Two advisors quoting the same $20m sale can be $520,000 apart.

What does a sell-side advisor charge?

Three components, in almost every case. A work fee paid during the engagement, a success fee paid at closing, and reimbursement of out-of-pocket expenses. The work fee is usually credited against the success fee, so it is a timing device rather than an addition: 77% of advisors deduct engagement fees from the success fee (Axial, 2026 M&A Fee Guide, 331 responses collected in Q2 2026, published 23 June 2026), while Firmex's survey puts the share at 57%, up from 53% the year before (Firmex, Global M&A Fee Guide 2024-2025, an online survey completed by 456 middle-market professionals in December 2024 and January 2025). Two surveys, two answers: both are quoted here because the gap itself is information about how soft this data is.

Where the market has moved is the shape of the work fee, not its existence. In the December 2024 survey, monthly fees remained the most common structure at 30% of firms, just ahead of one-time fixed retainers at 28%, with milestone fees payable at defined points in a process growing to 13% and hourly fees to 9% (Firmex, December 2024). By Q2 2026, 71% of advisors charged some form of upfront fee, and nearly a third charged none at all, up from 19% in 2024 (Axial, June 2026).

The success fee is then quoted on one of three shapes. A declining scale, the Lehman formula and its variants, where the percentage falls as the price rises. A flat percentage regardless of size. Or an accelerator, where the percentage rises above an agreed target price. In the December 2024 survey the split was 41% declining, 27% flat and 21% accelerator, with the accelerator share up from 16% a year earlier (Firmex, December 2024). By Q2 2026 flat-percentage structures had risen to 36% of a different survey's responses, with declining scales and flat fees together accounting for 79% (Axial, June 2026).

The classic Lehman scale is 5% of the first $1m of price, 4% of the second, 3% of the third, 2% of the fourth and 1% of everything above $4m. Double Lehman doubles each tier to 10, 8, 6, 4 and 2%, and is the version still in general use below roughly $25m to $50m of value. Both were designed in an era of much smaller deals, which is why the effective rate on a modern middle-market transaction rarely resembles either headline number.

How much is the success fee?

On the most recent grid this article has verified in full, the average effective success fee ran from 4.8% on a $5m transaction to 4.0% at $10m, 3.4% at $20m, 2.7% at $50m, 2.0% at $100m and 1.9% at $150m (Firmex, Global M&A Fee Guide 2024-2025, surveyed December 2024 and January 2025). The $5m rate fell from 5.5% a year earlier; at nearly every other size the level barely moved. European fees ran below North American ones throughout, by 1.1 percentage points on smaller deals narrowing to 0.4 at the largest.

The average is the least interesting number in that grid. At $20m, in the most recent survey to publish a distribution rather than an average (Firmex, surveyed December 2023), 80% of the quotes fell between 2.0% and 4.6%. That is a spread of 2.6 percentage points across the middle four-fifths of the market, or roughly $520,000 of fee on the same $20m sale, before anyone has argued about what is included in the price the fee is calculated on. The 2024-2025 edition published no distribution at all, so that range is quoted at its December 2023 vintage.

Dispersion that wide is not noise. Advisors rate three inputs into a success fee quote as very important in a near-tie: transaction complexity at 66% of respondents, the risk of not closing at 66%, and engagement size at 65% (Firmex, December 2024). A quote is therefore partly a statement about the business being sold. A number at the top of the range on a clean asset is a different signal from the same number on a concentrated, founder-run one.

Two definitional points move real money and are settled in the letter, not the market. First, what the percentage applies to, and when it is paid: 53% of advisors take the full success fee at closing regardless of when the seller is paid, while 41% collect the fee on deferred components such as seller notes and earnouts as and when the seller receives them (Firmex, December 2024). Second, the tail, the period after termination during which a transaction still triggers a fee, which is commonly 12 to 24 months.

What is the work fee, and is it credited?

In the December 2024 survey, $5,000 to $10,000 a month remained the most common monthly fee, named by 45% of the advisors who charge one, while the share charging $16,000 or more a month roughly doubled to about 15%. One-time fixed retainers moved the same way: the most common level stayed at $5,000 to $10,000, but 38% of firms charging them now set the fee above $26,000, up from 26% a year earlier. Larger firms and those with higher minimum transaction sizes charge above those averages (Firmex, December 2024).

Advisors are direct about what the fee is for. It funds the preparation work that happens before any buyer sees anything, and it tests whether the seller is committed. Firms describe it as a commitment fee more often than as compensation, and the surveyed comments repeat the point that engagement fees rarely cover the cost of a broken deal.

The number to negotiate is not the monthly figure. It is whether the fee is credited in full against the success fee, whether it stops if the process is paused, and what happens to it if the seller withdraws. Those three answers change the economics of a stalled process far more than the difference between $7,500 and $10,000 a month.

How good is this fee data?

Thinner than the confidence with which it is usually quoted, and worth saying so.

There is no regulator, exchange or professional body that publishes middle-market M&A advisory fees. Every figure above comes from a vendor survey of advisors: Firmex, a data room provider, and Axial, a deal network. Both survey the sell side about what it charges, which is self-reported pricing rather than observed engagement letters or realized fees. Neither audits the answers.

The two survey families also answer different questions. The most current, Axial's, was fielded in Q2 2026 and published in June 2026, but reports fee structures rather than fee levels: it says how many advisors use a declining scale, not what the scale is. The most current grid of fee levels is Firmex's M&A Fee Guide 2024-2025, the eighth annual edition, drawn from 456 advisors across 49 countries and surveyed in December 2024 and January 2025; this article verified that grid in full against Firmex's published Global Edition on 12 October 2026, which is why every fee level above carries the December 2024 vintage. So the shape of the market is current to this year and the price of it is coming up on two years old, in a period when the share of firms raising at least one type of fee fell from 47% in 2023 to 34% in 2024 in the global survey (Firmex, December 2024), and to 30% in Firmex's North American edition (289 advisors in the United States, Canada and Mexico, published 1 May 2025 on 2024 data), while 37% of firms said they planned a fee increase in 2025. No fee-level survey fielded in 2025 or 2026 has been published by any named source this article could verify as at October 2026.

The widely circulated fee-band tables are older still. The most linked one dates to a September 2016 article last updated in March 2024 (Divestopedia, A Summary of M&A Fees for Sell-Side Transactions), and puts retainers on sub-$100m transactions at $50,000 to $150,000 in total with a guideline that they should not exceed 15% of combined fees. The academic literature on advisory fees is real but almost entirely concerned with large public-company transactions, which price nothing like a $20m private sale.

Treat every number in this article as a well-sourced central tendency with a wide distribution around it and a date attached, not as a rate card.

What does the fee buy?

Work that happens before a buyer exists, and tension once one does. The preparation phase produces the normalized earnings picture, the buyer universe, and the materials. The process phase produces competition, which is the only mechanism that moves a price above the first credible offer. Neither is visible in a fee schedule, which is why fee comparison across two proposals is rarely a like-for-like exercise.

The comparison that is like-for-like is structural. A declining scale pays the advisor a lower marginal rate on exactly the dollars a seller most wants fought for. An accelerator, where the percentage rises above an agreed target, inverts that. Roughly one advisor in five now offers the accelerator shape, up from one in six a year earlier (Firmex, December 2024), so it exists in the market and is available to ask for.

The other structural question is what the fee attaches to. If a meaningful part of the consideration is contingent, the treatment of earnout dollars in the fee schedule decides whether the advisor is paid on money the seller may never receive.