How to Value an Agency Without Overpaying
Most first-time buyers do not lose money on the deal they walk away from. They lose it on the one they said yes to at the wrong number. This workshop shows you how to build a defensible valuation you can put in front of a seller and hold.
Valuation is simple. We choose to make it complex.
There is an entire service industry built on making this feel like corporate finance. Investment bankers and accounting firms spend twenty five days filling in a template and charge tens of thousands of dollars for it. Meanwhile the seller reads a report that compares their nine person agency to WPP.
- You start with a number in your head. Almost nobody sits down to value a business from a blank slate. Everything you already know about the company becomes part of your preconception, and once you have a preconception, your valuation follows it. If you like the company, your model will find a way to like the price. Here is the test: if your valuations are unbiased, you should raise your estimate as often as you lower it. Most buyers almost never lower it.
- You are anchored to the wrong comparables. I have seen a Carta 409A valuation for a two million dollar digital agency built on EBITDA multiples from WPP and Omnicom. Companies that grow through acquisition, trade publicly, and report to shareholders. That report suggests similar agencies trade between six and forty six times. It is nonsense, and it is the number the seller now believes.
- You are valuing revenue instead of earnings. Brokers list a revenue multiple because it makes the number look bigger. It is pageantry. The same advertising agency, run through a published revenue multiple range, values out anywhere from five hundred thousand to twenty two million. That is not a valuation. That is a spread wide enough to hide any answer you want.
- You lose the deal in the add-backs. Every dollar you let a seller add back raises the price by the multiple. At 4x, accepting a soft ten thousand dollar travel add-back costs you forty thousand dollars. Most buyers negotiate hard on the multiple and then give it all back one line item at a time.
- You treat the multiple as a negotiation. It mostly is not. The multiple is a function of the bracket the business sits in. A one million dollar EBITDA agency and a five million dollar EBITDA agency are not the same business at different scale. They are in different risk brackets, with different buyer pools, and they earn different multiples.
What you'll leave with
The earnings question settled. SDE or EBITDA, which one applies to the agency in front of you, and exactly what changes when the owner stays versus leaves.
The Multiple Staircase. The five brackets, the multiple range each one earns, and why the same dollar of profit is worth more in the bracket above.
The add-back battle. The six add-backs sellers bring to every deal, which ones are fair, which ones are the Porsche problem, and the adjustments that run the other direction and lower the number.
Three methods, triangulated. Earnings multiple, discounted cash flow, and market comps. How to run all three and why anchoring on one is how you get the precisely wrong answer.
Built for buyers who need a number they can defend.
Your number is a guess. You're looking at a real set of financials and landing on a multiple because it sounds about right. You want a method you can show your work on.
The seller is anchored high. They've read a report someone else paid for and it values them like a holding company. You want to move them without losing the deal.
You've never fought the add-backs. You negotiate hard on the multiple, then accept the adjustment schedule as presented. That's where the price actually got made.
You're pricing your first or your fifth. Whether this is deal one or deal five, every acquisition comes down to one number you have to say out loud and hold.
A working session with Peter Lang
The number is not a mystery. It is a method.
Ninety minutes, live on Zoom. The recording goes out to everyone who registers, so sign up even if the time does not work.