Business Services
Valuation Multiples

How Business Services Companies Are Valued: SDE vs. Adjusted EBITDA

Before any multiple in this article means anything, you need to know which earnings figure it sits on top of. Business services straddles the same divide most owner-operated sectors do. A single-location shop under roughly $500,000 of earnings, where the owner still runs the operation day to day, is usually quoted on seller's discretionary earnings. Think of a two-person creative agency, a single-route pest control operator, or a small staffing desk placing candidates the owner sourced personally. SDE adds the owner's full compensation back to profit, on the logic that a buyer stepping into the owner's chair captures that salary along with the business.

Once a company runs on a management layer, the convention shifts to adjusted EBITDA, which adds back only the portion of owner compensation above a market rate for the job actually being replaced. The gap is not a technicality. Pay yourself $200,000 to run a business a general manager could run for $110,000, and an SDE presentation adds back the full $200,000 while an EBITDA presentation adds back $90,000. Same business, same honesty, two different earnings bases. Every multiple in this article is quoted against EBITDA unless we say otherwise. Our guide to EBITDA and business valuation basics walks through how adjusted earnings get built and which add-backs survive a buyer's review.

The Number Everyone Quotes, and What GF Data Actually Publishes

Type “business services EBITDA multiple” into a search bar and you will get 7.4x, attributed to GF Data. We went looking for the primary source. We could not find one, and that turned out to be the most useful thing we learned writing this guide.

Here is what GF Data publishes in public. The Association for Corporate Growth, which owns GF Data, reported in its Q4 2025 release that “average purchase price multiples held steady year-over-year at 7.2x trailing 12-month adjusted EBITDA,” drawn from “297 completed transactions for the full year” reported by contributing private equity firms, covering “private equity-sponsored M&A transactions in the $10 million–$500 million enterprise value range.” GF Data's own Q3 2025 report puts the figure at 7.5x for that quarter against 6.9x in the second. Neither breaks out a single industry. ACG notes that its subscriber reports contain drilldowns by deal size, buyer type and industry. Those cuts are not public.

Notice the trap in that paragraph before going further. 6.9x, 7.2x and 7.5x are all genuine GF Data figures for 2025, and they differ because they describe different periods. Anyone quoting “GF Data's multiple” without naming a period is not being precise enough to argue with.

So where does 7.4x for business services come from? CapitalPad's compilation is the most transparent version we found, and it names its underlying documents: a GF Data ESOP Advisor Special Report from Q3 2025 and a CIBC US Middle Market Monitor from Q1 2026 sourced to GF Data, covering $10 million to $250 million of enterprise value, with 1,333 cumulative observations from 2003 to 2025. On that compilation, business services runs 7.2x to 7.4x from 2021 through 2024, 7.5x through the first three quarters of 2025, and 7.4x for the full year, holding “within a range of a few tenths year after year.” We could not obtain either underlying document. The figure may well be accurate. It is not publicly verifiable, and we are not going to present it as though it were.

Praxis Rock repeats the same number with an addition: “GF Data reports business services multiples hit 7.4x in 2025, tying the highest level in their database history.” We could not confirm that record claim from anything GF Data has published publicly. An earlier version of this article repeated it. We have taken it out.

Meanwhile, one 7.4x in this space is properly sourced, and it describes something else entirely. ACG's own Middle Market Growth reported GF Data analysis of 360 transactions completed since the third quarter of 2024 finding that “sellers that used a sell-side QoE saw TEV/EBITDA multiples of 7.4x on average, compared with 7.0x for those that didn't undertake a QoE process.” All industries, not business services, and it measures how prepared a seller was rather than what sector they were in. We cannot prove the advisory-site number began life here. We can say the sourced 7.4x and the unsourced 7.4x are the same digits describing different populations, and that only one of them has a document behind it.

The sub-sector ladders disagree with each other too, which is easier to show than the composite because both are public. Praxis Rock puts pest control at 5x to 10x and staffing at 4x to 8x. FISART, which cites GF Data, IBBA Market Pulse, BizBuySell and BVR's DealStats Value Index alongside its own advisory experience, puts pest control at 5.0x to 8.5x with a 6.5x median and staffing at 3.5x to 7.0x with a 5.0x median, and adds landscaping at 3.5x to 6.0x (4.5x median) and IT managed services at 5.0x to 10.0x (7.0x median). Neither discloses a sample size. They differ by a turn and a half at the top of pest control and a full turn at the top of staffing. Both cannot be right.

