Things you've probably already noticed and never had a name for

Ads that worked in 2021 stopped working, and nothing on your end changed.
Every new person's first thirty days look good. Then they don't.
Raising the budget made the numbers worse instead of bigger.
More traffic showed up and revenue didn't move with it.
You can't say, with confidence, which ads made you money last month.

What's actually happening

Meta is extraordinarily good at one specific job: finding the people who were already going to buy from you. Your retargeting pool. People who looked last week and didn't act. Lookalikes built from your existing customers. That's harvesting — collecting demand that already exists.

Harvesting is cheap and fast, and it makes whoever runs the account look like a genius. Including you, in 2021. It works the same way whether you're shipping product, booking jobs, or filling a program — the pool is different, the physics are not.

But that pool has a size, and the size is set by how many customers you already have — not by how much you're willing to spend. Push budget past it and Meta doesn't stop. It starts showing your ads to people who've never heard of you and aren't thinking about this today.

That isn't harvesting anymore. That's paying to create demand. Longer payback, worse first-click numbers, and it needs completely different ads. Same operator, same budget, different physics.

Your account never told you when you crossed that line. There's no alert. The blended number just gets worse.

Most accounts at your size run both of these out of one budget, on one line, judged against one target. They are not the same business.

Harvest

Who it reaches: People already close to buying
Supply: Finite, tied to your customer base
Payback: Fast
The ad that works: A reason to finish what they started

Demand creation

Who it reaches: People who've never heard of you
Supply: Large, and expensive to reach
Payback: Slow, and rougher on the way
The ad that works: A reason to care at all

Then two things removed your ability to see any of it. In April 2021 Apple's tracking changes cut a large share of conversion data out of accounts without server-side measurement, and Meta's default reporting window contracted from 28 days to 7. And percentage-of-spend fee structures meant nobody advising you was paid to say the words stop spending more.

So from the inside, running out of cheap demand and getting worse at your job produce an identical dashboard. You've spent a year or two assuming it was the second one.

It isn't one explanation for one problem. It's the explanation for all of them.

"It worked great in 2021 and then stopped." The pool wasn't exhausted yet — and you could still see what was happening.
"The first month with every agency was good." They inherited warm demand that was already converting.
"They raised the budget and it collapsed." The increase pushed straight through the ceiling into cold traffic.
"More traffic didn't move revenue." Traffic isn't demand. Harvest volume was already capped.
"I can't tell what's working." Apple removed the instrument. Not your competence.
"The attribution tool made it worse." It gave you a third number without separating the two engines.

Roughly nine percent of American businesses with employees ever cross a million dollars in revenue. You're inside a wall that stops nine out of ten people who get as far as you got. That's worth knowing, because for the last year you've been reading it as a verdict on you.

The diagnostic

Five business days. Read-only access to your ad account and your revenue source of truth — no changes, nothing touched. You get a written document at the end, and the order of the work matters more than anything in it.

Your measurement, rebuilt Server-side check, and one agreed source of truth that reconciles Meta against wherever your revenue actually gets recorded — store, CRM, booking system, or books.
Where your cheap demand ends The ceiling in your account, with a number on it and the evidence behind the number.
Your spend, split into its two engines How much is harvesting, how much is creating demand, and what each one is actually returning.
The real constraint, named plainly Media buying, offer economics, retention, or creative volume — including when the honest answer is that it isn't media buying.
The document, which is yours You keep it and everything in it, whether or not we ever work together.

Not included, deliberately: a geo holdout. A holdout needs two to four weeks to produce a number you could defend, so it's the first thing month one does — not something to promise inside five days.

What it costs

$2,000

If you go on to work with us, it comes off your first month. If you don't, you keep the document and we're square. That's the whole arrangement, and this is the only place on this page it gets mentioned.

When to close this page

We turn down more accounts than we take, and these are the reasons why. If one of them is you, a diagnostic would tell you something you can already read here for free.

What's left after you deliver can't fund customer acquisition at any realistic cost. No one running Meta ads can fix that, and buying traffic faster would just lose money faster.
The customers you buy don't stay, don't come back, and don't refer. That's a retention problem wearing a media costume, and it gets worse with more spend, not better.
You're spending under $2,000 a month. There isn't enough account history to analyse, so there's nothing yet to diagnose.
You've never run ads. Different problem, different service — we can't analyse an account that doesn't exist.
You want someone to also handle the website, the email, and TikTok. We do one thing.

If you want a guarantee before you'll start, we're the wrong people. Nobody honest can promise a number they haven't seen the account for.

Who does the work

I've been buying Meta ads since 2011, starting when I was seventeen. That's longer than most of the agencies currently pitching you have existed. A few hundred accounts since then, across most kinds of business, which means I've watched what's described above happen enough times to recognise it before I open the account.

Then I spent years on the other side of it, building and running my own ecommerce brands with my own money. That matters more than the track record does: I know what a bad month costs you, because I've had them with my own inventory and my own payroll rather than a client's.

The first call is a fit check, and someone on our side runs it. Their only job on that call is to work out whether this is actually your problem. They don't run ad accounts and they won't pretend to — ask them something technical and they'll tell you they don't know and bring it to me.

Everything after that call is me. I go into your account. I do the analysis. I write the document, I'm on the call where we go through it, and I run the ads afterward if it goes that way. No account manager, no handoff, no junior inheriting you in month two. Agencies do this in the opposite order — the person who charmed you on the sales call is not the person who ends up in your account, and you usually find that out around week six.

Those brands did well enough that I don't need to take every account that applies, and I don't. I kept doing this because it's the work I'd choose anyway, and one person can only be in a limited number of accounts before the work stops being worth what you paid for it. That's a real ceiling on our side rather than a scarcity line — and what it buys you is a straight no when this isn't your problem, and a straight answer in month three when something isn't working.

The application

Eleven questions, about four minutes. Every one gets read by a person, usually within a day.

Two of them ask what happened with your last agency and what you think is really holding growth back. Those are the ones that decide whether this is worth either of our time, so answer them like you'd answer a friend who asked over a beer — not like a form.