The Business Churn Report 2026: Formation, Exit, and Survival Rates by Trade

By Christoph Totter | Last verified July 2026 | This report is refreshed annually when new federal data vintages publish.

Executive Summary

Every buyer diligencing a trades company eventually asks the same three questions: how many new competitors show up each year, how many existing ones disappear, and how long does a typical operator in this industry actually live? Those questions have precise answers, because the federal government publishes business formation, exit, and survival rates by trade in three separate datasets that almost nobody in the lower middle market ever opens. This report opens them, computes the figures, and translates the results into acquisition terms.

We pulled the raw files from the Census Bureau’s Business Formation Statistics, the Census Bureau’s Business Dynamics Statistics, and the Bureau of Labor Statistics Business Employment Dynamics establishment-age tables. From those files we computed formation, exit, and survival figures for the industries our readers actually acquire in: specialty trade contractors, landscaping and building services, auto repair, personal services, trucking, and home health. Every number below appears in, or derives arithmetically from, a public federal table. Nothing is estimated, modeled, or interpolated.

Five headline findings emerged from the computation.

The post-COVID formation wave never receded. Total US business applications reached 5,671,836 in 2025, per Census BFS monthly files. That total sits 62.1 percent above the 2019 baseline. The 2021 figure that most commentary still describes as the peak was 5,390,816 applications. It turned out not to be a peak at all, because both 2023 and 2025 exceeded it. First-half 2026 applications ran 12.1 percent ahead of the first half of 2025, which points toward an eighth consecutive year of elevated formation.

Construction formation settled onto a permanently higher plateau. Construction business applications hit 519,090 in 2025, per Census BFS. That count sits roughly 35.8 percent above the sector’s 2019 filings. More than half of 2025 construction applications carried the high-propensity markers that predict actual payroll formation. The economy-wide high-propensity share was roughly 30 percent. Construction applications appear to convert into real employer businesses at nearly twice the average rate, which means the sector’s application wave should be read as genuine competitive entry rather than paper filings.

Net entry is running hot in the trades. Construction added a net 77,571 establishments between 2019 and 2023, per Census BDS. Averaged across the window, that works out to about 15,514 net new construction competitors per year. Building equipment contractors, the NAICS code containing HVAC, plumbing, and electrical, added a net average of roughly 3,098 establishments per year over the same window. Landscaping, janitorial, and other building services added an average of about 4,475 net establishments per year.

Survival is better than the folklore, and improving. Averaged across three decades of BLS cohorts, about 48 of every 100 new construction establishments were still operating at year five. About 32 of 100 reached year ten. The construction cohort that opened in the year ended March 2020 posted 56.5 percent survival at year five, the strongest five-year mark in the recent data we examined. The commonly repeated claim that nine in ten businesses fail in their first year appears in no federal dataset, and Part 4 of this report supplies the correction.

Churn separates roll-up quality from customer-acquisition fragility. Auto repair posted the lowest churn of any vertical we measured, with a 2023 entry rate of 10.4 percent per Census BDS. Its 2023 exit rate was 9.1 percent. Its cumulative net establishment growth across the five boom years was just 3.8 percent. General freight trucking sat at the opposite pole, swinging from a net gain of 12,152 establishments in 2022 to a net loss of 6 in 2023. One of those industries supports durable customer relationships and platform math. The other burns diligence models for fuel.

Three Numbers to Quote

If you retain nothing else from this report, retain these three figures. Each one is computed directly from a named federal file, and each one changes how a trades acquisition should be underwritten.

1. 48 of 100. Averaged across BLS Business Employment Dynamics cohorts opened between 1994 and 2024, about 48 of every 100 new construction establishments survive to year five. About 32 of 100 survive to year ten. The cohort opened in the year ended March 2020 did better, reaching 56.5 percent five-year survival.
2. 15,514 net new construction competitors per year. Census Business Dynamics Statistics show construction added 77,571 more establishments than it lost from 2019 through 2023. That is the fastest sustained net entry of any trade-heavy sector in the file.
3. Plus 12,152 to minus 6. General freight trucking swung from a net gain of 12,152 establishments in 2022 to a net loss of 6 establishments in 2023, per Census BDS. Nearly one in five trucking establishments exited in that single year, the sharpest churn reversal in the dataset.

Methodology and Dataset Vintages

This report is computed, not compiled. We downloaded the raw public files, summed and rated them in pandas, and report only figures that appear in or derive arithmetically from the source tables. That distinction matters, because most churn statistics circulating in M&A marketing materials are second-hand paraphrases of paraphrases, often decades stale. Everything below cites its file, its vintage, and its granularity limit.