Draw the conclusion out, because it is the argument of this article arrived at from the data side rather than asserted. A category whose headline multiple traces to no public primary source, whose published sub-sector ladders differ by more than a turn, and whose own overall figure moves half a turn between quarters is not behaving like a market with a price. Open the drawer and you find pest control routes, staffing desks, janitorial contracts, landscaping crews, fire-alarm inspection books, elevator portfolios, creative agencies, security guard companies, testing labs and equipment rental yards. There is no reason a single number should describe all of that, and the evidence says none does.

One thing GF Data's public reporting does establish, and it matters more than the sector question. CapitalPad's compilation of the size bands puts $10 million to $25 million enterprise value deals in a 6.1x to 6.4x range across 2021 to 2025 and $100 million to $250 million deals in an 8.5x to 10.3x range, with the long-run spread between platform buyouts above and below $100 million averaging 2.6 turns. Before vertical enters the conversation at all, deal size moves the number by more than two and a half turns.

Who This Article Is For

You own a business services company, broadly defined, and you want to know where your specific business sits before a category average misleads you. Business services is one of five sectors Salt Creek covers. Our sector expertise page has the others, and this article goes deeper on valuation than our guide to M&A advisors for business services, which covers the buyer landscape and the process itself. Most owners we work with run founder-owned or family-owned companies with roughly $2 million to $75 million in revenue and at least $500,000 of adjusted EBITDA. Below that range the mechanism below still applies, but your likely buyer is an individual operator or a small regional group rather than a private equity platform, and the top of every range here is not your market.

Contracted Revenue Is Not Recurring Revenue

This is the whole argument, so it gets stated on its own before the taxonomy.

Owners, brokers, and a fair number of buyers use “contracted” and “recurring” interchangeably. 37th and Moss, in a piece written specifically about the confusion, defines recurring revenue as “income that is expected to continue in the future on a repeatable schedule” that is “predictable, stable, and contractual,” and contracted revenue as income arising from an agreement where “the agreement may not guarantee utilization of a good or service, and subsequently may not guarantee predictable revenue.” The line the piece draws: “While recurring revenue is always contracted… not all contracted revenue is recurring.” It adds that master service agreements “do not inherently guarantee revenue as they often lack requirements for utilization.”

That is a consultancy's framing. The stronger evidence is that public companies in this sector describe themselves in exactly these terms in filings they are legally obligated to get right.

Take the recurring end first. Rollins, the largest pure-play pest control operator in North America, states in its 2025 Form 10-K that “as of December 31, 2025, approximately 75% of our business was recurring services, 10% was ancillary services, and 15% was one-time services,” and describes recurring services as “ongoing pest prevention treatment under a scheduled service agreement” where “relationships often extend over multi-year periods.” Nothing in that sentence requires a salesperson to do anything for the revenue to show up next quarter.

Now the contracted end. Robert Half, a $5.4 billion staffing and consulting firm, writes in its 2025 Form 10-K that “the Company's clients will frequently enter nonexclusive arrangements with several firms, which the client is generally able to terminate on short notice and without penalty.” That is not a critic's characterization of staffing. It is the largest specialty staffing firm in the United States telling its own shareholders what its contracts are worth. In the same filing, U.S. revenues fell 7.7% to $4.17 billion and full-year net income fell 47.1%.

The most useful disclosure of all may be EMCOR's. In its 2025 Form 10-K, the mechanical and electrical contractor warns that “amounts included in our remaining performance obligations may not result in actual revenues” because “many contracts are subject to cancellation or suspension on short notice at the discretion of the client.” Remaining performance obligations are, definitionally, signed contracts. A company with a multibillion-dollar contracted backlog is telling investors that contracted does not mean collected. Comfort Systems says the same thing about its own numbers in its 2025 Form 10-K: “the predictive value of backlog information is limited to indications of general revenue direction over the near term.”

The Four Questions That Settle It

The test is mechanical, and your own agreements answer it in an afternoon.

• Does it auto-renew? Or does someone at the customer have to affirmatively sign again? ABM describes its agreements as “typically re-bid upon renewal.” Comfort Systems describes its commercial service agreements as having “terms of one or more years, with automatic annual renewals, and frequently include 30- to 60-day cancellation notice periods.” Read that second one twice. It auto-renews and the customer can walk on 30 days' notice, which is the cleanest single illustration in this article that auto-renewal alone does not make revenue durable. Auto-renewal answers who has to act. It does not answer what exit costs.
• Is there a volume or spend floor? Or can usage go to zero without anyone breaching anything? This is the question that separates a route agreement from an MSA, and it is the one owners most often assume in their own favor.
• What does exit cost the customer? Nothing, a notice period, or an affirmative cancellation. ABM discloses that it “primarily provide[s] services pursuant to agreements that are cancelable by either party upon 30 to 90 days' notice.” Comfort Systems' service agreements “frequently include 30- to 60-day cancellation notice periods.”
• Who has to act for the revenue to continue? If the answer is nobody, it is recurring. If the answer is your salesperson, it is a pipeline with a contract stapled to it.

