring as the busiest season and 31 percent winter (36). Full-service operator surveys from 2013 to 2022 find winter busiest and fall slowest overall, with spring the slowest in the Mountain region and the busiest in the Midwest (36). Single-site data suggest peak-to-trough ratios of roughly 1.5x to 1.9x for retail-heavy formats in snow markets (36). At the subscription-heavy end of the industry, Mister’s 2025 quarterly revenue moved in a band of $261 million to $265 million (4). The seasonality a study should model is the seasonality of its retail line, and it should note that 60 to 75 percent of annual member growth typically arrives in the first half of the year, so a site opening in the fourth quarter ramps more slowly than one opening in March (4).

Vehicle population and trade-area demographics. The base is large, slow-growing and aging, and almost none of that changes wash demand. S&P Global Mobility counts 289 million light vehicles in operation in 2025, up 3 million in a year, at a record average age of 12.8 years (16). FHWA reports 297.5 million registered vehicles of all types for 2024, but its 2024 table switched from estimated to state-reported data, and part of the jump from 2023 is methodological (17). The two series should never be blended. Census ACS data imply about 1.83 vehicles available per household; dividing FHWA registrations by households would give 2.24, because registrations include commercial, government and duplicate vehicles (18, 17). Trade-area studies should use the ACS household-vehicle tables and say which one.

The aging fleet is a tailwind for repair shops and, on the only direct evidence, a mild headwind for washes, since newer-vehicle owners wash more (1). Electric vehicles are about 1.1 percent of light vehicles in operation and have no public primary data on wash frequency at all; the claim that EV owners wash 15 percent more often circulates without a source, sample or year, and one widely copied citation to a “ChargePoint 2025 EV Driver Behavior Report” refers to a document that does not appear to exist (16). We treat EV adoption as demand-neutral in every base case. The one structural point is that an EV removes the fuel stop, which weakens the capture model of a convenience-store in-bay and does nothing to a standalone tunnel.

Population growth drove supply into the Sun Belt, and supply has now caught up with it. The South grew 6.0 percent from April 2020 to July 2025, nearly double the national 3.1 percent, with the growth concentrated in the outlying counties of metro areas, which is exactly where express tunnels cluster (19). Private car wash establishments rose 26 percent in Texas and 20 percent in Florida between early 2021 and early 2026, and Mister’s 548 stores at year-end 2025 were led by Texas at 98, Florida at 86 and California at 66 (20, 4). The in-migration that justified that supply is slowing. National growth fell to 0.5 percent in the year to July 2025 on a collapse in net international migration, and Florida’s net domestic migration fell from 310,892 in 2022 to 22,517 in 2025 (19). For a lender, a Sun Belt address is no longer a positive credit factor on its own. The test is the competitive census and the permit pipeline inside the drive-time trade area, set against a growth rate that is a fraction of what it was when the corridor was planned.

The demographic thresholds that circulate in the industry are consistent with one another and validated by nothing. No public operator 10-K or investor presentation discloses a population, household or income threshold; Mister described only a rigorous process for opening greenfield locations and, in the same quarter, acknowledged site selection errors at a subset of stores (4). The broker and manufacturer guides converge on about 30,000 people and $60,000 median household income within three miles, fewer than two competing tunnels in the same ring, and 25,000 or more vehicles a day on the frontage (15, 14). That consistency likely reflects copying rather than independent testing. The thresholds work as screening floors. They are not demand estimates. The trade area itself should be drawn as a drive-time polygon of five to ten minutes on the road network, with the one-, three- and five-mile rings reported for the lender’s familiarity; a former operator who competed against Mister describes a three-mile radius and a nine-minute drive time as roughly equivalent in the suburbs and not at all in Houston or Atlanta (14, 37).