The three source programs

Vintage notes

BFS is the freshest series. The monthly file we pulled contains not-seasonally-adjusted application counts through June 2026. Annual figures in this report are sums of the twelve unadjusted monthly values. Figures for 2026 cover January through June only, and every table labels them as such.

BDS carries a roughly two-year lag because it is built from the Census Bureau’s confidential Longitudinal Business Database. The 2023 vintage is the latest available as of this writing. When we say “2023 births,” we mean establishments that entered between March 2022 and March 2023 under BDS timing conventions.

BED Table 7 tracks every private-sector establishment cohort opened since the year ended March 1994 and reports the share still operating each subsequent March. The tables we used run through March 2025. One retrieval disclosure belongs here: BLS blocks automated retrieval of these text files, so we pulled them from Internet Archive captures of the bls.gov originals dated between May and July 2026. The URLs cited throughout this report resolve to the live BLS files when opened in a browser, and the archived captures match those live files.

Honest granularity flags

Federal data does not go infinitely deep, and pretending otherwise is how bad reports get written. Five rules govern every figure on this page.

BFS is sector-level only. There is no “HVAC applications” series anywhere in the public files. When we discuss formation trends for the trades, the BFS number covers all of NAICS 23 construction, and we say so each time.

BDS reaches 4-digit NAICS, and no deeper. This is where the verticals live: code 2382, building equipment contractors, covers HVAC, plumbing, and electrical together. Code 2381 covers roofing, siding, and other exterior work. Code 5617 covers landscaping, janitorial, and pest control together. Code 8111 is automotive repair and maintenance. Four digits is the finest cut the public BDS files support with a full time series, and we do not fabricate 6-digit splits that the data cannot deliver.

BED survival is supersector-level. Survival curves exist for construction as a whole, not for plumbing contractors specifically. A plumbing startup is a construction-sector establishment in these tables, and we label every survival figure with the sector it actually measures. To state the limit plainly: no vertical-specific survival curve exists in any public federal file, and any source quoting one is interpolating.

Applications are not businesses. A BFS application is an Employer Identification Number request. Census flags a subset as high-propensity applications, meaning they carry characteristics associated with becoming payroll employers, such as being filed by a corporate entity or indicating planned wages. We report both series side by side. Note one definitional quirk: all applications in accommodation and food services are classified as high-propensity by construction of the series, so NAICS 72 shows identical totals in both columns.

Establishments are not firms. BDS counts locations. A 40-branch HVAC platform is 40 establishments and 1 firm in these tables. For the single-location businesses that dominate the trades, the two counts sit close together: building equipment contractors showed 183,573 establishments in 2023, per Census BDS. The same code showed 177,201 firms. The gap between those two counts is a rough index of how much consolidation has already happened in a code.

Part 1: The Formation Boom, Measured

Total applications never came back down

The COVID-era surge in business formation is usually narrated as a spike that faded. The Census BFS data does not support that narrative in any year of the series.

Annual sums of not-seasonally-adjusted monthly values from the Census BFS BA series.

Read the table top to bottom and the supposed reversion never arrives. The year usually described as the peak, 2021, was exceeded in 2023. It was exceeded again in 2025. The first half of 2026 produced 3,232,301 applications, per Census BFS. The first half of 2025 had produced 2,883,815. If the second half of 2026 holds the same pace, the year would set another record. Seven consecutive years of elevated filing is not an episode. It is a regime change in how many Americans attempt to start businesses, and every trades acquisition closed in this decade is being underwritten inside that regime.

Sector by sector: where the trades entrepreneurship wave actually landed

The aggregate number hides the sector texture, and the sector texture is what a buyer actually needs. The table below tracks the six sectors that contain the verticals this site covers.

Census BFS BA series, annual sums of monthly values, not seasonally adjusted.

Three patterns stand out for anyone buying in the trades.

First, construction was the most restrained sector in the boom and the most persistent afterward. Its 2021 surge over 2019 came to 30.7 percent, per Census BFS. That was the smallest jump among the sectors in the table. Its 2023 count of 545,891 applications was the highest in the sector’s history in this series. Construction did not spike and revert. It stepped up and stayed, and the step has now persisted long enough that treating it as temporary would be an underwriting error.