A contract that auto-renews, carries a volume floor, and requires an affirmative cancellation is recurring in the sense buyers pay up for. A contract that must be actively renewed, carries no floor, and can simply lapse is contracted in name only. Both look identical on a summary page in a confidential information memorandum, which is precisely why diligence teams stop reading the summary and start reading the contracts. The gap between the two is one of the more common reasons a business services deal gets repriced after the letter of intent.

What you can do about it is bounded but real. You cannot retroactively make a staffing MSA behave like a route agreement. Over a multi-year planning window you can shift mix toward structures that behave more like the recurring end: RPO and MSP-style embedded workforce arrangements instead of transactional temp placement, retainer or subscription pricing in an agency or facilities business instead of pure project billing, multi-year agreements with volume commitments instead of annual renewals with none. None of that happens quickly, which is why it belongs in the window described in our guide on when to start exit planning rather than in the twelve months before a sale.

A Taxonomy of the Filing Cabinet

Here is the category broken into the drawers that actually price differently, with the evidence we could verify for each and an honest note where we could not.

Route-Based Services: Pest Control, Hygiene, Waste

Route businesses sit at the top of the range because the contract does the selling. FISART puts pest control at 5.0x to 8.5x with a 6.5x median. Broker-published figures from CT Acquisitions show a steeper ladder by scale: 7x to 10x for mid-market tuck-ins at $500,000 to $2 million of EBITDA, 9x to 12x for multi-market regional operators at $2 million to $10 million, and 11x to 15x at platform level above $10 million, “with premier acquisitions reaching 17x.” We label those broker-published deliberately. They come from an advisory firm's marketing content rather than a transaction database, and should be read as the shape of the market, not a quote.

The same source puts a threshold on it: “70%+ recurring is the threshold where buyers materially shift their underwriting.” Below the line, an operator gets priced like a general home services business. Above it, the contract book itself becomes the asset. CT Acquisitions also quantifies the split directly, valuing recurring revenue at 8x to 10x EBITDA against 4x to 5x for one-time job revenue, and puts pest control at 7x to 10x against 4x to 7x for HVAC and 4x to 6x for plumbing. Both of those are relationship-driven, locally dominant trades. Both sell projects rather than subscriptions.

Rollins is the proof at scale. Its 75/10/15 recurring, ancillary, and one-time split is the composition a buyer is trying to find in a private operator. Worth noting honestly: Rollins does not publish a numeric customer retention rate in its 10-K. It discusses retention qualitatively and describes a route management system that increases “customer retention through on-time and rapid response service.” If the largest operator in the sector does not disclose a churn number, you should be skeptical of any published sector-wide retention statistic, including ones you may have seen quoted with a decimal point.

Commercial hygiene and uniform rental run the same playbook. Cintas discloses that “over 95% of the Company's revenue is derived from fees for route servicing,” delivered across “approximately 12,500 local delivery routes, 484 operational facilities and 12 distribution centers” using roughly 24,500 vehicles. Cintas even defines its own compensation peer group as companies with “route-based delivery of products and services” rather than as business services companies generally, which is a fair summary of this article's thesis coming from inside the category.

Facilities Services: Janitorial, Landscaping, HVAC and Mechanical

Facilities is where owners most often assume they are in the pest control drawer and find out otherwise. ABM Industries, at $8.75 billion of fiscal 2025 revenue the largest public pure-play facilities services company in the U.S., discloses that it “primarily provide[s] services pursuant to agreements that are cancelable by either party upon 30 to 90 days' notice,” and that those agreements are “typically re-bid upon renewal” and “obtained through a competitive bid process as well as pursuant to work orders.” ABM also states that it has “historically had a high rate of client retention” despite the cancellation terms, which is the honest version of the story: the relationships are sticky, the contracts are not binding, and a buyer prices the contracts.

FISART puts landscaping at 3.5x to 6.0x with a 4.5x median, the lowest named vertical on its business services table. We did not find a comparable published multiple range for commercial janitorial specifically, and we are not going to derive one from the landscaping figure and call it data.

