TL;DR
Agency valuations swing from 2× to 6× on factors most buyers never check. What drives the multiple, what destroys it, and the diligence list.
→ See how this applies to your business (free 30-min call)Marketing agencies are for sale everywhere right now, at prices that range from a bargain to indefensible for businesses that look almost identical on the surface. The reason the spread is so wide is that the number people quote — revenue — is nearly meaningless, and the numbers that actually determine value take a week of digging to find.
I've looked at a lot of these, both as a buyer and as an operator who has rebuilt agency delivery from the inside. Here's what actually drives the price and what to check before you commit.
The Multiple, Honestly
Small agencies trade on a multiple of seller's discretionary earnings — profit plus the owner's salary and personal expenses run through the business. Not revenue. Revenue multiples are what sellers quote when the profit number is unflattering.
Rough ranges in the current market:
Larger agencies with real EBITDA above roughly $1M shift to EBITDA multiples and attract private-equity-style buyers, where the ranges climb further. But most listings you'll see are firmly in the first two brackets.
The Five Things That Destroy Value
Before you look at anything else, check these. Any one of them can turn a 4× business into a 2× business.
1. Client concentration. If one client is more than 20% of revenue, you're not buying an agency, you're buying a relationship — and relationships follow the founder out the door. Get the revenue-by-client table for the last 36 months. Look for a top client that's been growing as a share of the total; that's a business becoming more fragile, not less.
2. Founder-led sales. Ask who closed the last ten clients. If the answer is "the owner" ten times, revenue stops the day they leave. This is the single most common reason acquisitions underperform. Whatever the seller says about referrals and reputation, the pipeline is a person.
3. Project revenue disguised as recurring. "Monthly retainer" that's really a series of projects with no contract and 90-day average tenure isn't recurring revenue. Get actual client tenure distribution. Median tenure under 12 months means you're buying a treadmill.
4. Undisclosed churn. Revenue can be flat while the client roster completely turns over. Ask for a cohort view: clients who started in each of the last eight quarters, and how many are still active. Flat revenue with 60% annual churn is a sales machine attached to a leaky bucket, and the sales machine is the founder.
5. Deferred delivery liability. Money collected for work not yet done. Annual prepayments look great in the bank account and are a liability you inherit. Net it out of the purchase price explicitly.
You're not buying revenue. You're buying the probability that revenue continues after the person who made it happen stops answering the phone.
The AI Question, Both Directions
This is the newest and most consequential variable in agency valuation, and it cuts both ways.
The discount. An agency whose margins depend on billing hours for work that AI now does in minutes is holding a melting asset. Content production, basic ad copy, reporting assembly, first-draft creative — if that's the P&L, the price should reflect a business whose pricing power is actively eroding. Ask what their delivery cost per client has done over the last 24 months. If it hasn't fallen, they haven't adopted, and their competitors have.
The premium. An agency that has genuinely rebuilt delivery around AI — fewer people per client, faster turnaround, systems the client can't easily replicate — has expanding margins and a real moat. That's worth paying up for. Verify it, though: ask to see the actual delivery process, not a slide about it. "We use AI" describes a ChatGPT subscription. "Our lead response system contacts every client's inbound lead within 90 seconds and books qualified appointments automatically" describes infrastructure.
The arbitrage that makes agency acquisition interesting right now: buy a solid, boring, relationship-driven agency at 2.5× with intact client relationships, then rebuild delivery on AI systems. Same revenue, materially lower delivery cost, and a service the incumbent competitors can't match. That's a real strategy — but only if you can actually execute the rebuild, which is a different skill from buying.
The Diligence Checklist
Work through this in order. Stop and reprice if any answer is bad.
Three years of P&L and tax returns. Not a summary deck. Reconcile the two — gaps between what's reported to the IRS and what's in the deck are informative.
Revenue by client by month, 36 months. Concentration, seasonality, churn, and tenure all fall out of this one table.
Contracts. Notice periods, auto-renewal, assignability. A contract that terminates on change of control makes your acquisition conditional on client consent — find that out now, not at closing.
Delivery cost per client. Hours or headcount per account. This is the margin story and the AI story.
Who owns the client assets? Ad accounts, pixels, domains, CRM data. If the agency owns them, that's leverage; if clients own them, switching away from you is trivial.
Team retention risk. Who actually does the work, are they employees or contractors, and are they staying? Get a sense of it directly if the seller allows, and expect some attrition regardless.
The last five lost clients, and why. Sellers are surprisingly candid here, and the pattern tells you what's broken.
Pipeline. Not "we have a lot of interest." Named opportunities with stages and dates. If there isn't a CRM to show you, the pipeline is a story.
Structuring the Deal So You Don't Overpay
Nobody sane pays all cash up front for a service business whose value walks out the door.
If You're the One Selling
The mirror image, and it takes 12-18 months of deliberate work:
The Underlying Point
Agencies are valued on the durability of their revenue, and durability comes from systems rather than relationships. That's true whether you're buying, selling, or just running one. The agency that responds to every client lead within 90 seconds, tracks every ad dollar to a booked job, and delivers with a fraction of the headcount isn't just more profitable — it's worth a materially higher multiple, because the revenue survives the founder.
That's the model we've built at Thinxster: AI callers on the front line, GoHighLevel pipelines carrying the attribution, and delivery that scales without linear headcount. It's carried $102M+ in tracked client revenue at a peak ROAS of 9.2× — and structurally, it's the version of an agency that's worth buying.
If you're buying an agency and want a second opinion on whether its delivery model survives the next three years — or you're selling and want to know what to fix first — [book a free strategy call](/book).
Free Weekly Briefing
One AI Marketing Tactic.
Every Tuesday. Free.
What's actually working across our client accounts right now — ROAS moves, follow-up sequences, creative angles. The stuff that isn't in any blog post yet.
No spam. Unsubscribe anytime. 1,200+ business owners already in.