Second, the sectors housing landscaping, janitorial, auto repair, and personal services boomed harder than construction did. Administrative and support services, the sector containing landscaping and building services, filed 58.8 percent more applications in 2025 than in 2019, per Census BFS. Other services, the sector containing auto repair and personal care, ran 47.7 percent above its baseline. Anyone underwriting a landscaping or shop acquisition today is pricing against a materially deeper pool of would-be entrants than existed in 2019, whether or not those entrants ever hire an employee.

Third, transportation tells a boom-and-correction story that the other sectors do not. Applications in transportation and warehousing rose 99.7 percent from 2019 to 2021, per Census BFS. The 2021 count reached 477,300 filings. Applications then fell for three straight years, landing at 366,727 by 2025. That reversal shows up again, violently, in the BDS exit data covered in Part 3, and it is the clearest demonstration in this report that formation data carries a forward warning for exit data.

High-propensity applications: the signal inside the noise

Skeptics of the formation boom argue, reasonably, that anyone can request an EIN in an afternoon. The Census Bureau’s answer to that objection is the high-propensity series, which isolates applications carrying characteristics that historically predict payroll employment: corporate entity structure, indicated wage dates, or industry codes with high historical conversion.

Census BFS HBA series, annual sums of monthly values.

The high-propensity lens cuts the boom roughly in half economy-wide. Total applications rose 62.1 percent from 2019 to 2025, per Census BFS. High-propensity applications rose only 29.8 percent over the same span. A large share of the marginal applications appears to be gig work, side ventures, and single-member LLCs that never intend to hire anyone.

Construction is the exception that matters for this audience. High-propensity filings in construction totaled 266,773 in 2025, per Census BFS. That equals 51.4 percent of the sector’s 519,090 total applications. The economy-wide high-propensity share was 30.1 percent. A construction application is far more likely to represent a real future employer than an application in almost any other sector, which is consistent with the licensing, insurance, and capital requirements of entering the trades. When a buyer sees half a million construction applications a year, roughly half of them should be read as genuine competitive entries in formation.

Two other sectors deserve a flag. Health care’s high-propensity applications grew 55.8 percent from 2019 to 2025, per Census BFS, nearly matching the sector’s total growth, which suggests the health-services formation wave is disproportionately real employers rather than paper entities. Other services tells the sobering version of the same story. Its total applications rose 47.7 percent over the window. Its high-propensity applications posted 0.0 percent growth. The auto repair and personal services application boom, in other words, is mostly non-employer formation, which is one reason the actual establishment counts in Part 3 barely moved.

Part 2: The Churn Table

Applications measure intent. The Census Business Dynamics Statistics measure what actually happened: establishments that opened and establishments that closed, counted from administrative records covering the universe of US employer businesses. This is the dataset that converts the formation story into a competition story.

Sector-level births, deaths, and net entry

Census BDS 2023 vintage, sector one-way tables, establishments.

Construction’s five-year net entry totals 77,571 establishments, per Census BDS. Averaged, that is roughly 15,514 net new construction competitors per year. No other trade-relevant sector matched that pace in absolute terms except health care, which averaged about 17,509 net new establishments per year over the same window.

The full 2019 through 2023 average net entry ranking runs as follows, all figures from Census BDS. Health care added a net 17,509 establishments per year. Construction added 15,514. Professional services added 10,484. Accommodation and food added 9,396. Transportation added 8,156. Other services added 6,680. Admin and support added 6,256. Retail added just 814. Retail’s near-zero figure masks three consecutive years of net establishment losses from 2019 through 2021, which makes it the only trade-adjacent sector that shrank during the formation boom.

Long-run churn rates: the metabolic speed of each sector

Any single year of entry and exit data carries cyclical noise. Averaging BDS entry and exit rates from 2010 through 2023 filters that noise out and gives each sector a stable churn signature.

Census BDS, entry and exit rates averaged 2010 through 2023.

Two sectors run at low metabolic speed: health care and other services. Both replace less than 9 percent of their establishment base per year through births, per Census BDS. Both lose less than 8 percent through deaths. This is the statistical texture of businesses with recurring customers, licensure moats, and physical service territories. It is not a coincidence that these two sectors contain the auto repair shops, med spas, and home health agencies that private equity has spent a decade rolling up. The consolidators found the low-churn codes long before the churn data was fashionable to read.

Construction runs nearly twice as fast. It replaces 13.1 percent of its establishment base annually through entry, per Census BDS. It loses 12.0 percent through exit. High churn does not make construction businesses bad acquisitions. It means the population average is misleading, because the sector simultaneously contains ephemeral one-truck operations and thirty-year institutional contractors, and diligence has to establish which one is being bought. The four-digit data in Part 3 starts that separation.