Mechanical and HVAC service is the exception inside facilities, and the reason is contract structure again. Comfort Systems describes commercial, industrial, and institutional service agreements as having “terms of one or more years, with automatic annual renewals, and frequently include 30- to 60-day cancellation notice periods.” The renewal mechanism is closer to a route agreement than to a janitorial re-bid. The cancellation window is not, and an owner claiming route-like durability on the strength of auto-renewal language should expect a buyer to read the next clause. An HVAC business whose revenue is mostly annually renewing maintenance agreements is a different asset from one whose revenue is mostly replacement projects, even at identical EBITDA. That is the same distinction CT Acquisitions is pricing when it puts HVAC at 4x to 7x against pest control's 7x to 10x.

Compliance-Driven Services: Fire and Life Safety, Elevator, Environmental

This drawer has a feature none of the others do: a third party requires the customer to buy. APi Group describes itself in its 2025 Form 10-K as providing “statutorily mandated and other contracted services” and explains the mechanism plainly: building codes “require testing, inspections, repair, maintenance and specific retrofits of building fire suppression and sprinkler systems, which generates recurring revenue related to those services,” and “inspections are often required by legislation or insurance mandates, providing a strong recurring revenue stream.” The company describes its strategy as “selling inspection work first,” on the view that inspection creates “a stickier customer relationship that leads to recurring revenue, higher margins, and growth opportunities.”

Read that as an underwriting model, not a marketing line. When a fire marshal, an elevator inspector, or an environmental permit compels the service, churn is structurally capped by something other than customer satisfaction. APi also notes that its inspection-led model requires capital expenditures “typically less than 1.5% of total net revenues,” which matters for free cash flow conversion and therefore for what a leveraged buyer can pay.

We did not find credible private-transaction multiple data for lower middle market fire and life safety, elevator service, or environmental compliance businesses. Public comparables in this space trade well, but public multiples on diversified multibillion-dollar companies are not a proxy for a $3 million EBITDA regional inspection business, and we are not going to present one as if it were.

Staffing and Recruiting

Staffing sits at the bottom of the range, and it is not because staffing firms are badly run. FISART puts the sector at 3.5x to 7.0x with a 5.0x median. Broker-published bands from Auxo Capital Advisors break it out by model: light industrial staffing at 4.5x to 7.0x, direct hire and executive recruiting at 4.0x to 7.5x, professional and administrative at 5.0x to 8.0x, IT and technical at 6.0x to 9.5x, RPO and MSP-enabled workforce solutions at 6.5x to 10.0x or better, and multi-service staffing platforms at 7.0x to 10.0x or better. The pattern inside those bands is the same one running through this whole article. The two highest bands are the two models with embedded, multi-year, workflow-integrated contracts. The lowest band is transactional placement.

The cyclicality is not theoretical either. Bureau of Labor Statistics Current Employment Statistics data, which we pulled directly from the agency's public API rather than from a secondary write-up, shows payroll employment in employment services (NAICS 5613, the staffing and temp help sector) peaking at 3,928,100 in March 2022, bottoming at 3,137,700 in October 2025, and standing at 3,207,500 in June 2026. That is a 20.1% peak-to-trough decline in the sector's own headcount over three and a half years. Over the same window, employment in services to buildings and dwellings (NAICS 5617, which covers janitorial, landscaping, and pest control) set an all-time high of 2,316,900 in March 2026. One category. Two sub-sectors moving in opposite directions by roughly twenty points of employment.

Kforce's 2025 Form 10-K shows what that looks like in a single company's P&L: Flex revenue down 5.3% to $1.30 billion, direct hire revenue down 11.1% to $25.7 million, and gross margin down to 27.2% from 27.4%, with the decline attributed in part to a shift in the mix away from direct hire. Note the second-order effect there. Direct hire revenue carries near-100% gross margin, so a mix shift out of it compresses reported margin even when the underlying business is stable, which is exactly the kind of thing a buyer's quality of earnings analysis is built to isolate.

Scale matters too, and it is worth knowing what you are competing against. The American Staffing Association reports that roughly 2.2 million temporary and contract employees worked for U.S. staffing companies in an average week in 2024, that staffing firms hired 12.7 million temporary and contract employees over the course of 2023, and that there were about 27,000 staffing and recruiting companies operating close to 54,000 offices. A fragmented market with 27,000 competitors and no contractual switching costs is a structurally hard place to defend a premium multiple.

