The multiple in the headline is a large-cap print. Yours has not been published since 2025.

Why does the number in the press not apply?

Because it is measuring a different market. Transactions above $250m of enterprise value averaged 12.2x enterprise value to EBITDA in the first quarter of 2026 (Capstone Partners, Capital Markets Update, 4 June 2026). Global trailing enterprise value to EBITDA held near 10.7x over the same period, with North American large and mega-cap transactions reaching 11x to 13x (VRC, What's the Deal With Deals, published 14 May 2026 and updated 23 June 2026, citing PitchBook, as at Q1 2026).

Those are the numbers that circulate. They are accurate, they are current, and they describe transactions several size bands above the business most owners are asking about. Scale is worth turns, and reading a large-cap print as a valuation of a smaller company is not optimism; it is a category error.

The awkward part is what sits underneath. As at August 2026, the most recent published table of middle-market multiples by size band covers the first nine months of 2025. The publisher's most recent release of any kind was February 2026, reporting full-year 2025. There is no current, public, size-band print for a business between $10m and $250m of enterprise value. An anchor at this size has to be built rather than looked up.

What should the anchor be?

The last published median for the relevant size band, stated with its date, and then adjusted for what is known to have changed since. That is a less satisfying answer than a number, and it is the honest one.

The baseline is the table below: average enterprise value to EBITDA running from 5.9x at $10m to $25m of enterprise value to 10.0x at $100m to $250m, a spread of 2.8 turns across the range (GF Data, published 27 January 2026, covering the first nine months of 2025). The full-year 2025 print across all bands was 7.2x, on 297 private-equity-sponsored transactions, a count down 23% on the year (GF Data, published 18 February 2026).

Then the adjustments, each of which has a current source. Financing capacity at the small end has tightened: senior debt for a borrower under $10m of EBITDA now clears at 2.00x to 2.50x against 2.00x to 3.00x a year earlier (SPP Capital Partners, July 2026), and a buyer who can borrow less pays less. Practitioners expect valuations to hold steady in aggregate over the six months from the start of the third quarter of 2026 (ACG and GF Data, Q3 2026 Market Pulse Survey, published 15 July 2026). And an increasing share of the headline number is being bridged with structure rather than cash: 24% of private-target deals outside life sciences now carry an earnout, of which closer to one in five earnout dollars is actually paid (SRS Acquiom, 7 July 2026).

An anchor built that way survives contact with a buyer. A number lifted from a press release does not, and the cost of the second approach is not embarrassment. It is a process that launches at a price the market will not meet, stalls, and re-prices in public.

Is the median even the right target?

It is the right anchor and the wrong ambition. The median is what a business is worth to a buyer who is not competing for it, and roughly half of transactions clear below it by construction.

The current evidence is that the median is holding while the distance from it widens. Most respondents to the latest middle-market practitioner survey expect valuations to remain steady over the following six months, and they rank mismatched buyer and seller price expectations as the second-largest risk to middle-market activity. One respondent put the split without decoration: multiples for A-grade targets are insanely high, and lower-grade companies are not getting bids (ACG and GF Data, Q3 2026 Market Pulse Survey, 15 July 2026).

Both halves of that sentence matter to an owner. The premium band is real and is being paid. The lower band is not clearing at a discount; in places it is not clearing at all. Which means the useful question is not what multiple the market is paying, but which of those two populations the business is currently in, and what would move it.

What actually produces the premium?

Quality and tension, in that order, and neither is created during a process. Quality is what makes a buyer want the asset. Tension is what makes them pay for it before someone else does. A process with one of the two produces a fair price. A process with neither produces a re-trade.

The quality work is unglamorous and mostly financial. A vendor quality of earnings commissioned early is the highest-return spend available on a founder-owned business carrying heavy add-backs, because it pre-empts the retrade and removes the largest single category of deal killer before a buyer is in the room. Customer concentration, where a single customer above 30% to 40% of revenue is a frequent cause of failure at this size, is cheaper to address before a teaser than after a letter of intent. Management depth, retention and succession decide whether the business is a platform or an add-on, and that classification is worth turns on its own.

The tension work is process design. A broad auction suits a clean, recurring-revenue business. A concentrated or confidentiality-sensitive business is better served by a targeted list of ten to thirty pre-qualified buyers, because a wide list that produces no second bid has cost confidentiality and bought nothing. Either way the buyers have to be qualified for financing before they see a management presentation, since certainty of close is worth real basis points in a market where lenders are underwriting to coverage.

The uncomfortable corollary is that the premium is not always available. If the buyer list cannot produce a credible second bidder, no asking price will manufacture one, and the right advice is to fix the business or wait rather than to launch and discover it.

What happens to the part you do not win?

It gets structured, which is not the same as being paid. Where price will not close a gap, the market currently closes it with earnouts, seller notes and rollover equity, and practitioners report a decent amount of that in 2026 as sponsor-to-sponsor processes stall on valuation (PitchBook News, 6 July 2026).

Price the structured part before accepting it. Contingent consideration is worth roughly a fifth of face on the published payout evidence, so a headline that is a fifth earnout is not the headline. And the certain part is less certain than it looks: 93% of deals carry a purchase price adjustment mechanism and 89% of those with one recorded an actual adjustment (SRS Acquiom, 19 May 2026).

Price to the median, sell to the premium. The first half is an argument with your own expectations, made privately, before anything is launched. The second half is an argument with the market, and it is won by the preparation that happened before the argument started.