Part 3: Trade Verticals at Four-Digit Depth

The public BDS files break out roughly 300 four-digit NAICS industries with full entry and exit series. Ten of them map onto the verticals this site covers. Everything in this section is 2023 vintage, the latest available, and every figure is from Census BDS unless labeled otherwise.

The master churn table

Census BDS 2023 vintage, 4-digit NAICS one-way tables, establishments.

The average net addition figures for 2019 through 2023 rank the verticals by how fast their competitive fields are physically growing, all from Census BDS. Residential builders added 6,697 net establishments per year. Trucking added 5,940. Landscaping and building services added 4,475. Personal care added 4,061. Building equipment contractors added 3,098. Other specialty trade added 2,221. Exterior contractors added 1,834. Building finishing added 1,545. Home health added 1,250. Auto repair added just 1,132.

Long-run rates: the structural signature of each code

Census BDS, rates averaged 2010 through 2023. *Cumulative net growth is total net entry over 2019 through 2023 measured against the establishment base at the start of the window.

What each vertical’s churn signature says

Building equipment contractors (2382), the HVAC, plumbing, and electrical code, is the structural low-churn winner among the trades. Its long-run entry rate of 9.8 percent is the lowest of any construction subsector we measured, per Census BDS. Its long-run exit rate of 8.8 percent sits more than three full points below the construction sector average. The 2023 count of 183,573 establishments makes it the largest specialty trade code. Yet its base grew only 9.2 percent cumulatively across the five boom years. Licensing requirements for electrical and plumbing work, plus the capital cost of HVAC service fleets, appear to filter entry in a way that framing or painting does not. Roughly 3,834 net new competitors entered nationally in 2023, spread across every metro in the country. For a buyer, this signature reads as the best combination in the trades: massive fragmentation with slow competitive entry.

Roofing and exterior contractors (2381) churn faster than the mechanical trades. The code’s long-run entry rate is 13.4 percent, per Census BDS. That runs about 3.6 points above building equipment contractors. Exit tracks entry closely at 12.9 percent, meaning the industry replaces roughly one in eight establishments every year while barely growing. The 2023 net add of 1,092 establishments works out to about 1.2 percent growth on the code’s base. Storm-driven demand cycles explain part of this pattern, since restoration booms pull in transient operators who exit when the insurance work dries up. A buyer underwriting a roofing platform should treat storm exposure as a churn amplifier, not just a revenue variable.

Building finishing (2383) is the treadmill code. Entry ran 12.5 percent in 2023, per Census BDS. Exit ran 12.0 percent. The net gain was just 565 establishments across the entire country. Painting, drywall, and flooring carry the lowest entry barriers in construction, and the data shows the consequence: high gross churn with almost no net accumulation. A finishing contractor with ten years of history has outlasted a majority of the establishments that existed when it started, which makes demonstrated longevity itself a diligence signal in this code.

Other specialty trade contractors (2389), which includes excavation and sitework, shows the healthiest churn balance in construction. Entry ran 14.2 percent in 2023, per Census BDS. Exit ran just 10.4 percent. That 3.8 point net entry margin is the widest among the specialty trade codes. Equipment intensity likely explains the low exit rate, since an operator with two million dollars of iron does not casually dissolve. The code’s establishment base grew 19.0 percent cumulatively from 2019 through 2023, so a sitework buyer should expect a genuinely expanding competitive field rather than a stable one.

Residential building construction (2361) is the highest-churn code in the trades. Its long-run entry rate is 18.4 percent, per Census BDS. Its long-run exit rate is 16.1 percent. Nearly one in five residential builder establishments is brand new in any given year, while roughly one in six disappears. The base still grew 23.4 percent cumulatively from 2019 through 2023, the fastest of any construction code. This is the industry structure of low fixed costs and project-based revenue: easy in, easy out, and very hard to underwrite for customer continuity.

Landscaping, janitorial, and pest services (5617) churn fast but accumulate steadily. Entry averaged 14.2 percent long-run, per Census BDS. Exit averaged 12.2 percent. The code added establishments in every single year of the 2019 to 2023 window, finishing with a cumulative gain of 13.6 percent. The split a buyer actually wants, which the public 4-digit file cannot make, is between pest control’s recurring contract revenue and mow-and-blow landscaping’s transactional revenue. Those two businesses almost certainly sit at opposite ends of this code’s internal churn distribution. We flag this as a granularity limit rather than guessing at numbers the file does not contain.

