What a business valuation multiple is, and what moves it

A valuation multiple is the number you multiply a business’s annual earnings by to get a price — so a business earning £200,000 a year valued at 4× is priced at £800,000. The part almost everyone misses is that the multiple is only half the sum, and the earnings figure it is applied to moves the answer just as much. The same business, at the same 4×, is worth £576,000 or £356,000 depending purely on which profit figure you multiply — and this page works that through in full, along with the six things that genuinely move a multiple up or down, and how to build a defensible range of your own instead of borrowing an industry average that describes nobody.

What a multiple is, and what it is applied to

Strip away the language and an earnings-multiple valuation is one line of arithmetic:

Value = normalised earnings × multiple

That is the whole method. A business earning £200,000 a year, valued at 4×, is priced at £800,000. Everything else in a valuation — the normalising, the add-backs, the negotiation, the due diligence — is an argument about one or other of those two numbers.

Because there are only two inputs, people fixate on the multiple. It is the memorable one, the one that gets quoted in the pub and asked about on forums. But the multiplication is symmetric: a 10% movement in the earnings figure changes the price by exactly as much as a 10% movement in the multiple. And the earnings figure is far more elastic than most owners expect, because there is no single number called “the profit”.

The same business, three different earnings figures

Take a small services business turning over £850,000. Its accounts, prepared the way small company accounts usually are, show an operating profit of £84,000. Now rebuild it the way a buyer would.

Nothing about the business changed between the top of that table and the bottom. The same vans, the same customers, the same bank balance. But now apply a single agreed multiple of 4× to each of the three figures:

A spread of £240,000 on one business, with one multiple, on one day. That is why “we were offered 4×” is a sentence with no information in it until somebody says four times what.

The two figures that matter in practice are SDE and EBITDA, and they encode a different assumption about the buyer. SDE adds one working owner’s entire package back, because it assumes the buyer will step into the job and take the owner’s pay themselves. EBITDA leaves a market-rate replacement salary in as a real cost, because it assumes the buyer will hire someone. Neither is wrong; they answer different questions, and they carry different multiples precisely because they mean different things. Our SDE vs EBITDA guide works through which one a given buyer will reach for.

Converting between them, so you can compare like with like

Because SDE is the larger number, an SDE multiple is always the smaller multiple for the same price. The business above, priced at £576,000, is on 4× SDE. Expressed against EBITDA, the same price is:

£576,000 ÷ £89,000 = 6.47× EBITDA

4× and 6.47× are the identical deal. If you take a multiple you heard quoted against EBITDA and apply it to your SDE, you will overvalue your business by a wide margin and then spend the sale process being talked down from a number you invented. The first discipline of using multiples is that you never quote one without naming its base.

The industry multiple table you came here for, and why it cannot help you

Most people who look up valuation multiples are really looking for one thing: a table with their sector on the left and a number on the right. There are plenty of those tables published. This page does not contain one, and the reason is not caution — it is that the table cannot do the job you want it to do, and the arithmetic showing why is short enough to work through here.

A published sector multiple is an average. To produce it, somebody took a set of completed sales in a sector and averaged the multiples those businesses achieved. That is an honest thing to compute. The problem is what it does when you apply it to one specific business.

Two businesses, same trade, same earnings

Take two businesses in the same sector, both with normalised SDE of exactly £200,000. On a published table they are the same row.

Now suppose — purely to demonstrate the arithmetic — that A sells at 4.5× and B sells at 2.5×. Those two figures are illustrative; they are chosen to make the sum legible, not quoted from anywhere.

The average is arithmetically perfect and describes neither business. It underprices A by £200,000 and overprices B by £200,000. As a proportion of the figure it hands you, it is wrong by £200,000 ÷ £700,000 = 28.6% in both directions.

And notice what has happened to the two owners. A’s owner, armed with the sector average, leaves £200,000 on the table and never finds out. B’s owner goes to market at £700,000, gets interest, gets an offer conditional on due diligence, and then watches the offer fall apart when a buyer works out that 41% of the revenue sits with one customer who is on no contract at all. Both owners were let down by the same number.

Why the spread does not average out with more data

The instinctive objection is that a real published average is built from hundreds of sales, not two, so it will be more reliable. It will be a more reliable average. It will not be more reliable for your business, because the thing driving the spread is not sampling noise — it is real, permanent differences between the businesses in the sample. A larger sample measures the average of those differences more precisely. It tells you nothing about which end of them you are on.

