A buyer I watched close last year had priced out an excavation company at $4M asking. The wire went out at $3.2M, on about $1.4M in SDE. That's 2.3x. The DSCR at that price came in at 2.4, more cushion than the buyer strictly needed but exactly the kind of number that makes a lender comfortable saying yes.
I pull deal data out of the searcher communities I'm in pretty regularly (partly because I'm nosy about what people actually pay versus what the valuation guides say they should pay). Over the last year or so I've tracked roughly 30 closed or LOI'd deals with enough detail in the announcement to compute a real multiple and, in a lot of cases, a real DSCR. It's a biased sample: these are announced wins in a community that coaches people through the process, so failed deals and overpays are underrepresented. But it's still more useful than "service businesses trade at 2 to 4x," which is the range every valuation guide gives you and none of them explain.
The multiples, by category
Rounded numbers, mostly $500K to $8M in purchase price, from searchers buying small businesses across a range of trades.
| Category | Purchase price | SDE multiple |
|---|---|---|
| Landscaping | ~$1.2M | ~2.9x |
| Auto repair | $960K–$1.1M | 2.7x–3.7x |
| Commercial cleaning | ~$950K | ~2.7x |
| Glass | ~$1.2M | ~2.3x |
| Janitorial-adjacent services | ~$225K | ~2.3x |
| Plumbing | ~$2.1M (business only) | ~3.3x |
| Industrial equipment services | ~$2.6M | ~3.3x |
| E-commerce services | ~$3.5M | ~2.8x |
| Recruiting | ~$950K | ~2.4x |
| Clinical practice | ~$5.5M | ~4.3x |
| Festival-adjacent retail | small, distressed seller | ~1.4x |
The clinical practice at 4.3x is the highest multiple in the table, and it's not close. Medical-adjacent businesses with recurring patient relationships get paid a premium for that. The 1.4x retail deal at the bottom was a distressed seller, and it shows. Strip out those two, and pretty much everything else clusters between 2.3x and 3.7x no matter what the business actually does.
Scale changes the number
2 of the biggest deals in this dataset were electrical contractors, and they're the exception that explains the rule. One priced at $7M against about $1.4M in owner earnings, right at 5x. Another set its guaranteed minimum at $6M against SDE, which works out to 4.8x. Both were 50-year legacy shops with real commercial pipelines already built, the kind of scale that starts to look like what private equity pays for at rollup speed in HVAC, plumbing, and electrical.
Everything else in the dataset, the $1M to $3M range where most searchers actually buy, sits closer to 3x. I think that's the honest takeaway: there's a PE floor somewhere around 5x for trades once they hit real scale, and there's a completely different, lower market underneath it that most first-time buyers are shopping in. If a broker quotes 5x on a $1.5M deal, I'd ask what makes it different from the $1.5M deals closing at 2.8x.
The multiple depends on what you count
I've watched the same deal read 2 different ways depending on where the seller note sits. One buyer's numbers on a small consumer products company came out to 2.9x excluding the seller note and 3.8x including it, same purchase price, same SDE, just a different definition of what's in the numerator. Another buyer, on an auto repair shop, described their multiple as landing somewhere between 3.5x and 3.7x depending on the calculation, without ever saying which one they meant.
People report it casually, without specifying which number they mean. When someone tells you their multiple, ask if it includes the seller note or the working capital. The number moves more than you'd expect.
DSCR gets engineered
The DSCRs I've tracked run from about 1.4 on the tightest deal in the set up to roughly 2.4 on the loosest. Most closes land around 1.5 to 1.8. That's where lenders get comfortable without the buyer giving up too much on price. I've also heard, secondhand and only a couple of times, of a deal dying around 1.3 because the bank just wouldn't move (small sample, so hold that one loosely).
What surprised me going through this data is how often DSCR gets built after the fact, standby periods and forgivable notes doing as much work as the purchase price itself. A 2-year standby on the seller note buys you 2 years of lower payments while you stabilize the business. A forgivable tranche tied to hitting SDE targets does something similar on paper, without actually reducing what the seller nets if the business performs.
One deal in the set had a $2M forgivable note tied to a 4x multiple on SDE, sitting on a full 3-year standby. Another had a note that was half forgivable, tied to matching the prior year's revenue. It shows up in the DSCR.
The asking price is a starting position
The gaps between list and close in this dataset are wide enough that I've stopped treating a broker's asking price as anything more than an opening bid. One deal listed at $2.9M, a 6x multiple, and closed at $1.7M, 3.4x. The excavation company from the top of this post listed at $4M and closed at $3.2M. In another, an ask near $1.4M closed at $950K.
The clearest picture of how that gap actually closes is an industrial equipment services deal that started at a $3.2M ask. The buyer's first move was pure comps: he showed the seller what similar businesses were trading at and got the ask down to $2.9M before either side touched the financials. The second move came out of quality of earnings, where the SDE his team verified came in $150K to $200K below what the CIM claimed.
That took the price to $2.6M, 3.3x on the real number. He led with comps, since that move is fast and doesn't feel like an attack. QoE followed, and it's the one the seller can't really argue with once an accountant has looked at the books. If you're building your own case for a lower number, a guide on negotiating the purchase price walks through both moves in more detail.
Treat this as a floor
I'd hold all of this loosely. Every deal here is a win somebody was proud enough to post in a coached community, which means the failed LOIs and the deals that died in diligence are basically invisible in this dataset. I'd guess the real spread of outcomes is wider and uglier than what's above. But real closed numbers, even biased ones, still beat a valuation guide's range every time someone asks me what a landscaping company or a plumbing shop is actually worth right now.