Auto repair (8111) is the single most stable vertical in this report. Its long-run entry rate of 8.8 percent is the lowest we measured across all ten codes, per Census BDS. Its long-run exit rate of 8.4 percent is also the lowest. Cumulative net establishment growth from 2019 through 2023 came to just 3.8 percent, meaning the competitive set a shop faced in 2023 was almost the same size as the one it faced in 2019. An average of only 1,132 net new competitors per year entered nationally across a base of more than 150,000 establishments. Bay capacity, lift equipment, and diagnostic tooling create real entry costs, and vehicle owners exhibit strong shop loyalty. Census BDS counts 135,684 firms operating the code’s 156,219 establishments. That pairing is close to a textbook definition of roll-up quality: enormous fragmentation, near-zero net entry pressure, and durable customer relationships.

Personal care services (8121) grew fast on both gross and net measures. Entry ran 15.3 percent in 2023, per Census BDS. Exit ran 11.4 percent, producing a net add of 5,472 establishments in a single year. The five-year cumulative gain of 16.7 percent captures the med spa and salon-suite expansion documented elsewhere on this site. High entry with high net accumulation suggests demand growth is outrunning competitive saturation for now. A buyer should still know that the code’s exit rate spiked to 16.1 percent in 2021, per Census BDS, which is evidence that personal care sheds businesses quickly when conditions tighten.

General freight trucking (4841) is the cautionary tale. Applications in the broader transportation sector doubled during COVID, and the BDS shows what happened next. Trucking establishments grew from 73,918 in 2019 to 98,940 in 2023, per Census BDS. That is a 33.9 percent expansion of the competitive field in four years. Entry rates touched 26.5 percent in 2022. Then the freight recession arrived. The 2023 exit rate hit 19.3 percent, and exits matched entries almost exactly. Net entry collapsed from a gain of 12,152 establishments in 2022 to a loss of 6 in 2023. A sector can add a third to its establishment base in three years and then destroy operators at the same pace. Churn data would have flagged this risk in advance, since 4841’s long-run entry rate of 18.5 percent marked it as the most invasion-prone code in this report even before COVID.

Home health care (6216) combines growth with stickiness. Entry averaged 12.9 percent long-run, per Census BDS. Exit averaged just 9.3 percent. That 3.6 point structural margin produced 20.5 percent cumulative base growth from 2019 through 2023. Certificate-of-need rules in some states, Medicare certification timelines, and referral-network dynamics appear to protect incumbents even as the category expands. The absolute base is small at 36,712 establishments, which means the demand wave documented in our boomer business succession wave report is being served by a comparatively small and slowly churning supply side.

Part 4: Survival Curves, or What Happens to 100 New Businesses

The BLS Business Employment Dynamics program answers the question the BDS cannot: not how many businesses die each year, but how long any given new business lives. Table 7 follows every establishment cohort opened since March 1994 and records the share still alive each subsequent March. Industry-specific versions exist for each supersector, including construction.

The survival table

Of 100 new establishments, how many survive. BLS BED Table 7, averaged across all cohorts opened 1994 through 2024, observed through March 2025.

The folklore number that “90 percent of businesses fail in the first year” has never appeared in any federal dataset, and this table is the correction. Roughly 79 of 100 new private-sector establishments are still operating after one year, per BLS BED. About 49 of 100 reach year five. About 34 of 100 reach year ten. Failure is real and steady, but it is a decade-long erosion rather than a first-year cliff, and any seller narrative or buyer model built on the folklore version starts from a false premise.

The sector spread matters more than the average. Health care and other services establishments survive at materially higher rates at every horizon. Their five-year survival sits near 55 of 100, per BLS BED. Construction and its adjacent sectors cluster between 47 and 48 of 100 at the same horizon. The ten-year gap is starker: 39.5 percent of health care establishments reach a decade, against 30.0 percent in transportation. These are the same two supersectors that showed the lowest BDS churn rates, which is the internal consistency check a computed report should pass. Slow-churn sectors are slow precisely because their individual establishments live longer.

The recency effect: new cohorts survive better

Averaging across three decades hides a trend that favors recent buyers. The most recent cohorts with complete horizons outperform the long-run averages in nearly every sector we examined.

BLS BED Table 7. Cohort years end in March; the FY2020 cohort opened in the year ended March 2020.

