A note before the numbers: every figure in this piece comes from audits we ran, calls with people who buy and sell these businesses, and real transactions. Names and identifying details are changed. Nothing here is invented, and where we are estimating, I say so.

Three quotes, one agency

Call him Dave. He runs a small shop in a mid-size market: $194,000 in revenue last year, and about $128,000 of that ended up being his once you count salary, draws, and the personal spending that ran through the business.

Ask three different people what Dave's agency is worth and you will hear three different-sounding answers. The photographer across town thinks in revenue, because what she'd really be buying is Dave's client list to bolt onto her own operation. A business broker quotes a multiple of owner earnings. A consolidator quotes a multiple of EBITDA, and his multiple sounds the most generous of the three.

Here's what took me years to internalize: those can all be roughly the same dollar figure. The label changes with the size of the agency and who's buying. Most owners never get shown the translation, and that gap is where bad deals live.

The lens changes with your size

Your sizeThe lens buyers useThe honest range
Under ~$300K revenuePercent of revenue, a book sale0.25x to 0.5x revenue for most books. 0.6x is the realistic top, even when the earnings behind it are strong
$300K to $1M revenueMultiple of SDE, owner earningsRoughly 1x to 3x depending on how the business runs
$1M+ revenueMultiple of adjusted EBITDA2x to 4x, climbing with scale. Real platform sales have printed 4.6x and 5.9x

Why does the small end trade on revenue? Because the buyer keeps your clients and drops nearly everything else. Your editing costs, your software, your admin all disappear into her existing operation. She isn't buying a company, she's buying bookings. That's also why even a tired book has a floor: the agency across town will pay a quarter of revenue for clients she doesn't have to win one at a time.

And the top of that band is real, not conservative. We brokered one of these sales ourselves: a solo operation doing $285K in revenue sold for $115K, a 0.40x, and finding the buyer took more than six months. At this size the buyer isn't buying a company, he's buying himself a job, and people who want that job and have the cash are scarce. However strong the earnings underneath, a book under $300K does not clear much past 0.6x of revenue.

0.0x0.2x0.4x0.6x0.8xSale price as a multiple of revenue →Where most books clear0.6x: the realistic ceiling0.40x: a real sale
Under ~$300K of revenue the sale is a book sale. Most clear between 0.25x and 0.5x of revenue; 0.6x is the realistic top even on strong earnings. The gold dot is a sale we brokered ourselves, and it still took six-plus months to close.

Past $1M, the buyer is usually keeping the machine and replacing you, so the conversation shifts to what the machine earns after paying someone to do your job. That number is adjusted EBITDA, and it matters because private equity will not seriously engage below roughly $3 million of it, combined. Nobody in this industry gets there alone, which is the entire consolidation thesis.

Start with the right earnings number

SDE, seller's discretionary earnings, is net profit plus everything you take out of the business: salary, draws, the vehicle, the travel that was really personal. It's the number an owner-operator buyer actually gets to live on, so it's the number they multiply.

Revenue tells a buyer almost nothing on its own. We've audited two agencies within $50K of the same revenue where one kept 45% and the other kept 11%. And the net profit line on your P&L usually understates you, because your accountant's job is to make that number small. Dave's books showed $67,000 of net profit. The audit found $128,000 of SDE. The difference was real money he was already taking, just wearing different labels. What Your Agency Is Actually Worth walks through the add-backs in detail.

What moves you inside the band

Two agencies with identical SDE can price a full turn apart. Our model scores the gap on the things buyers actually probe in diligence, weighted by how hard they price them:

Owner-dependence carries the most weight, by design. The 30-day test: if you step out for a month, does revenue grow, hold, dip, or stall? Who closes new business? Does work reach a client without your hands on it? A buyer pays for what keeps running after you leave the room.

Recurring revenue is next. Per-listing work restarts at zero every month. Retainers and clients who book every month don't. Getting recurring past half your book is worth real turns on the multiple.

Client concentration gets priced twice. It drags your score like the others, and past 50% from one client or brokerage we take a flat 15% off the multiple on top. That mirrors buyer behavior we've seen directly: at that level they either restructure the deal around the risk or walk.

The rest: documented systems, profit trend over two years, team stability, and pricing power, including what your average order says about it. Each moves the number less than the big three, and together they decide whether you price near the floor or the ceiling of your size band.

The 2.83x trap

Now the translation problem that costs owners real money.

