CAC Spiked After You Scaled? Check Where Meta Is Actually Delivering Before You Kill the Campaign

You pushed the budget up, CAC jumped, and now you're sitting there with your finger over the off switch wondering if the whole campaign is cooked. Before you kill it, I want you to ask one boring question that most people skip: where is Meta actually sending these ads now?

Nine times out of ten, that question is the entire diagnosis. The campaign isn't broken. The creative isn't suddenly bad. The delivery quietly drifted, and your CAC followed it.

I've watched this exact pattern play out enough times that it's now the first thing I check, not the last. So here's the teardown of what's really happening when CAC spikes right after a scale, and the diagnostics we run before anyone touches the kill switch.

What actually happens when you scale

Here's the mechanism, because once you see it you can't unsee it.

When you increase budget meaningfully, you're asking the algorithm to go find more people, faster. To spend that bigger number it widens the net. And widening the net is not neutral, because it tends to drift toward whoever is cheapest and easiest to get a response from.

The classic version: delivery starts piling into an older audience, say 55-plus, who happen to click at a high rate but convert at a low one. From Meta's side that can look like a perfectly good audience, lots of engagement, cheap clicks. From your side it's a disaster, because those clicks don't turn into purchases and your cost per acquisition balloons.

I saw this with a pet brand we were unpicking. They'd run a deliberate test, pushed Meta budgets up hard to find the ceiling, and CAC spiked. The assumption in the room was that they'd simply hit the limit of profitable spend and the channel was tapped out. The actual cause, found in the postmortem about a week later, was that the platform had started sending most of the ads to the 55-plus bracket with the highest click-through rate and the lowest conversion rate. Nobody changed the targeting. The delivery shifted on its own. They weaned spend off that audience and CAC came straight back into line.

That's the whole trap. The spike reads like "we're out of room" when it's really "the algorithm is buying the wrong clicks". Two completely different problems, and you only get to tell them apart if you look at where delivery went.

The delivery-diagnostics checklist

This is the run I'd do after any budget increase, in order, before concluding the campaign is done. None of it is fancy. It's just looking in the places people don't look.

• Age breakdown. Pull the age report on the campaign for the period since you scaled and compare it to the period before. If a band you don't actually sell to has quietly swallowed a big share of spend, there's your culprit. The 55-plus high-click-low-buy pattern is the usual offender, but it can be any band that's cheap to reach and weak at converting.
• Placement breakdown. Same exercise across placements. A scale often dumps a chunk of new budget into the cheapest inventory, and the cheapest inventory is frequently the lowest-intent. If spend has tilted hard toward placements that historically convert poorly for you, that drift alone can explain the CAC move.
• CTR versus CVR divergence. This is the tell that ties it together. Healthy delivery has click-through and conversion moving roughly together. The failure mode is CTR holding up or even rising while conversion rate falls off a cliff. That gap is the signature of the algorithm chasing cheap engagement from people who were never going to buy. When you see clicks up and purchases down, you're not looking at a creative problem, you're looking at an audience problem.
• New-versus-existing split. Check whether the extra spend is actually finding new buyers or just re-touching people already in your world at a worse rate. Scaling is supposed to bring incremental customers. If the bigger budget is mostly recycling, your blended CAC will rise even though nothing "broke".

Run those four and you'll almost always land on the same answer: delivery moved somewhere it shouldn't have, and the fix is to steer it back, not to switch everything off.

Fixing it without nuking your signal

Here's where people overcorrect and make it worse, so this is the part to be careful with.

The instinct, once you spot the bad audience, is to slam on hard exclusions and tight age and interest restrictions everywhere. The problem is that aggressive constraints starve the conversion data Meta needs to optimise, and you can choke the signal feeding your CAPI setup in the process. You fix the CAC symptom for a week and quietly degrade the thing that was making the account work in the first place.

The cleaner move is to steer delivery rather than cage it. A few ways I'd approach it, lightest touch first:

• Wean spend off the offending audience deliberately. In the case above, the fix wasn't a dramatic teardown, it was easing budget away from the band that was eating spend without converting, and letting delivery rebalance. You're nudging, not amputating.
• Use structure to shape who you reach, not just exclusions. A lot of "signal engineering" is really just thoughtful campaign architecture, breaking things out so each campaign or ad set is pointed at a genuinely distinct audience instead of dumping everything into one bucket and hoping. Done well, the structure itself reaches a cleaner audience without you having to bolt restrictions onto every layer.
• Protect the conversion event. Whatever you change, keep the purchase signal flowing cleanly back to the platform. If you're running CAPI, the goal is to keep that server-side data intact and let the algorithm optimise toward real purchases, so it self-corrects toward buyers rather than clickers. Tightening so hard that conversions dry up is the one own-goal to avoid.

The mindset is: clean the signal, don't sever it. You want Meta optimising toward the right outcome with good data, not flying blind inside a cage you built in a panic.

The reframe most people miss

There's a second layer here that changes how urgent the spike even is.

CAC is only half of the equation. The number that actually matters is payback, how long it takes a customer to pay you back and then put you into profit, and that depends just as much on margin and lifetime value as it does on acquisition cost. A higher CAC at a healthier margin can hit the exact same payback window as a lower CAC at a thin one.

I've seen brands where CAC was genuinely lower at a much larger monthly spend than it had been at a fraction of it, because the things underneath it improved: conversion rate, the audience the creative was pulling, the team, the tooling. Scale didn't doom their CAC. The drift did, when it happened, and they caught it.

So before you decide a campaign has to die because CAC crossed some line, check two things together. One, where delivery actually went, using the checklist above. Two, whether the new CAC still clears your payback target given your real margins. Plenty of "kill it" moments are actually "steer it" moments wearing a scary number.

And if you genuinely are capped, if delivery is clean, the conversion data is intact, and the maths still doesn't work, then fine, that's a real ceiling and the answer is a different channel or a different market. But you've earned that conclusion instead of guessing at it. Most people reach for "the channel is tapped out" because it's the easy story, when the boring truth was a delivery report they never opened.

Where to from here

So next time the budget goes up and CAC jumps, sit on the off switch for an hour. Pull the age report, pull the placements, look at whether clicks and conversions have come apart, and ask the dull question first: where is this actually being delivered now?

What would you find if you opened that breakdown on your last big scale, a campaign that broke, or one that just got pointed at the wrong people?