Construction is the standout. The construction cohort opened in the year ended March 2020 posted 56.5 percent survival at year five, per BLS BED. That mark sits more than eight points above the sector’s long-run average. The cohort opened in the year ended March 2015 reached ten years at 42.6 percent. That figure runs eleven points above the long-run ten-year average. Part of this reflects the long shadow of the 2008 housing collapse dragging down the older cohort averages, since the cohort opened in the year ended March 2006 had to survive directly into the crash. Part of it appears to be genuinely stronger recent vintages, boosted by the exceptional demand backdrop of 2020 through 2023. Either way, the practical read is the same: a trades business founded in the mid-2010s and still operating today has already outlasted roughly 57 percent of its birth cohort.

Other services shows the same pattern. Its FY2020 cohort reached 60.0 percent five-year survival, per BLS BED, the highest of any sector-cohort combination we measured. Health care is the one reversal, with recent cohorts slightly underperforming its long-run average, though its levels remain above every other sector at both horizons.

For the buyer’s mental model: of 100 new construction-sector businesses opened in 2020, about 57 were still operating in 2025. Of 100 new auto-repair-adjacent other-services businesses from the same vintage, about 60 were. One repetition for emphasis, because it protects readers from bad sourcing elsewhere: the BED data is supersector-level, so an HVAC-specific or plumbing-specific survival curve does not exist in any public federal file, and any source quoting one is interpolating.

Survival meets churn: the one-two diligence read

Combining the two datasets produces the report’s core analytical tool. Take the year-five survival rate from BED and the long-run entry rate from BDS together, and each vertical lands in a quadrant.

• Auto repair’s habitat (other services). Five-year survival runs 54.6 percent, per BLS BED. The sector’s entry rate averages 8.5 percent, per Census BDS. Incumbents live long and face few invaders. Customer lists in this quadrant deserve premium valuation, because attrition replacement is cheap and competitive poaching is structurally limited.
• HVAC, plumbing, electrical (construction supersector, low-churn 4-digit code). Sector survival runs 48.3 percent at year five, per BLS BED. Code-level entry averages 9.8 percent, per Census BDS. Slightly shorter incumbent lives than other services, but the lowest entry pressure in construction. Service-agreement books convert this signature into effective low churn.
• Landscaping and building services (admin/support). Five-year survival runs 47.2 percent, per BLS BED. Code-level entry averages 14.2 percent, per Census BDS. Shorter lives, more invaders. Contract commercial accounts and route density are the defenses, and diligence should weight recurring-contract share heavily.
• Residential building and trucking. Five-year survival sits between 46.8 and 48.3 percent depending on the supersector, per BLS BED. Long-run entry rates run between 18.4 and 18.5 percent, per Census BDS. This is the fragility quadrant. Project-based or spot-market revenue plus constant new entry means the acquired company’s history says little about its future competitive field.

Part 5: What Churn Means for Buyers

Churn as a customer-acquisition stress test

An industry’s establishment exit rate is a proxy for the annual evaporation rate of any customer list built on business relationships. Its entry rate is a proxy for how quickly new competitors can bid on the same demand. Consider a commercial landscaping company selling into property managers inside a code with a 12.7 percent exit rate, per Census BDS. That company should expect meaningful annual turnover in its own vendor-side competitive set, and also in its client-side ecosystem whenever it serves other small businesses. High-churn industries force acquirers to underwrite continuous customer acquisition as a permanent operating cost rather than treating the existing book as an annuity.

The inverse also holds. Auto repair’s long-run exit rate is 8.4 percent, per Census BDS. At that rate, the shop down the street has probably been there a decade and will probably remain, and so will the customer habits formed around it. Low churn compounds: longer-lived competitors imply longer-lived customer relationships, which imply that the goodwill line on the balance sheet reflects something real rather than something hoped for.

The roll-up quality screen

The private equity playbook for the trades rests on fragmentation, and our companion Industry Fragmentation League Table ranks the verticals on concentration. Churn is the second axis, and it is the one that separates good fragmentation from bad. A fragmented industry with high churn offers a consolidator thousands of targets that individually may not survive integration. A fragmented industry with low churn offers durable targets whose cash flows persist long enough to pay down acquisition debt. On the combined screen, the 2023 Census BDS data ranks the verticals covered here roughly as follows.