Call him Marcus: $1 million in revenue, $225,000 of SDE, pays himself $67,500. A consolidator who is currently buying agencies in this industry pays around 2.83x, and that number is real. Marcus hears it and starts doing happy math on his $225K.

Wrong earnings. The consolidator prices what the business earns after paying someone to replace Marcus. Figure $80,000 for that hire. Now the earnings are $145,000, and 2.83x of that is about $410,000.

Our model quotes the same agency $330,000 to $420,000. The consolidator's bigger multiple and our smaller one describe nearly the same wire transfer. His sounds better because the earnings under it got smaller. So when anyone quotes you a multiple, the only question that matters is: a multiple of what?

Interactive

What would the offer that shows up uninvited look like for you?

Drag to your adjusted earnings: yearly profit after paying yourself a market salary. Every number below comes from the same valuation engine we run on real agencies.

$250K
The consolidator’s standard offer, 2.83×$708K
Same earnings, prepared and shopped, up to 4.1×$1.03M
Systematized and re-rated, up to 5.4×$1.34M
The spread between the first row and the last is the gap this whole piece is about. The top two rows cap at 4.6× and 5.9×, the two real platform sales on record. The full engine, every lever drivable, is just below.

The same shop, cleaned up

Run Marcus forward two years: recurring past 50%, a sales function that doesn't need him, documented ops, the client base spread out. Same revenue, but the rebuild moves the earnings too, because retainers end the discounting and packages lift the average order. Call it $300,000 of SDE now, roughly $220,000 after the replacement salary. The model prices that version at $740,000 to $940,000 today, and $950,000 to $1.3 million once the systems have a year of verified history behind them. A broker who sells these businesses for a living independently quoted us $1.1 to $1.2 million for exactly that mature profile. Nobody negotiates that gap for you. You build it, and the building takes roughly 18 to 24 months of deliberate work.

Here is the whole picture on one chart: the band the model prices inside, the real transactions that pin it, the consolidator's standing 2.83x, and both versions of Marcus.

1x2x3x4x5x6x7x8x$30K$100K$300K$1M$3MAdjusted EBITDA (log scale) →SystematizedClean ceilingMessy floor2.83x standardMarcus todayMarcus, cleaned up
The band the model prices inside, over adjusted EBITDA. Gold dots are real transactions and audits; the dashed gold line is the consolidator's standing 2.83x. The blue dots are Marcus from this article, before and after the two-year rebuild. Hover any dot for the numbers.

And here is the engine itself. Every constant below is the one running in production, so what this prints is what the assessment would tell you.

Interactive

Run an agency through the pipeline yourself.

These are the same constants and the same arithmetic as the live valuation engine. Drag any agency through it and watch which stage decides the number.

$1M
$250K (25%)

Seller's discretionary earnings: profit before you pay yourself.

$80K

A real salary is credited on top of the profit cap, up to $150K, once revenue clears about $1M. Distributions with no salary line get no credit.

6
2

Contractors count half toward credibility. They scale with volume.

55%

Where the nine drivers land between messy (0%) and clean (100%).

Claimed profit25% of revenue
$250K
Credible band for this team45% of revenue plus salary credit: the most the engine believes before an audit
$530K
Underwritten profitWithin the credible band. Used as given.
$250K
Minus owner replacement salary8% of revenue, $50K to $120K
− $80K
Adjusted EBITDAThe number buyers multiply
$170K
Multiple band at this sizeFloor 2.4× to ceiling 3.7×
position 55% → 3.09×
GuardrailsFloor $150K to $220K · ceiling $1.1M
not needed
Value today$463K$589K
Systematized targetSame earnings, target curve$589K$922K
The credible band for your team build is 45% of revenue plus your reported salary (credited up to $150K past ~$1M revenue): the engine underwrites any claim above it and flags the gap for a P&L review instead of pricing a number nobody would pay. A slider is not an audit. The real version of this, run against your actual P&L, is the full assessment.

What this math can't see

The model runs on what you tell it, and owners underestimate themselves in a predictable direction: they forget add-backs. Dave's self-reported profit missed $28,000 of personal spending that legitimately counts. The free assessment gets you a working range in three minutes. The audit exists because a buyer will spend $15,000 to $25,000 per brand on diligence, and walking into that conversation with verified numbers is the difference between being priced on evidence and being priced on doubt.

Run your own numbers. Sixteen questions, the same math you just read, and you'll see exactly which levers move your multiple and by how much.