There is a second, quieter problem. To build a sector table you need a sector definition, and small businesses resist being sorted into one. A business that installs and then maintains equipment might sit under construction, under facilities management, or under maintenance services, and those are three different rows. Its maintenance contracts might be 70% of its profit while its installation work is 70% of its revenue, in which case the row it lands in is decided by which figure the classifier used. Meanwhile the thing a buyer will actually price — that the maintenance revenue recurs and the installation revenue does not — appears in no sector table anywhere.

What you should take from a published table

They are not useless; they are just not a valuation. Read them as a sanity check on the order of magnitude, and nothing more. If everything you can find suggests businesses like yours change hands somewhere in the low single digits of SDE, and your own working lands at 11×, you have made an error somewhere and should go and find it. That is a genuine use. Taking the midpoint of the table and calling it your price is not.

The rest of this page is the alternative: understanding what the multiple is actually measuring, then building a range from your own business’s characteristics that you can defend line by line when a buyer pushes on it.

A multiple is a statement about risk and durability

Here is the sentence that makes multiples make sense. A multiple is the number of years of current earnings a buyer is willing to hand over up front in order to own those earnings from then on.

Multiple = years of today’s earnings a buyer pays to own tomorrow’s

At 3×, a buyer is paying three years of profit to own the business. If the profit holds, they have their money back in three years and everything after that is return. At 6× they wait six years before they are square. Nobody waits six years for earnings they think are fragile.

So the multiple is not a measure of how good the business is, or how hard the owner worked, or what it cost to build. It is a measure of one thing: how confident a buyer is that the earnings will still be there. Every factor in the next section is a factor because it changes that confidence.

What a wrong multiple costs a buyer

Work it from the buyer’s side and the logic becomes obvious. Take SDE of £200,000 and see how long the buyer waits to get their money back — first if the earnings hold, then if the earnings drop by 20% to £160,000 after completion, which is exactly what an owner-dependent business does when the owner leaves.

The same 20% shortfall does more damage the higher the multiple goes: half a year of extra waiting at 2×, a year and a half at 6×. A high multiple and a fragile earnings stream are a combination a buyer cannot accept, because the high multiple is precisely what removes their room to absorb a bad year. That is the mechanism behind every discount in the next section — it is never a punishment, it is a buyer sizing their exposure.

And why most buyers cannot pay a high multiple even if they want to

There is a hard constraint underneath this. A large part of most small business purchases is borrowed, and the borrowing is repaid out of the earnings being bought. So the price has to service itself. If the buyer of the £800,000 business above has borrowed £500,000 over five years at, say, 9% — again, an illustrative rate, not a quoted one — then in year one the business must find £500,000 ÷ 5 = £100,000 of capital plus £500,000 × 9% = £45,000 of interest, which is £145,000 against £200,000 of earnings. That leaves £55,000 of headroom before the buyer has taken a penny of income or replaced a single vehicle.

Now run it at £160,000 of earnings and the headroom is £15,000. The deal has not failed, but it has stopped being comfortable, and a buyer who has run that sum will not go above a multiple where it still works. This is why price and deal structure are one negotiation rather than two, and it is the same arithmetic our management buyout guide runs from the buying side.

The six things that actually move your multiple

If the multiple measures confidence in the durability of the earnings, then the things that move it are the things that make those earnings more or less likely to survive the change of ownership. There are six that do most of the work, and every one of them is visible in your own business today.

Concentration, costed

Two of those deserve arithmetic rather than description, because owners routinely underrate them. Start with concentration. Take Business B from earlier: revenue £900,000, SDE £200,000, largest customer 41% of revenue.

The overheads do not leave with the customer, which is why losing 41% of revenue takes 74% of the earnings. Now put that beside a price. A buyer who paid 4× on the pre-loss figure has handed over £800,000 for a business now earning £52,400 — a payback period of £800,000 ÷ £52,400 = 15.27 years. They did not buy a 4× business. They bought a 15× business and found out afterwards.

No buyer who has run that sum pays the same multiple for B as for A, and no amount of arguing about sector averages changes it. That single sum is why concentration is the fastest way to lose a multiple and, incidentally, why it is usually the most valuable thing an owner can spend two years fixing before a sale.

