$186k of ad spend later, one channel turns out to be carrying all of it

How did you handle the deals that touched more than one channel? You mention two social leads that saw a search ad first. If you're crediting the first touch you'll flatter search, and if you're crediting the last you'll flatter whatever closed the loop. With 22 closings the assignment rule could move the whole conclusion.

Also curious how you valued the SEO retainer, since organic doesn't carry a tracking number in the same way.

63% unusable is the part that scares me and I haven't spent a dollar yet. Does the intake field problem exist in every CRM or is that a configuration thing you can set right on day one?

Numbers like $186k of media are a different planet from where I'm working, but the plumbing point translates down. My whole marketing budget is postcards and a Google Voice number and I still don't know which mailing produced which call, because I used the same number twice.

ember, in most systems you can make a field required and lock it from editing after creation. The failure ledger describes, where a reassignment overwrites the original source, usually comes from an automation someone set up two years ago and forgot. It's fixable in an afternoon if you know it's happening.

The thing I keep circling is that you paid $78,000 in retainer over a year for reporting that didn't answer the only question that mattered. Was any of that on them contractually, or does the agreement just promise campaign management and reporting in general terms?

Your three way attribution check is the right instinct and I'd still be careful with it. 22 closings is a small sample and the channels weren't running at equal spend, so search was being judged on $61,000 and social on $76,000. If search had a natural ceiling on volume, which it usually does because search demand is finite, you may have found that you were overspending on social relative to available search inventory rather than that social doesn't work.

The test that would settle it is whether you can push search spend up 50% and hold cost per closed deal. Four months in at $5,100 blended suggests you partly can. Ask the agency for the impression share number on your top search campaigns. If you're already at 80% or 90% impression share, there is no more search to buy and the reallocated money has to go somewhere else.

What is impression share? Genuinely asking, I'm two weeks into reading about any of this.

The line I'd underline is that the agency didn't fight the restructure. Everyone who has priced a service knows what their fee is protecting. When a client finally asks the right question and the vendor folds without argument, the fee was soft.

orchard, or they'd rather have a client on a performance deal for four years than a retainer client who churns in eight months. Folding fast can be a longer game.

How is the per contract fee defined in the restructure? Signed contract or closed file. I've watched a per lead arrangement go bad because the definition of a lead was a form fill, and the volume arrived and the quality didn't. Signed contract is a better anchor and it still leaves room, since a contract you cancel in inspection counts the same as one you close.

On the construction side we solved the same problem with job numbers on every purchase order. Nobody buys anything without one. Sounds like your lead ID is the identical idea and the failure mode is identical too, somebody's in a hurry and skips the field.

$22,000 per closed deal on social and you ran it for nine months before checking. I'm not throwing stones, I ran mailers to a list I never scrubbed for eleven months. But I'd want to know what the agency's monthly report actually showed during those nine months. Was social producing great cost per lead numbers that made it look like the winner?

meridian's question is the one I'd put ahead of the attribution work. The definition of the triggering event in a performance fee is where these arrangements go wrong, and it needs to sit next to a clawback if the contract cancels inside some window. Also worth settling in the same document who owns the ad accounts, the conversion history, and the tracking numbers, because the moment your fee structure gets tied to performance, the value of that history rises for both sides. Have someone who does contracts for a living read it.

Reading this from the other direction. If I ever build a service business in this space, the lesson seems to be that the client who measures you properly is the client worth having, and most of them can't. That's a strange market to enter.

Answering my own question since nobody bit, impression share is the percentage of the available ad auctions for your keywords where your ad actually appeared. If it's low you're leaving volume on the table, if it's near 100 there's nothing left to buy. Correct me if that's off.

rigging, that's right in substance. It's reported per campaign and there are separate figures for how much you lost to budget versus to rank, which matters because losing share on budget means spend more and losing on rank means fix your bids or your quality.

The bit I want to steal is judging SEO on a two year window and saying so out loud to the vendor. Every content retainer I've paid died in month five because I got impatient at exactly the point where it was starting to matter.

For someone about to be an agent rather than an investor, does this whole approach scale down? I'm not going to have $186k of media. Is the plumbing worth building at $800 a month?

halyard, at $800 a month you need one tracking number per campaign and a spreadsheet. The plumbing scales down further than people expect. What doesn't scale down is the sample size, so you'll be making decisions on four closings a year and you have to admit those are guesses.

The most common thing I see, and I've been running a service business for five years, is a client who wants to know cost per deal and has never once written down which deal came from where. ledger did the unglamorous half and that's why the numbers exist.

One caution though. When you move an agency to per contract fees you change what they optimize for. Expect the lead volume to shift toward whatever converts fastest to a signature, which is not always the deal you want to be in. Watch your average spread on the deals that come in over the next six months, not just the count.

thatch is right and I'll say it from inside the agency chair. Under a retainer I optimize for the client staying. Under a per contract fee I optimize for contracts, and if your acquisitions team signs sloppy stuff I get paid for it. The version that works is a modest base that keeps me solvent, a per contract fee, and a clawback if the file dies before closing. ledger, does yours have the clawback?

Every trade I know bills this way already. Deposit, progress, retention held back until it passes. Nothing new under the sun.

I'm building a service offer around exactly this gap, the reconciliation layer between the ad platforms and the CRM, and this thread is the best argument for it I've read. ledger, six weeks of your time to build it. If someone had offered to do that for a fixed fee, what would you have paid?

The reports looked fine for nine months and were useless, and that detail matters more to me than the rest of the thread. Nine months of clean-looking numbers and none of it held up. So a report arriving on time and full of numbers isn't evidence of anything.

juniper, that's the whole post honestly.

Question I'd have been too embarrassed to ask in person. When you say you backfilled three months out of call recordings, does that mean somebody sat and listened to hundreds of calls? Or was there a shortcut. I ask because I'd assume the shortcut exists and I'd rather know before I'm the one with the headphones on.