Agency owners hear the multiples long before they understand them. Someone in your network sold for eight times profit, someone else got three, and nobody can quite explain why two similar businesses earned such different prices. The answer is rarely luck and never the awards shelf. Buyers pay for profit they believe will survive the handover, and every part of that belief is built, or broken, by your numbers.
This piece walks through an agency sale from the buyer's side of the table: how the price actually gets constructed, which signals move it, what due diligence digs into, and why the numbers you present at the start of a deal follow you around for years afterwards. None of it requires a corporate finance background. It does require starting far earlier than most owners think.
How the price is built
Almost every agency deal starts from the same sum: adjusted EBITDA multiplied by a multiple. EBITDA stands for earnings before interest, tax, depreciation and amortisation. Strip away the acronym and it is simply the profit your agency makes from doing the work, before the cost of any borrowing and before the accounting entries that spread big purchases over several years. It is the buyer's best proxy for the cash the business throws off in a normal year.
The multiple is the number of years of that profit the buyer is willing to pay for today. Pay four times EBITDA and, if nothing changes, the buyer earns their money back in four years. Which means the multiple is really a measure of confidence. The more certain a buyer is that the profit repeats without you standing in the middle of it, the more years of it they will hand over upfront.
The published ranges are wide precisely because confidence varies so much. First Page Sage, which tracks agency transactions, put the average marketing agency multiple at around 6.3 times EBITDA in 2025. FE International's 2026 valuation guide is more useful for our purposes because it splits the range by revenue model: project-only generalist agencies at 2 to 4 times, agencies with a mix of retainers and projects at 4 to 6 times, retainer-heavy agencies with more than 60% recurring revenue at 6 to 9 times, and specialist agencies serving premium sectors such as B2B software or healthcare at 7 to 12 times.
Read that list again, because it is the most important paragraph in this article. The same profit is worth three times as much in one agency as another. The difference is not the quality of the work or the strength of the brand. It is how the revenue is structured and how believable the profit is.
“The multiple is not a market rate. It is a confidence score, and your numbers are the evidence.”
Adjusted EBITDA: where the arguing starts
The word 'adjusted' carries a lot of weight. Nobody expects your statutory accounts to show the exact profit a new owner would enjoy, so both sides normalise the number first. Costs that genuinely will not recur get added back: the one-off legal dispute, the office move, the rebrand that never shipped. Personal spending that has been running through the company comes out too. Each of these is called an add-back, and every add-back that survives scrutiny increases the EBITDA the multiple gets applied to.
The biggest adjustment is usually you. If you have been paying yourself a small salary and taking the rest as dividends, which is sensible tax planning while you own the company, your accounts overstate the profit a buyer inherits, because they will have to pay someone a full market salary to do your job once you leave. That salary gets inserted into the costs and EBITDA comes down. It cuts the other way too: if you have been paying yourself well above the market rate for your role, the excess is added back and EBITDA rises. What no buyer accepts is a fantasy replacement cost in either direction. They benchmark the role against the market and adjust to that number, not yours.
Every add-back needs evidence. On larger deals the buyer commissions a quality of earnings review, a forensic exercise where an accountant tests whether your stated profit is real, repeatable and correctly categorised. A 'one-off' cost that appears in three of the last four years gets reclassified as a normal operating cost. Your adjusted EBITDA drops, and because the multiple applies to that number, every pound of failed add-back can take five or more pounds off the price. The sellers who do best walk in with a conservative, documented adjusted EBITDA that survives the review untouched. It reads as competence, and competence is exactly what the buyer is trying to price.
The signals buyers weigh
Once the profit number is agreed, everything else in the negotiation is about how safe that profit is. Five signals do most of the work, and they will feel familiar because they are the same things that make an agency pleasant to run.
Recurring revenue leads. Retainers and rolling contracts are profit a buyer does not have to re-win. Project revenue has to be sold again from a standing start every quarter, by a team that no longer includes the person who sold it last time. This is why the FE International bands hinge on the recurring percentage, and why shifting your mix from one third recurring to two thirds recurring can do more for your eventual price than several years of top-line growth.