Marketing, Creative, and Professional Services

Agency retainers look recurring on a summary page. In practice they are frequently cancelable on 30 to 60 days' notice and tied to specific creative or account personnel, which puts them closer to a staffing relationship than to a route agreement despite the label. The same applies to most consulting practices, where the revenue is a series of engagements and the asset walking out the door at 6pm is the asset.

Korn Ferry's fiscal 2026 Form 10-K is instructive on how a professional services business tries to solve this, describing a model that serves clients “through both project-based and recurring engagements” and “balances project-based work with recurring revenue streams.” A firm at Korn Ferry's scale invests heavily in converting project revenue into subscription and program revenue for exactly the reason this article keeps returning to.

We did not locate reliable private transaction multiple data specific to lower middle market marketing agencies or consulting practices, and are not going to publish a range we cannot support. What we can tell you with confidence is the underwriting question: how much of the revenue survives if the founder stops taking client calls, and can you prove it with something other than an assertion?

IT Services and Managed Services

MSPs are the closest thing inside the broader services universe to a pest control route, and they get priced accordingly. FISART puts IT managed services at 5.0x to 10.0x with a 7.0x median, the highest median on its business services table and a full two and a half turns above staffing's. The mechanism is identical to the one described throughout this article: multi-year contracts with defined minimums and auto-renewal terms. Our MSP valuation multiples guide covers that premium and what erodes it.

The interesting edge case is IT staffing, which is often mistaken for managed services because both sell technical labor. Auxo puts IT and technical staffing at 6.0x to 9.5x, ahead of light industrial but behind a true MSP contract book. The gap is the contract, not the skill level of the worker.

Equipment Rental and Testing, Inspection, and Certification

Equipment rental is the drawer where the EBITDA multiple is the wrong lens on its own. United Rentals' 2025 Form 10-K describes a fleet carrying $22.5 billion of original equipment cost and reports fleet productivity, a metric combining rental rates, time utilization, and mix, up 2.2% for the year, with 69% of equipment rental revenue coming from key accounts. A rental business converts EBITDA into free cash flow at a very different rate than a route business does, because the fleet has to be replaced. Two companies at the same EBITDA and the same headline multiple can deliver wildly different cash to an owner after maintenance capex, which is why buyers in this vertical underwrite fleet age and utilization before they underwrite the multiple.

For testing, inspection, and certification businesses, public comparables trade well and private data is genuinely thin. We could not find credible private-transaction multiple data for lower middle market TIC companies. If you run one, the honest starting point is that published data for your situation does not really exist, and a conversation about your actual contracts and accreditation scope will tell you more than any range we could print.

Vertical Comparison at a Glance

Two kinds of figures sit in this table and they are not equivalent. The FISART column is transaction-guided data covering a broad private sample. The broker-published column comes from advisory firms' own marketing content. Where a cell is blank, we did not find data we were willing to publish.

Read down the last column. Every ceiling in the table is set by the same question asked in different vocabulary: does the revenue renew on its own? Pest control's ceiling is a recurring share threshold because the contract renews itself. Compliance services' ceiling is a statutory mandate because the government renews it for you. Staffing's ceiling is concentration and contract structure because nothing renews it at all.

Net Revenue Retention Is the Underwriting Metric

Software buyers have used net revenue retention for a decade. Business services buyers have caught up, and the owners have not.

Two numbers matter and they answer different questions. Logo churn counts customers lost as a percentage of customers at the start of the period. Net revenue retention measures revenue from last year's customer base this year, including expansion and price increases, divided by revenue from that same base last year. A route business can lose 9% of its logos and still post 104% net revenue retention if the remaining customers took a price increase and bought an ancillary service. That business is worth meaningfully more than one at 100% logo retention and 96% net revenue retention, where the accounts are stable but the pricing is eroding.

Here is the part owners underestimate. Most lower middle market service businesses cannot produce either number. The field service software holds current customers, not a clean history of who left and when. Billing lives in QuickBooks with no customer cohort tagging. Price increases were applied inconsistently by branch. When a buyer asks for cohort retention and the answer is a spreadsheet assembled the week before, the buyer does not simply accept it.

What a buyer does instead is reconstruct it, and the reconstruction is always less flattering than your recollection. The standard approach is to pull a full customer-level revenue extract for the last three fiscal years, match customers by account ID rather than name (names change; accounts do not), and calculate what percentage of year-one revenue reappeared in year two and year three. Accounts that cannot be matched get treated as churn. Then they call twenty to thirty of your customers during confirmatory diligence and ask two questions: are you under contract, and would you stay under new ownership. That is what actually sets the retention assumption in the model.