Traffic counts and site capture. The driver every study leads with ranks last, because its statistical support is the thinnest on the list. The traffic-capture method, expected washes equal to AADT times a capture rate, is described by appraisers as an intuitive tool and is the default in most lender studies (33). The capture rates it uses trace to two sources. The first is a December 2014 analysis by Retail Petroleum Consultants of express and flex-serve washes in California, Arizona and Nevada using 2011 to 2013 data, which found express capture of about 1.7 percent at 10,000 vehicles a day falling to about 0.5 percent at 80,000, with flex-serve capture about 60 percent as high (9). The second is a set of forum posts and trade articles by consultant Robert Roman, who reports observed capture between 0.4 and 1.81 percent with an average near 0.7, a conveyor range of 0.7 to 1.4 percent, in-bay automatics near 0.4 percent, and a regression in which movement in traffic explained 6.0 percent of movement in volume (10). Roman calls the method naive. It is the industry’s best-known number, and its primary source is a forum thread.

The thresholds are no better. For an express tunnel the minimum frontage count cited runs from 20,000 vehicles a day in a manufacturer’s 2026 guide to 25,000 from Sonny’s to 30,000 from Autobell’s chief operating officer (15, 11). A 50 percent spread between floors means the threshold is a screening heuristic, and the sources are describing different things: a viable site versus an A-plus site. In-bay thresholds cluster at 12,500 to 20,000 and self-serve at 5,000 to 15,000 (10, 15). DRB, whose point-of-sale data cover thousands of sites, argues that car count has almost zero predictive value, which directly contradicts the manufacturers that publish the thresholds (8).

The Retail Petroleum curve is the one decision-relevant finding, because of its shape. At 80,000 vehicles a day and 0.5 percent capture, implied volume is about 400 cars a day; at 10,000 and 1.7 percent, about 170. An eightfold increase in traffic buys less than two and a half times the volume, because stacking and site size cap what can be processed, and because the share of passing drivers who live in the trade area falls as the road gets busier. The data predate the membership era and should be revalidated against current comparables, but nothing published since contradicts the shape. Roman adds a second correction that most studies miss: capture changes over a site’s life, from new-customer trial to loyalty, and in his worked example it rose from 0.34 percent in month three to over 0.8 percent by year three (10). A capture rate lifted from a mature comparable therefore overstates year-one volume at a new build.

The site factors that change capture have consistent direction and no published effect size. Speed at or below 45 miles an hour, full two-direction access without a median or U-turn, signalized or corner entry, visibility, and stacking for a dozen or more vehicles beyond the pay stations are supported in the same direction by operators, manufacturers and data vendors (11, 14, 15). Going-home side, co-tenancy with a grocery or big-box anchor and daytime population are directionally accepted and quantified by nobody; one 2026 investor guide’s claim that left-turn entry costs 30 to 50 percent of potential visits is unsourced (15). A study that applies a numeric uplift for any of these is using an assumed value and should disclose it as one.

Two measurement points matter more than any benchmark. State DOT counts are two-way unless expressly directional, and most are factored estimates from short counts rather than continuous ones (17); applying a one-way capture benchmark to a two-way count doubles the projection. And a count taken on an undivided road does not transfer to a divided one. Roman’s corner-lot formula discounts each road’s AADT by 35 percent before adding them, and the trade press puts the loss from a divided or obstructed road at half the count (10, 11). Usable traffic, not published traffic, is the input.

Corridor saturation and member churn are the two risks every car wash study names and almost none quantifies. The best way to size them is the one case where the whole cycle has been documented in a bankruptcy court.

ZIPS Car Wash filed for Chapter 11 in the Northern District of Texas on February 5, 2025 with $653.9 million of funded debt and about $1 million of cash (22). It operated more than 260 locations in 23 states. It had about 625,000 unlimited members generating more than two-thirds of roughly $303 million in annual revenue, and its members per site had doubled from about 1,080 in May 2020 to about 2,250 in June 2024 (22). By every metric a feasibility study would ordinarily cite, the member engine worked. What it could not do was outrun the corridor. The first-day declaration cited about 900 new car wash sites opening nationally every year for five years, competition, labor inflation, softer discretionary spending, off-market leases and churn (22). Annual cash interest had risen from $59 million to $93 million, and the term loans had matured on December 31, 2024 (22). The company sought to reject leases on 41 locations, about 16 percent of its fleet, cut about $279 million of debt, and emerged on April 30, 2025 under lender ownership with roughly 230 sites (22). A year and a half later its real estate was still the widest-priced of any major tenant in the net-lease market, at a 7.15 percent average asking cap rate against 5.36 percent for Quick Quack (25).

