Issue #16 · Ops & Om · Health of Business

Start small. Prove what works. Then scale the winner.

That's the advice. It's what I believed. And it's what I did — right up until I realized the order was completely backwards.

The algorithm doesn't reward caution. It rewards data. Starve it at the start and you spend the entire campaign paying for the privilege of learning nothing.

I've been running Meta ads for a supplement brand — a real client, a real budget, and not a big one. We started where small budgets start: $500 behind a single ad. My instinct was to protect it. Don't spend much until you know what hits, then pour fuel on the winner.

Here's the problem with that plan. The knowledge I was waiting for was the exact thing the spend was supposed to buy. I was rationing the fuel and wondering why the engine wouldn't turn over.

Here's what we'll cover in this issue:

When a campaign launches, the platform knows nothing about who converts. Every early dollar is buying information: which audience, which creative, which placement actually produces a customer. The information is the asset. The sales are the byproduct.

Underfund that stage and you don't get a cheap test. You get a slow, blurry one. The system never collects enough signal to find the pattern, so it keeps guessing — and you keep paying for the guesses. Weeks later you've spent real money and still can't name the ad that works.

So we flipped the order. Heavier spend up front, while the system is learning. Then, once it's clear what converts, pull back — because at that point you're not paying for discovery anymore, you're paying for delivery. The ramp goes down, not up.

That's the whole sequence: fund the learning, then taper into efficiency. Starting low and increasing "once we knew what hit" quietly guaranteed we would never know what hit.

We started at $500 on one ad. That account now runs two ads against $3,500 in total spend. Not because we got braver — because we stopped rationing the part that teaches.

The second half of the problem is memory. An ads dashboard tells you what happened. It will not tell you what you learned. So every campaign restarts from the same instincts and the same blind spots as the last one.

So we built a tracker. Every test goes in: what we ran, what we spent, what we expected, what actually happened, and what it changed our minds about. It isn't a report — it's a training file. Each campaign makes the next one sharper, because the reasoning lives somewhere it can be reused instead of re-derived.

And I don't run it alone. I brief the AI to operate like the most advanced media buyer I could hire — not a chatbot narrating metrics back to me, but an operator with a point of view and the receipts to back it up.

Try this prompt:

That third instruction is the one that matters. It's what makes the thing a loop instead of a line.

Every time the tracker learns something, I go rewrite the project instructions the AI runs on — so the next session starts from what we know instead of from scratch. Longtime readers will recognize the move. It's the same end-of-session ritual from the last issue, pointed at ads instead of email.

Skip it and the loop springs open quietly. The tracker goes stale. The instructions describe a strategy you've already outgrown. And the AI keeps confidently executing last quarter's playbook while you wonder why the magic wore off.

Here's the proof it's a system and not a lucky account: I've since rebuilt the whole thing for a different client — a consulting offer with nothing in common with supplements. Same $500 starting point. It scaled faster, because the loop didn't need building, only re-pointing.

And yes, $3,500 is a small number. It's what the client could commit to. That's exactly the point. The discipline is what transfers, not the budget. Someone spending fifty times that without a loop is just buying fifty times more guessing.

If you have ads running right now and you can't say, in one sentence, what they've taught you — that's the loop to close first.

Bring us the account and we'll build it: what to spend, in what order, and what to write down so it compounds instead of evaporating.

P.S. — Next launch, stop asking "how little can I risk?" Ask "how fast can I buy the truth?"

Join beehiiv live on July 16th at 1PM ET for a first look at the future of audience-led business.

This isn’t just another feature launch (though there will be plenty of those). It’s a look at a more connected future for creators and brands that are tired of juggling disconnected tools, platforms, and data.

If you care about building an audience online, this is worth your time.

Reserve your spot.

If you found this issue helpful, please forward it to a friend or colleague who might benefit from it. Sharing is caring!

Did someone forward you this issue? Don't miss out on future insights — subscribe to the Ops & Om newsletter here!