Client concentration is the fastest way to lose a multiple. Once a single client passes roughly 20% of revenue, buyers start modelling what happens if that client leaves the month after completion. Past 30%, many will discount heavily, restructure the deal so that you carry the risk, or walk away altogether. It is an uncomfortable finding because concentration usually comes from doing excellent work for someone who kept buying more. It is still a risk you are asking the buyer to purchase, and they will price it like one.
Owner dependence is the signal founders underestimate most. If the biggest client relationships, the new business pipeline and the final quality check all run through you, the buyer is not acquiring a business. They are acquiring a job with you still in it, and they will structure the deal to keep you in that job for years. The fix is unglamorous and slow: a second tier of leadership, named account owners the clients actually deal with, and sales that close without you in the room.
Margin quality is the health check. Buyers want gross margin, meaning what is left after the cost of the people and tools that deliver the work, comfortably above 50%, and an EBITDA margin above 20%. The UK context makes this a genuine differentiator: BenchPress, The Wow Company's benchmarking survey of UK agencies, found that average gross margin among agencies turning over £1m or more fell to 39% in 2024, the lowest it has recorded. An agency holding a stable 55% gross margin is not just more profitable than average. It is visibly better run than most of the market, and buyers notice.
Clean books tie the whole story together. Three years of reconciled, accruals-based accounts, monthly management accounts, and revenue recognised in the month the work was done rather than the month the invoice went out. None of this raises the price by itself. Its absence lowers the price every single time, because mess reads as risk and risk comes off the multiple.
| Signal | What good looks like | What worries a buyer |
|---|---|---|
| Recurring revenue | 60% or more on rolling retainers or contracts | Revenue re-sold from scratch every quarter |
| Client concentration | Largest client under 20% of revenue | One client above 30%, or the top three above half |
| Owner dependence | Sales and delivery led by the team; the owner takes real holidays | Every key relationship routes through the owner |
| Margins | Gross margin above 50%, EBITDA margin above 20%, stable over three years | Margins that only look healthy after generous adjustments |
| Books and contracts | Three years of clean accruals accounts and assignable client agreements | Cash-basis records, handshake deals, surprise liabilities |
Where due diligence bites
Due diligence is the buyer's inspection period between agreeing a price and completing the deal: weeks of detailed questions covering your financials, tax, contracts, employment and systems. Owners tend to fear it as an interrogation. It is closer to an audit of your last three years of habits, and it rewards boring, consistent bookkeeping more generously than almost anything else in business.
None of these is automatically fatal. Every one of them either trims the price, stretches the timetable or shifts risk back onto you through warranties, which are the legal promises you make about the state of the business, and which follow you after completion. The pattern across all of them is the same: problems disclosed and fixed early are negotiating points, while problems discovered by the buyer's accountants are trust failures. Trust failures are what actually kill deals.
Earn-outs: living with the numbers you showed
Very few agency sales are all cash on completion day. The standard shape is a payment upfront with the balance paid over the following years, part of it contingent on the business hitting agreed targets. That contingent slice is the earn-out. Aventis Advisors, an M&A firm that publishes research on deal structures, puts the typical earn-out at 10 to 30% of the total price, and in agency deals the upfront portion commonly lands between half and 80% of the headline number, with the remainder earned over one to three years.
Here is the part owners miss. The earn-out targets are set from the numbers you presented during the sale. Push an aggressive adjusted EBITDA to win a bigger headline price and you have just set yourself a profit target you must now hit for two or three years, inside someone else's company, under someone else's cost decisions. Sellers who inflate their numbers negotiate a prison and call it a price. Honest, conservative numbers do the opposite: more of the price arrives on completion, and the targets that remain are ones the business can actually reach.
The size of the earn-out is itself a confidence score. Buyers stretch the contingent portion when they doubt the profit will survive without you. Every point of owner dependence and client concentration you remove before the sale converts earn-out money into completion money, which is the only money you should ever fully count on.