Three lessons follow. A member share above two-thirds did not protect the credit, because a member book with 7 to 8 percent monthly churn is a 13-to-18-month annuity that has to be refilled from a retail pool the new competitors are also fishing in. Supply growth at the corridor level, not the national level, is what broke the model, and the corridors that broke were the ones where ZIPS and its competitors had all read the same site guide. And the leverage that financed the 2021 and 2022 build cohort was sized to a stabilized year that the ramp never reached.

The same cycle is visible across the rest of the record. Driven Brands sold its 383-site U.S. car wash business to Whistle Express for $385 million, including a $130 million seller note, in April 2025 after recording impairments on it; Centerview’s materials for the Mister transaction put that sale at about 7.7x 2024 EBITDA (21, 24, 5). Mister itself slowed greenfield growth, acknowledged site selection errors and was taken private at a multiple its bankers benchmarked against oil-change chains (4, 5). The private equity money has not left the sector: KKR agreed in June 2024 to pay $850 million for a minority stake in Quick Quack, and the top 25 operators hold about 70 percent of the sites in the industry’s Top 50 (23, 24). It has become more selective about corridors, which is the point.

The ZIPS numbers give a lender a usable downside case: a corridor where the operator’s own member density doubled over four years while per-site economics fell far enough to reject one lease in six. That is the stress to apply wherever the subject would be the third tunnel in its submarket, or wherever the sponsor’s pro forma stabilizes in year two.

The SBA data, so far, do not show the cycle. Resolved SBA car wash loans since fiscal 2010 have charged off at rates at or below the program average on every published tabulation: 4.4 percent of 7(a) and 504 loans combined through March 2026 on one, 5.25 percent of 7(a) loans against 7.45 percent for all industries on another (31). The windows and program mixes differ and should not be averaged. Both share one caveat. The average time to charge-off is about four years, so the fiscal 2021 and 2022 vintages, the largest in the program’s history for this industry, are not yet resolved (31). The current rates understate the ultimate losses of the boom cohort by an amount nobody can yet measure.

Everything above is input. What a lender receives is projected cars per day, members and revenue for each of the first five operating years, and the method that connects the two is where most car wash studies fail the file.

Build demand as two groups. Member washes are members times 2.3 to 2.6 washes a month; retail washes are unique retail visitors times two to five washes a year (2, 8). The 13.6 washes a year figure is a 2016 historical anchor, and a study that uses it should label it as one (1). The professional-penetration input is 73 percent on the most recent reading and 79 percent on the prior one; pick one, disclose the sample basis, and do not average them (2, 3).

Run all three site-volume methods and show them side by side. Traffic capture, AADT times a capture band drawn from the Retail Petroleum curve and then overridden by comparable-site evidence, is the upper-bound sanity check (9, 10). Trade-area penetration, vehicles in the drive-time polygon times professional penetration times frequency times the subject’s fair share of conveyor capacity at the opening date, is the demand ceiling (14, 34). Comparable calibration, named competitor sites with observed car counts and member counts, governs the selected figure. When the three disagree by more than 20 percent, the study explains the gap rather than averaging it. A capture rate that exceeds the subject’s fair share without strong site-level justification is the single most common reviewer flag we see.