The practical instruction is unglamorous. Two years before you sell, make sure every invoice carries a stable customer ID, tag each account as recurring or one-time at the point of sale rather than in hindsight, and keep a monthly record of accounts added and lost. That data hygiene is worth more at close than most of the operational improvements owners spend the same period on, because it converts a discount for uncertainty into a number the buyer can underwrite. It is the same principle behind a clean sell-side quality of earnings report.

Contract Assignability: The Clause That Reprices Deals

You have 340 customer agreements. You have been telling buyers, accurately, that 78% of revenue is under contract. Then the buyer's counsel reads them and finds that 60% contain an anti-assignment or change-of-control provision requiring the customer's written consent before the agreement transfers. Your contracted revenue just became a consent campaign.

This is the single most common reason a services deal reprices between the letter of intent and closing, in our experience, and the mechanics are worth understanding before a buyer explains them to you. As Aird & Berlis sets out, a general anti-assignment clause restricts transfer of the contract itself and is typically triggered by an asset sale but not by a share sale, because in a share sale the contracting entity does not change. A separately drafted change-of-control clause is the one that bites either way: language deeming “any change of control… resulting from an amalgamation, corporate reorganization, arrangement, business sale or asset [sale]… an assignment or transfer” reaches the economic transfer even when the legal entity stays the same. Closing without required consent can constitute default, with consequences up to termination of the contract.

Three practical consequences follow.

First, deal structure stops being a tax question alone. If most of your material agreements carry plain anti-assignment language with no change-of-control trigger, a stock sale walks around the problem. If they carry change-of-control language, structure does not save you and consents have to be gathered. Buyers know which is which before they sign the letter of intent; sellers frequently do not.

Second, consent campaigns leak. Asking a customer for written consent to an ownership change is, functionally, telling them you are selling. Some will use it to renegotiate price. A few will use it to run a bid. That is why buyers push for consents to be a closing condition and sellers push to make them a covenant with a materiality threshold, and why the negotiation over which contracts count as “material” is worth real money.

Third, this is fixable in advance and almost nobody does it. When contracts come up for renewal in the two years before a sale, adding permitted-assignment language for a transfer to an affiliate or successor costs nothing at signature and is nearly impossible to add once a process is live.

Customer Concentration, and Why Your Lender Is Stricter Than Your Buyer

Concentration gets discussed as a single risk. It is two, priced by two different parties on two different schedules.

The buyer's version is a valuation discount. Highland Global, a business valuation firm, publishes an illustrative ladder: 25% to 35% concentration draws up to a 10% discount, 36% to 50% draws 15% to 20%, and above 50% draws a 25% to 35% discount or a change in deal structure. We are flagging plainly that Highland Global does not disclose the empirical basis for those figures, so read them as a practitioner's rule of thumb rather than a regression result. Auxo Capital describes the same effect in mechanism terms without numbers, noting that higher concentration “usually reduces buyer confidence” and produces “a lower multiple, more earnout exposure, larger escrow, or additional diligence conditions.”

The lender's version is a hard line, and it arrives earlier. A buyer can decide that a 30% customer is acceptable at the right price. A credit committee funding the acquisition often cannot, because the loan has to survive that customer leaving, not merely be priced for it. In practice that shows up as reduced advance rates against receivables from the concentrated account, a concentration covenant capping the top customer at a stated share of trailing revenue, or a lower total leverage multiple that caps what any buyer can bid regardless of enthusiasm. The sequence matters to you: a buyer who wants your business at 7x and can only finance 5.5x of it will restructure the offer with a larger seller note or a bigger earnout, not withdraw. Your headline multiple survives. Your cash at close does not.

This is also why concentration hurts more in staffing and facilities than in route businesses at identical percentages. A 30% customer under an auto-renewing agreement with a volume floor is a different credit than a 30% customer on an MSA with no minimum and a 30-day out. The percentage is the same. The collateral is not.

Density: Why 400 Stops in One Metro Beats 400 Stops in Four States

Two pest control companies each service 400 commercial accounts and each generate $1.4 million of EBITDA. One operates entirely inside a single metro. The other is spread across four states from a legacy of opportunistic acquisitions. They do not sell for the same multiple, and the reason is arithmetic rather than preference.

Route density drives the variable cost line directly. Rollins lists “route density to manage variable costs” among the specific competitive advantages its scale confers, and describes a route management system that reduces “miles driven and associated costs while increasing customer retention through on-time and rapid response service.” A technician with 12 stops within a 15-mile radius bills more hours and drives fewer miles than one with 8 stops across 60 miles. That difference flows straight to gross margin, and it compounds: the denser operator can promise same-day response, which wins commercial accounts, which increases density further.