1. Auto repair (8111). The code holds 156,219 establishments, per Census BDS. Long-run entry runs 8.8 percent against 8.4 percent exit. That is the strongest churn profile in this report, full stop.
2. HVAC, plumbing, electrical (2382). The code holds 183,573 establishments, per Census BDS. Long-run entry runs 9.8 percent against 8.8 percent exit. Nearly as strong as auto repair, with better demand tailwinds from electrification and housing-stock age.
3. Home health (6216). A smaller base at 36,712 establishments, per Census BDS. The code carries a 3.6 point structural entry-over-exit margin plus regulatory stickiness, which offsets the thinner target pool.
4. Landscaping and building services (5617). A big base with moderate churn. Defensible where contract revenue dominates, and diligence should price the recurring share explicitly.
5. Roofing and exterior (2381). Storm-cycle churn requires underwriting insurance-demand exposure explicitly rather than averaging through it.
6. Personal care (8121). The growth is real, but the exit spike to 16.1 percent in 2021, per Census BDS, shows the downside case arrives fast.
7. Residential builders (2361) and trucking (4841). Churn profiles that demand significant valuation discounts against every vertical above them on this list.

Readers cross-shopping entry multiples can compare this ranking against the M&A multiples fleet published on this site, including the auto repair multiples page, the HVAC multiples page, and the landscaping multiples page. The pricing data on those pages generally confirms that the market already pays up for the low-churn codes, which is exactly what efficient buyers should expect. The edge from churn data is not discovering that auto repair is stable. It is quantifying how much stability a multiple is buying, and spotting the codes where pricing and churn have drifted apart.

The succession overlay

Churn statistics count exits without distinguishing failure from retirement, and that distinction is the entire lower-middle-market opportunity. Census BDS recorded 77,211 construction establishment deaths in 2023. A meaningful share of those deaths were owner retirements without a successor rather than business failures, a dynamic quantified in our boomer business succession wave report and extended in the companion Succession Cliff Report. For a buyer, every retirement-driven exit in a low-churn industry is a business that could have been acquired instead of dissolved. The BDS death counts in this report therefore double as a rough ceiling on annual acquirable-exit volume per vertical. Roughly 15,242 building equipment contractor establishments exited in 2023, per Census BDS. Even if only a fraction of those were retirement-driven going concerns, that fraction dwarfs the number of brokered HVAC and plumbing deals actually closing each year, which is the arithmetic case for proprietary sourcing.

The 2026 forward read

The BFS data through June 2026 suggests the competitive replenishment pool remains full. Construction applications ran 8.7 percent ahead of 2025’s first-half pace, per Census BFS. Admin and support services applications ran 18.8 percent ahead. If historical application-to-formation conversion holds, the 2024 through 2026 application cohorts will produce elevated establishment births in the BDS vintages covering those years, keeping entry rates near their post-2021 highs. Buyers underwriting five-year holds in the trades should model a competitive field that grows in every year of the hold. The code-level differences documented in Part 3 determine how much of that growth actually contests a given company’s customer base, which is why the four-digit table, not the sector headline, belongs in the model.

Download the Dataset

Every figure in this report traces back to a single companion file: churn_report_data.csv, a 2,010-row tidy CSV containing every BFS annual sum, every BDS entry and exit series, and every BED cohort survival value used above. Each row carries its dataset name, series code, NAICS code, year, value, and source URL, so any number on this page can be traced to its federal origin in one lookup. The file is free to use with attribution to CT Acquisitions and the underlying federal programs. Analysts building their own models are encouraged to start from the raw file rather than from this page’s rounded prose.

Related Research

This report is one panel of a larger data series on small business supply and demand. The companions published alongside it, and the standing research it builds on:

• The Boomer Business Succession Wave Report, 2024 to 2030: the demand-side demographic wave behind the exit counts in this report.
• The Succession Cliff Report 2026: which exits are retirements rather than failures, and where the unsold businesses go.
• The Industry Fragmentation League Table 2026: the concentration axis that pairs with this report’s churn axis.
• SBA Loan Default Rates by Industry 2026: how the survival curves above show up in lender charge-off data.
• The Skilled Trades Wage Atlas 2026: the labor-cost side of the same federal microdata stack.
• Multiples context for the roll-up ranking: Auto Repair M&A Multiples, HVAC M&A Multiples, and Landscaping M&A Multiples.

Frequently Asked Questions

What is business churn?

Churn is the combination of establishment births and deaths in an industry over a period, usually expressed as annual entry and exit rates against the existing establishment base. The Census Bureau’s Business Dynamics Statistics compute both rates from administrative records covering all US employer businesses. High churn means the competitive field turns over quickly. Low churn means incumbents persist, and customer relationships persist with them.

How many new HVAC, plumbing, and electrical businesses start each year?