Model the member book as a stock with churn. Members at the end of a month equal members at the start, plus gross additions from retail conversion, less voluntary and card-failure churn shown separately, with base churn at 7.5 to 8.0 percent a month and promotional cohorts carrying a 20 percent second-month haircut (6, 8). Conversion is tied to the competitors’ member density, using the Rinsed bands of roughly 3, 10 and 17 percent, not a national average (6). Stabilization for a typical site is 2,500 to 3,000 members at month 30 to 36; anything above 4,000 is justified by trade-area arithmetic, not by a sponsor’s experience elsewhere (8, 6). About 70 percent of members use a single home location, so the book is sized to the site’s own trade area, not the brand’s (2).

Underwrite the ramp to cohort evidence. Mister’s 10-Ks state that greenfield sites reach average unit volume after approximately three full years (4). Tommy’s Express franchise disclosure for 2025 shows sites open 12 to 23 months averaging about 51 percent of the revenue of sites open 60 months or more, 63 percent at 24 to 35 months, 81 percent at 36 to 47 and 89 percent at 48 to 59; first-year averages fell from about $1.44 to $1.48 million on 2021 and 2023 data to $1.09 to $1.14 million on 2024 and 2025 data (13). The practitioner convention of year one at 60 to 70 percent of stabilized and year two at 80 to 90 percent, which appears in our own published car wash research, sits well above the cohort evidence, and one advisory firm argues that real openings take 30 to 42 months against the 24 assumed in most pro formas (39). We now use year one at 50 to 60 percent of mature revenue as the base case and size the interest reserve to carry debt service through month 24 to 36. A two-year stabilization assumption is aggressive, and a study that shows it should show the local openings that support it.

Normalize weather on the retail line only. Count the station’s days with precipitation at or above 0.10 inch from the NOAA 1991 to 2020 normals, since trace precipitation rarely stops washing; haircut retail volume on those days by about 50 percent per the AMP finding and member volume by a small fraction of that; for snow markets, add a post-storm uplift only after calibrating it to comparable sites and run a mild-winter sensitivity case; for Gulf and Florida markets, run a wet-season case (28, 7). This is our method, not a published standard, and we say so in the report. Pollen and lovebug peaks are upside, not base case.

Screen saturation against supply at the opening date. Build the conveyor roster by hand, count tunnels per 10,000 households and people per conveyor in the drive-time trade area, include every permitted and proposed tunnel, and apply an explicit step-down when the subject would be the third or later tunnel in its submarket (14, 11). Check for moratoria, spacing rules and conditional-use requirements in the jurisdiction, and note that the same rules that threaten the entitlement protect the subject once it is built (26).

Report the retail exposure. State the share of projected revenue that is retail and therefore exposed to weather, price and competition, and run the downside with retail at the trailing panel trend, which has been negative in 19 of 24 months (6). Then show debt service coverage through month 24.

Separate the real estate from the business. Market rent at a net-lease cap rate values the land and building; EBITDA after market rent at a business multiple values the operation (25, 5). Capitalizing operator cash flow at a 6.3 percent cap rate values a business at a real estate rate, and a credit officer who has read the Mister information statement will notice.

A study built this way produces a forecast a credit officer can test line by line, and one that an SBA or USDA review will recognize, because those reviews ask the same questions: is demand split between members and retail, is the ramp supported, was supply measured at the opening date, and how much of the cash flow is exposed to the weather and to the competitor across the street.

Five conclusions follow for any car wash study written in the fourth quarter of 2026.

Cars per day should be underwritten as two numbers. A forecast that does not separate the member book from retail cars has not described the business, and it cannot be stressed, because the two lines respond to different drivers.

Membership should be underwritten as an annuity with 8 percent monthly churn, not as contracted revenue. Measured churn of 7.3 to 7.9 percent through 2026, half of it card failure, a median member lifetime of 14.5 months and the ZIPS record are the facts; survey intent and grand-opening counts are not inputs (6, 22).

Retail should be underwritten as declining and weather-exposed. The panels show retail revenue down in 19 of 24 months and down 9.6 percent in the most recent quarter; a base case that assumes retail frequency recovers to 2023 levels needs local evidence behind it (6).