Records, costed

Quality of records looks like the boring one on the list, and it has the most direct arithmetic of all, because it does not move the multiple at all — it moves the earnings figure the multiple is applied to. Every add-back you cannot evidence gets removed, and every pound removed is multiplied.

Price impact of a disputed add-back = the add-back × the multiple

Go back to the £9,000 rebrand in the first table. If you can produce the invoice and show it was a single project that will not recur, it stands, and at 4.5× it is worth £9,000 × 4.5 = £40,500 of price. If your records cannot separate it from ordinary marketing spend, it comes out, and so does the £40,500. The £3,000 of personal travel behaves the same way: £3,000 × 4.5 = £13,500, kept or lost on whether the expense claims distinguish it.

Four disputed add-backs totalling £30,000 move the price by £135,000 at 4.5×. That is what bookkeeping is worth in a sale year, and it is the one factor on the six-item list that a determined owner can genuinely fix inside twelve months.

The one that is not on the list

Revenue. Owners often assume a bigger top line buys a bigger multiple, and it generally does not by itself — two businesses on identical revenue can carry completely different multiples for every reason above, and the multiple is applied to earnings in any case. If you are working from a turnover figure, our guide to valuing on revenue sets out where revenue multiples come from and what they quietly assume about margin.

Building your own defensible range instead of borrowing one

So if the published table will not do it, what does? A four-step method that produces a range you can defend line by line. It will not hand you the certainty a table appears to, but what it gives you is better: every number in it has a reason attached, and reasons are what survive a buyer pushing back.

Step one: fix the earnings base and write it down

Decide whether you are working in SDE or EBITDA, normalise the accounts to get there, and then never mix them again in the same conversation. Write the figure at the top of the page with its label next to it. Half the confusion in small business valuation is two people multiplying different numbers and arguing about the multiple. Our full method for the rebuild is in how to value a small business.

Step two: anchor from comparables you can actually see

You need a starting point, and it has to come from somewhere real. The honest sources are all ones you have to go and get: completed sale data from an accountant or broker who works in your sector, businesses currently listed for sale that genuinely resemble yours, and any approach you or a competitor has had. Two cautions worth holding on to. An asking price is not an achieved price, so a listing tells you what a seller hoped for. And a headline figure that includes property, stock or deferred elements is not comparable to one that does not, so you have to know what is inside the number before you use it.

If you cannot find anything solid, say so and work the rest of the method from a deliberately wide anchor. A wide range that is honest beats a narrow one that is invented.

Step three: score the six factors against that anchor

Suppose your enquiries give you a plausible anchor band for businesses like yours of 3.0× to 4.5× SDE, midpoint 3.75×. That band is illustrative here — yours comes from step two, not from this page. Now take the six factors and mark each as a strength, a weakness or neutral for your business, and move the midpoint by 0.25× for each. The 0.25 weighting is a device to make the reasoning explicit and arguable; if you think concentration matters more than records in your trade, weight it more and say why.

Step four: turn the point into a range, then sanity-check it

A scored figure is still a single number, and a single number is the wrong shape of answer. Put a band of ±0.5× around it and price the ends.

Then apply the payback test from earlier as a reality check, because it is the test the buyer will apply. At the top of A’s range a buyer waits 5.5 years to get their money back, and A’s earnings are defensible enough that this is arguable. At the top of B’s range a buyer waits 3 years on earnings where one customer departure takes 74% of the profit — which tells you that even 3× will need defending, and that B’s real work is not the negotiation but the two years before it.

Notice what you now have that a table could never give you. Not just a number, but six specific sentences explaining it, each pointing at something in the business. When a buyer opens with a lower multiple, you are not reduced to insisting; you can ask which of the six they are marking you down on, and that is a conversation you can have with evidence.

Why the answer is a range, and how a sensitivity table works

Both halves of the sum are uncertain. The earnings figure moves as add-backs are challenged; the multiple moves as a buyer prices the six factors. A single valuation figure hides both uncertainties behind a false decimal point, and the moment anything shifts, the whole thing has to be rebuilt.

A sensitivity table shows the uncertainty instead of hiding it. Put the plausible earnings figures down one side, the plausible multiples across the top, and fill in every combination. It is the same arithmetic as before, done twenty times.

A full sensitivity table

Take a business whose normalised SDE lands somewhere around £200,000 depending on how the add-back arguments go, with an anchor range of 3.0× to 5.0×.