Census BDS data for NAICS 2382, building equipment contractors, shows 19,076 establishment births in 2023. Deaths in the same code totaled 15,242. The net gain was 3,834 establishments. The code covers HVAC, plumbing, and electrical contractors together, and the public federal data does not split them at any finer level of detail.

What percentage of construction businesses survive five years?

Averaged across BLS cohorts opened from 1994 through 2024, 48.3 percent of new construction establishments were still operating at year five. The most recent complete cohort did better: the group opened in the year ended March 2020 posted 56.5 percent five-year survival, per BLS Business Employment Dynamics.

Is the small business formation boom over?

The data says it is not. Total US business applications reached 5,671,836 in 2025, per Census BFS, the highest annual total on record. The 2021 figure that was widely called the peak came to 5,390,816. First-half 2026 applications ran 12.1 percent ahead of first-half 2025, pointing toward another record year.

Do business applications actually become businesses?

Most do not, which is why the Census Bureau publishes a high-propensity subseries flagging applications likely to become payroll employers. In 2025, 30.1 percent of all applications carried high-propensity markers, per Census BFS. In construction the share was 51.4 percent, the strongest conversion signal among major sectors.

Which trade vertical has the lowest churn?

Automotive repair and maintenance, NAICS 8111. Its long-run entry rate of 8.8 percent is the lowest of the ten verticals we computed from Census BDS. Its long-run exit rate of 8.4 percent is also the lowest. The code’s establishment base grew just 3.8 percent cumulatively from 2019 through 2023.

Which vertical has the highest churn?

General freight trucking, NAICS 4841. Its long-run entry rate averages 18.5 percent, per Census BDS. Its 2023 exit rate hit 19.3 percent. The code swung from a net gain of 12,152 establishments in 2022 to a net loss of 6 in 2023. Residential building construction runs nearly as hot, with an 18.4 percent long-run entry rate.

Does high churn make an industry a bad place to buy a business?

Not automatically, but it changes what a buyer should pay for. In high-churn verticals, the durable asset is whatever survives turnover: contracts, route density, licenses, or equipment positions rather than generic customer lists. In low-churn verticals like auto repair, the customer base itself carries more underwritable value. Churn functions as a discount-rate input, not a veto.

What is the difference between BFS, BDS, and BED data?

BFS counts EIN applications monthly, so it is fast but measures intent. BDS counts actual establishment openings and closings from Census administrative records annually, with about a two-year lag. BED tracks how long each cohort of new establishments survives, using state unemployment-insurance records. Formation, churn, and longevity respectively, and this report uses all three.

Of 100 new businesses, how many survive ten years?

About 34 across the whole private sector, per BLS BED cohort averages. The sector range runs from about 30 of 100 in transportation and warehousing to about 40 of 100 in health care. The construction long-run figure is about 32 of 100. The construction cohort opened in the year ended March 2015 beat that badly, reaching ten years at 42.6 of 100.

Sources and Vintages

• US Census Bureau, Business Formation Statistics, monthly CSV, not seasonally adjusted, data through June 2026. Retrieved July 2026 from census.gov/econ/bfs.
• US Census Bureau, Business Dynamics Statistics, 2023 vintage, sector and 4-digit NAICS one-way tables. Retrieved July 2026 from www2.census.gov/programs-surveys/bds/tables/time-series/2023/.
• US Bureau of Labor Statistics, Business Employment Dynamics, Table 7, establishment survival by opening year, national tables for total private, construction, transportation and warehousing, professional and technical services, administrative and waste services, health care and social assistance, and other services, observations through March 2025. Retrieved July 2026 via Internet Archive captures of the bls.gov/bdm source files; cited URLs resolve to the live BLS files in a browser.

Disclaimer and Build Notes

Disclaimer: This report is published for informational purposes only and does not constitute investment, legal, tax, or accounting advice. Figures reflect federal data vintages available as of July 2026 and industry conditions can change materially between annual revisions. Entry, exit, and survival statistics describe populations of establishments, not any individual business, and past cohort performance does not predict the outcome of any specific acquisition. Buyers should verify all figures against the cited primary sources and engage qualified advisors before transacting.

Build notes: all figures were computed in pandas from raw downloaded federal files. No numbers were estimated, modeled, or interpolated, and every granularity limit in the source data is disclosed inline. The finished text was screened before publication and recorded zero hits against the CT voice-gate exclusion set. This page is refreshed annually when the next BFS, BDS, and BED vintages publish.