Republic Services describes the same economics in waste, stating that its goal is “to develop the best vertically integrated market position to enable us to build density and improve returns,” and disclosing in its 2025 Form 10-K that approximately 67% of collected solid waste volume went to landfills it owns or operates. That internalization rate is density expressed as a percentage: revenue that stays inside the network rather than leaking to a third party.

For a buyer, density has a second meaning that matters more than margin. A dense operator is a platform, because the acquirer can fold a tuck-in's stops into existing technician territories at close to zero incremental route cost. A scattered operator is a collection of small businesses that happen to share a logo, and each geography has to be underwritten on its own. Same EBITDA, different strategic value, and the difference shows up in the multiple rather than in a footnote.

Labor: Pass-Through, Fixed-Price, and What a Wage Move Actually Does

ABM discloses that direct labor costs represented 68% of total revenue in fiscal 2025. Hold that number while you read the rest of this section, because it explains why contract type is a valuation input in labor-intensive services and merely an accounting detail elsewhere.

ABM describes two contract structures with opposite risk profiles. Cost-plus arrangements are ones in which “clients reimburse us for the agreed-upon amount of wages and benefits, payroll taxes, insurance charges, and other expenses associated with the contracted work, plus a profit margin.” Monthly fixed-price arrangements are ones in which “the client agrees to pay a fixed fee every month over a specified contract term.” Under cost-plus, a wage increase passes through and the margin percentage holds. Under fixed-price, the entire increase lands on the operator until the contract is re-bid, and ABM flags that risk explicitly, noting that when minimum wage rates rise it “may have to increase the wages of both minimum wage employees and employees whose wages are above the minimum wage.” That second clause is the one owners forget. Wage compression forces raises well above the mandated floor.

Run the arithmetic on a small janitorial business. Revenue $6,000,000. Direct labor at 68% is $4,080,000. EBITDA at 9% is $540,000. A state minimum wage move that raises the effective wage bill by 6%, including the compression effect on supervisors, adds $244,800 of cost. On fixed-price contracts with a year left to run, EBITDA falls to $295,200, a 45% decline, and the multiple applies to the new number. On cost-plus contracts, essentially nothing happens. Same business, same wage law, two very different valuations, and the only difference is which paragraph got signed three years ago.

A buyer will therefore ask what share of your revenue sits on pass-through terms, how long the fixed-price contracts have to run, whether your last two price increases stuck, and how much of your recent margin improvement came from holding wages below market rather than from operating leverage. The last question is the uncomfortable one. Margin earned by underpaying in a tight labor market is margin a buyer will not pay for, because they will have to give it back to keep the crews.

Working Capital: DSO and the Unbilled Problem

Services businesses look asset-light until you read the balance sheet. ABM's fiscal 2025 disclosure shows $1,223.0 million of billed trade receivables against $8,745.9 million of revenue, roughly 51 days of sales outstanding on commercial receivables alone. Add $273.6 million of unbilled trade receivables and $193.7 million of costs incurred in excess of amounts billed, and the company is carrying about $1.69 billion of contract assets, close to 70 days of revenue.

The unbilled line is the one that causes trouble in a deal. ABM's unbilled trade receivables more than doubled year over year, from $124.0 million to $273.6 million. In a private company the equivalent is work performed but not yet invoiced because the customer requires a purchase order, or the job is not closed out, or the billing clerk is behind. It is real value, it is not cash, and it is the single most contested item in a working capital true-up.

Why this costs money: your buyer sets a working capital peg, typically a trailing twelve-month average of the net working capital the business needs to operate. Deliver less than the peg at closing and the purchase price is reduced dollar for dollar. Commercial services businesses commonly get caught here because DSO drifts upward during a sale process, when the owner is in diligence meetings instead of on collections calls, and because unbilled WIP is easy to understate in a normalized peg calculation and expensive to correct after the fact. Our guide to the working capital peg in an M&A deal covers how the peg gets set and where sellers lose money on it.

Platform or Add-On: The Same Business, Two Different Prices

Your business is worth two different numbers on the same day depending on which side of a buyer's model it lands on, and knowing which one you are is worth more than any negotiation tactic.

Private equity buys most companies as add-ons now. PitchBook data reported by Cherry Bekaert puts add-on acquisitions at 72.9% of all U.S. buyouts in 2025, holding steady with the five-year average. We could not fetch PitchBook's report directly, so that figure comes to us through a secondary source that names PitchBook as its origin, and we are labeling it as such.