The two-year runway
All of this points in one direction: the sale is decided long before the sale process starts. Two years is the practical minimum, and it happens to be the legal one too. Business Asset Disposal Relief, the relief that taxes qualifying gains at 18% in 2026/27 instead of the standard higher capital gains rate of 24%, only applies if its conditions are met throughout the two years ending with the sale: the company must be trading, you must hold at least 5% of the shares and voting rights, and you must be an officer or employee. The relief covers a lifetime limit of £1m of gains, which at the 18% rate rather than 24% is worth up to £60,000 to you personally. Restructure your share classes or drop in a holding company at heads of terms, the document that sketches the deal before lawyers get involved, and you can restart that two-year clock at exactly the moment you cannot afford to.
Two years is also roughly what the operational work takes. You cannot manufacture three years of clean accounts retrospectively, but you can make the next 24 months immaculate. Concentration takes several sales cycles to dilute because the answer is winning other clients, not firing your best one. A leadership layer needs at least a year of visibly running things before a buyer believes it. Start this work when an offer is already on the table and almost none of it is fixable. Start two years out and nearly all of it is.
The agency you could sell is the one you want to own
There is a happy irony running through all of this. Every change a buyer would pay more for makes the business better to own in the meantime. Recurring revenue smooths your cash flow and your sleep. Lower concentration means no single email can ruin your year. A team that runs delivery without you gives you back your evenings and makes the eventual handover credible. Clean monthly numbers mean decisions get made on facts rather than nerve. Prepare properly for an exit and the worst outcome available is that you never sell, and simply own a calmer, more profitable agency that no longer needs you in the building every day.
You do not need to want to sell to act on any of this. You need to want the option, because options have deadlines, and this one is measured in years rather than weeks.
Start with an honest reading of four figures: your recurring revenue percentage, your largest client's share of revenue, your gross margin, and what your EBITDA becomes once a market salary for your own role sits in the costs. Those four numbers are what a buyer would see first, and they are also the clearest map of what two years of deliberate work could change.
Questions agency owners ask
How much is a marketing agency worth?
Almost every agency deal starts from adjusted EBITDA multiplied by a multiple, and the multiple is really a confidence score: the number of years of profit a buyer will pay for upfront. FE International's 2026 guide puts project-only generalist agencies at 2 to 4 times EBITDA, mixed retainer and project agencies at 4 to 6 times, retainer-heavy agencies with over 60% recurring revenue at 6 to 9 times, and specialists serving premium sectors at 7 to 12 times. The same profit can be worth three times as much depending on how the revenue is structured and how believable the profit is.
What is adjusted EBITDA?
EBITDA is profit before interest, tax, depreciation and amortisation, and the adjustment normalises it to what a new owner would actually inherit. Genuine one-off costs get added back, personal spending comes out, and a full market salary for your role gets inserted if you have been paying yourself a small salary and taking dividends. Every add-back needs evidence, because a failed add-back can take five or more pounds off the price for every pound of profit it claimed.
What do buyers look for when buying an agency?
Five signals do most of the work: recurring revenue at 60% or more, no single client above roughly 20% of revenue, a business that runs without the owner, gross margin above 50% with an EBITDA margin above 20%, and three years of clean accruals-based accounts. They are the same things that make an agency pleasant to run. Mess in any of them reads as risk, and risk comes off the multiple.
How much client concentration is too much when selling an agency?
Once a single client passes roughly 20% of revenue, buyers start modelling what happens if that client leaves the month after completion. Past 30%, many will discount heavily, restructure the deal so you carry the risk, or walk away altogether. Diluting concentration takes several sales cycles, because the answer is winning other clients rather than firing your best one, which is one reason preparation starts years out.
What is an earn-out when selling an agency?
An earn-out is the contingent slice of the price, paid over the following years only if the business hits agreed targets, and it typically runs at 10 to 30% of the total. In agency deals the upfront portion commonly lands between half and 80% of the headline number, with the remainder earned over one to three years. The targets are set from the numbers you presented during the sale, so inflating your adjusted EBITDA negotiates a prison and calls it a price.
When should I start preparing my agency for sale?
Two years out, at minimum. That is the legal window for Business Asset Disposal Relief, which taxes qualifying gains at 18% in 2026/27 instead of the standard 24% and is worth up to £60,000 across its £1m lifetime limit, provided its conditions are met throughout the two years ending with the sale. It is also roughly what the operational work takes: concentration needs several sales cycles to dilute, and a leadership layer needs at least a year of visibly running things before a buyer believes it.