Traffic should be context, not the model. The only published fit explains 6 percent of volume; capture bands belong in the sanity check, comparables govern the figure, and usable traffic replaces the published count (10, 9).

Supply should be measured at the opening date, in the corridor, by hand. The national build count is falling, which is context. The permitted tunnel a mile away, the incumbent’s 4,000-member book and the city’s spacing ordinance are what move the subject’s conversion, capture and entitlement (14, 26).

The purpose of a feasibility study in a car wash credit file is not to confirm the sponsor’s member curve. It is to tell the lender which part of cars per day is bought with a monthly fee and which is bought at the curb, how much of each the trade area can deliver, and what both are worth when it rains and when the third tunnel opens. That is the analysis MMCG Invest delivers in every car wash feasibility study prepared for SBA, USDA and conventional lenders.

How many washes a year does the average driver buy at a professional car wash?

The last public ICA estimate is 13.6 a year per professional washer, from 2016 (1). It no longer describes the market. Members wash about 2.6 times a month, roughly 31 a year; retail customers wash two to five times a year (2, 8). Model the two groups separately.

What capture rate should a car wash feasibility study use?

There is no published trade standard. The only multi-site dataset shows express capture falling from about 1.7 percent at 10,000 vehicles a day to 0.5 percent at 80,000, and the consultant whose posts produced the 0.7 percent rule of thumb reports that traffic explains 6 percent of volume (9, 10). Use a band tied to the traffic level, override it with named comparables, and run coverage at the low end.

How many members does a typical express tunnel stabilize at?

The industry averaged about 2,600 members per site at the start of 2025, peaking near 2,800 in summer (8). A conservative stabilized assumption is 2,500 to 3,000 at month 30 to 36. Mister’s blended figure near 4,140 includes its most mature stores and should not be borrowed (4).

How much does weather move a car wash?

Rain removes more than half of retail volume on the day, and member volume falls far less (7). At an operator with about three-quarters of sales from members, that produced a comparable-sales swing of roughly five points between a good-weather quarter and a bad one (4). The exposure is the retail share.

Does a new competitor automatically cut volume?

Not the first one. The published cannibalization data show almost no measurable damage from a single new competitor, about a 14 percent loss in review velocity for incumbents within a mile of a new express tunnel, and roughly double the per-incumbent damage once a third tunnel opens (14). Mister’s stores facing a competitor under a year old comped down low single digits while the rest comped up (4).

Is the car wash sector oversupplied?

Nationally, tunnel builds have fallen from 850 to 900 a year to about 550, and SBA charge-offs remain at or below the program average (11, 31). At the corridor level, several suburban submarkets are overbuilt by the trade’s own 25,000-people-per-conveyor test, and the fiscal 2021 and 2022 loan vintages have not yet seasoned (11, 31).

Do car wash memberships earn a lower cap rate or a higher multiple?

No published evidence supports a separately priced membership premium. The most member-dense operator sold at roughly 9.0x trailing EBITDA, in line with auto-service comparables, and net-lease asking cap rates track tenant credit, from 5.36 percent for Quick Quack to 7.15 percent for ZIPS in September 2026 (5, 25).

Does SBA require a feasibility study for a car wash?

SBA treats car washes as special-purpose property, which raises the 504 borrower contribution to 15 percent under the current SOP 50 10, and most study requirements come from the lender’s own credit policy (31). What the reviewer will test is whether capture exceeds fair share, whether the ramp is supported, and whether supply was counted at the opening date.

Michal Mohelsky, J.D., FMVA, Practicing Affiliate of the Appraisal Institute

Principal, MMCG Invest, LLC

Published November 2026. How this analysis was prepared.

Michal Mohelsky, J.D., FMVA (2026). What Drives Car Wash Volume: A Demand-Driver Framework for Car Wash Feasibility Studies. MMCG Invest, LLC. https://www.mmcginvest.com/post/what-drives-car-wash-volume-a-demand-driver-framework-for-car-wash-feasibility-studies