Why your paid Meta CAC is rising — and what we do about it

If you've run paid Meta for more than two years, you've probably noticed: the same audience that cost $14 to acquire in 2023 is closer to $26 today. The creative is better. The targeting is more sophisticated. The platform is more capable. And yet CAC keeps climbing.

This isn't because you're worse at running ads. It's because three structural things have shifted, and most playbooks haven't caught up.

1. The pixel signal is degraded — and it's not coming back

ATT happened in 2021. The slow degradation has been happening ever since. The pixel that fired reliably for a decade now fires for maybe 40–60% of mobile iOS conversions, depending on your category. Meta's machine learning makes up for some of this, but its bids are calibrated on signal it can't see.

What we do: server-side conversion API on every account, day one. Combined with first-party email data piped via CAPI Gateway or Stape, we typically recover 18–28% of attributed conversions that the pixel alone is missing. This isn't a one-time fix — the signal degradation continues, and CAPI hygiene needs ongoing maintenance.

2. Audience saturation is faster than it used to be

In 2018, a Lookalike 1% audience in a mid-sized market would scale to $50k/mo before fatigue. In 2025, that ceiling is closer to $15–20k for most categories. The platform has gotten better at finding intent, which means it converges on the most likely buyers faster — and then runs out of them faster.

What we do: layered audience strategy with explicit refresh cadence. We don't run a single Lookalike for more than 8 weeks without replacing it. We rebuild from post-purchase survey data every quarter — not from the same pixel-derived seed audience.

3. Creative volume requirements have 3× in three years

In 2022, a healthy paid Meta account refreshed creative every 4–6 weeks. In 2025, we see creative fatigue inside 7–14 days on scaled audiences. This is a function of Meta's bidding algorithm getting better at exposure: it finds your relevant audience faster, exposes them to the creative faster, and then exhaustion sets in faster.

What we do: hypothesis-led creative production at much higher volume — typically 10–14 net-new creatives per month for any scaling account. We use a structured creative testing framework where every new creative is tied to a written hypothesis (e.g., "social proof at the top of the funnel will outperform feature-led creative for this audience"). Anything that doesn't beat baseline inside 7 days gets cut.

What this means for your account

If you're seeing rising CAC, the diagnostic order matters:

1. First — check your conversion signal quality. Is CAPI live? Are you sending enriched data (email, name, IP) to recover modeled conversions? If not, fix this before anything else.
2. Second — audit your audience refresh cadence. If you're running the same Lookalike for 12+ weeks, that's likely a primary driver.
3. Third — count your creative production volume. If you're producing fewer than 8 net-new creatives per month at scale, you're under-fueling the algorithm.

Most "rising CAC" problems are one of the three above — and most are fixable inside 30 days of structured work. If you want a written audit of where your account sits against these three benchmarks, open an intake and we'll come back with a specific report inside one business day.