The pricing arithmetic follows from the structure. CT Acquisitions' pest control ladder shows tuck-ins at $500,000 to $2 million of EBITDA trading at 7x to 10x while platforms above $10 million trade at 11x to 15x. A sponsor that owns a platform valued at 12x can buy a tuck-in at 8x and, if the tuck-in's EBITDA survives integration, create four turns of value on the acquired earnings before doing anything operational. That spread is why add-on buyers can pay more than a first-time buyer and still like the deal, and also why they will not pay platform pricing for a business that cannot function as a platform.

What makes you a platform rather than an add-on is specific and largely knowable in advance: a management team that runs the business without the owner, financial reporting a sponsor can consolidate without rebuilding, a geography or vertical with enough remaining targets to justify a thesis, and a systems backbone that a future acquisition can be loaded onto. Fail those and you are an add-on, priced against what you add to someone else's platform. Our comparison of strategic buyers versus private equity buyers covers how the two buyer types diverge on price and structure, and our guide to why roll-ups are heating up covers the sponsor-side logic in more detail.

One caution on comparing offers. A platform bid and an add-on bid are often quoted on different consideration mixes, and a headline multiple that includes an earnout is not comparable to an all-cash one. Auxo Capital says it directly to staffing sellers: “a higher multiple with heavy contingent consideration can be less attractive than a lower multiple with cleaner cash at close.” Ask every bidder for cash at close as a percentage of enterprise value before you rank the offers.

What Else Moves the Number

Management depth and owner dependence. A business services company where the owner personally holds the key client relationships, prices every large job, or is the only person who can run scheduling transfers with risk attached, regardless of vertical. Buyers respond with earnouts, longer transition requirements, and rollover expectations rather than a lower headline number. Our guide on how a sell-side process compares to working with a business broker covers how that dependence gets structured around.

Documentation quality. GF Data reporting published by Middle Market Growth found, across 360 transactions completed since the third quarter of 2024, that “sellers that used a sell-side QoE saw TEV/EBITDA multiples of 7.4x on average, compared with 7.0x for those that didn't undertake a QoE process.” Read the caveat in the same breath as the number, because it decides whether it applies to you: the same reporting states that “the benefits were most notable for deals with enterprise values above $50 million, while smaller deals tended not to experience a valuation boost.” Most Salt Creek clients sit below that line, so treat the 0.4x as evidence that buyers pay for resolved uncertainty, not as a premium you should expect to collect. This is also, as the opening section sets out, the only properly sourced 7.4x in this sector's data. The unsourced one is the sector composite.

Time. Nothing in this article moves a multiple as reliably as starting early enough to fix the contract book, the customer data, and the assignment language before a buyer reads them. Our guide on how long it takes to sell a business covers realistic timelines, and the lower middle market M&A glossary defines the terms a buyer will use on you.

How Business Services Compares Across the Lower Middle Market

• Against the broader market. Skip the sector composite, since it has no public source, and compare on size instead. CapitalPad's compilation of GF Data's size bands runs 6.1x to 6.4x at $10 million to $25 million of enterprise value and 8.5x to 10.3x at $100 million to $250 million. Deal size does more verifiable work than sector membership.
• Against childcare. Childcare's broker-published ladder runs from roughly 2x at single-site owner-operated centers to 6x to 9x at institutional platform scale. The ceiling is lower than route services' for a structural reason: no childcare center operates on an auto-renewing agreement, so enrollment has to be re-won every year.
• Against recurring-revenue technology services. FISART's 7.0x median for IT managed services against 5.0x for staffing is the cleanest single illustration of this article's argument inside one dataset, and both are labor businesses selling technical skill.
• Against manufacturing. Our manufacturing valuation multiples guide covers a sector where the assets are on the balance sheet rather than in the contract file, which changes both the multiple and what a lender will advance against it.

Getting a Range Specific to Your Business

The ranges above tell you roughly where business services companies trade. They cannot tell you where yours would land, because that depends on what your agreements say about renewal and minimum volume, whether they carry change-of-control language, how concentrated your top five clients are, how much of your margin sits on pass-through terms, and how much of the business runs without you.

A preliminary valuation conversation is how a category range becomes an estimate that reflects an actual business. Jack and Connor handle these directly. We work on a success-fee basis with no retainer, so an early conversation carries no cost and no obligation to sell.

If you want to understand your options, start with our guide to M&A advisors for business services, then review how Salt Creek approaches business valuation and how a lower middle market sale process actually unfolds.