Read the corners: £480,000 at the bottom left, £1,100,000 at the top right. That is a spread of £620,000, and the high corner is £1,100,000 ÷ £480,000 = 2.29 times the low one. Every cell in that grid is defensible arithmetic. Anyone who hands you one figure from inside it and calls it the value has simply not shown you the other nineteen.

What the grid is for

Three things, and none of them is decoration.

• It tells you which lever to pull. Moving one row up, from £200,000 to £220,000 of evidenced earnings, is worth £900,000 − £810,000 = £90,000 at 4.5×. Moving one column right, from 4.0× to 4.5× at £200,000, is worth £900,000 − £800,000 = £100,000. Those are comparable in size — but the row move is usually done with bookkeeping and evidence inside a year, while the column move needs the business itself to change. That comparison is the single most useful thing the grid produces.
• It prepares you for the negotiation you will actually have. A buyer does not open by rejecting your valuation. They open by removing two add-backs and marking you down half a multiple, which moves you from £900,000 to £720,000 in one sentence. If you have the grid in front of you, you already know that cell, you know exactly which two inputs produced it, and you can argue the inputs rather than the conclusion.
• It stops one bad assumption compounding. A single figure carries its errors invisibly. A grid makes you state your earnings range and your multiple range out loud, which is where you notice that the top of your multiple range assumes a contract you have not signed yet.

Where to land

The useful output of a valuation exercise is not a price. It is a range, a written reason for each end of it, and a clear view of which of the six factors is costing you the most. With those three things you can decide whether to go to market now, or spend eighteen months moving a row and a column first — and that decision, rather than the number itself, is usually where the money is.

Two closing cautions. First, none of this covers what the price is finally paid in: cash at completion, deferred instalments, an earn-out tied to targets. A high headline multiple paid mostly on an earn-out is a smaller certainty than a lower one paid in cash, and comparing offers on headline multiple alone is how owners get caught. Second, the grid values the trading business. Surplus cash, property, stock and debt are all dealt with separately, and how they are dealt with is set out in the sale agreement rather than the valuation.

Arithmetic and general information only — not financial, tax, legal or investment advice. Your own figures, and your own accountant, decide what any of this means for you.

Frequently asked questions

What is a typical business valuation multiple for my industry?

There is no figure anyone can honestly give you, and a published industry average is close to useless for a single business. That average is taken across businesses with different owner dependence, different customer concentration, different margins and different quality of records, and those differences drive the price far more than the sector does. Two businesses in the same trade with identical earnings can justify very different multiples, so an average sits between them and describes neither. Use a published table only as a rough sanity check on the order of magnitude, then build your own range from your own business.

What is a business valuation multiple?

It is the number you multiply a business annual normalised earnings by to arrive at a price. A business earning 200,000 pounds a year valued at four times is priced at 800,000 pounds. In practical terms the multiple says how many years of current earnings a buyer will pay up front to own those earnings from then on, which makes it a measure of how confident the buyer is that the earnings will last.

Is a valuation multiple applied to turnover or to profit?

Almost always to a measure of profit rather than turnover, and which measure matters enormously. The two in common use are SDE, which adds one working owner full pay and benefits back on the assumption the buyer does the job, and EBITDA, which leaves a market-rate replacement salary in as a cost. The same business at the same multiple produces very different prices on the two bases, so a multiple quoted without naming its base carries no information.

What makes a business valuation multiple go up?

Six things do most of the work: low owner dependence, revenue spread across many customers rather than a few, recurring or contracted revenue rather than one-off projects, margins that hold steady across several years, growth a buyer can attribute to something repeatable, and records good enough that every add-back is evidenced. Each one is really the same question in a different form, which is how likely the earnings are to still be there after the owner has gone.

Why should a business valuation be a range rather than a single figure?

Because both halves of the sum are uncertain. The earnings figure moves as add-backs are challenged in due diligence, and the multiple moves as a buyer prices the risk factors. A sensitivity table that crosses a range of earnings with a range of multiples shows every defensible combination at once, and it also shows which lever is worth more, since a small rise in evidenced earnings can be worth as much as half a turn of the multiple.

Other guides

• How much does Etsy take from a sale?
• How to build a cash flow forecast
• Agency utilisation rate: the formula, and the denominator that decides it