CAC Up 40% in Six Months? How to Isolate Creative vs. Landing Pages vs. Targeting

Read three metrics together before changing anything: if CTR is falling, suspect creative; if CTR is stable but conversion rate is falling, suspect landing pages; if CPM is rising while CTR holds, suspect targeting or auction pressure. Then confirm with one isolation test per layer — in that order — instead of changing everything at once.

Context first: rising CAC is partly structural. Customer acquisition costs have risen 40–60% since 2023 across B2B (Alexander Group, June 2026). So some of your increase is the market. The diagnostic below separates the part you can fix from the part everyone is paying.

Which metrics implicate which layer?

CAC is a composite. Decompose it before touching anything:

CAC = (CPM ÷ CTR ÷ CVR) × (1 ÷ close rate), roughly — cost to be seen, times how many see it click, times how many clicks convert, times how many conversions become customers.

That last row matters: in 2026, some "CAC increases" are attribution definition changes. Rule out measurement before re-engineering the funnel.

What is the correct sequence of isolation tests?

Run one variable at a time, cheapest and fastest first:

1. Creative holdout (1–2 weeks)

Freeze audience and budget; introduce 2–3 new concepts against your current control in the same ad sets. If new creative materially beats control on CTR at similar CPM, fatigue was the driver. Creative age analysis usually confirms — performance decay curves by launch date are the tell.

2. Landing page A/B (2–4 weeks, traffic-dependent)

Keep media untouched; split traffic between the current page and one meaningful variant (one hypothesis, not a redesign). If CVR recovers on the variant, the page was the constraint. This is standard CRO practice — the common finding is not that a page "got worse," but that traffic mix shifted toward a segment the page never served well.

3. Audience cohort analysis (no test needed — read existing data)

Cut CAC by audience cohort, geography, and placement over the six months. Rising CAC concentrated in one cohort means saturation there, not systemic failure. Spread evenly across cohorts usually means auction pressure — check whether CPMs rose category-wide, which you cannot fix with targeting.

4. Down-funnel validation (read your CRM)

If platform metrics all look stable but real CAC rose, the leak is post-click: lead-to-opportunity conversion, routing delays, or lead quality mix. This is where platform-only reporting hides the answer — you need CRM-connected attribution to see it.

Why do teams get this diagnosis wrong?

Because each vendor sees one layer. The ads agency sees CTR and defends creative; the web team sees CVR and defends the pages; nobody owns the composite. The diagnosis above only works if one party can see — and is accountable for — the whole chain. That was the eCommission situation: unified attribution plus media under one team produced +1,089% Google Ads ROAS and −74% cost per conversion, and a companion data and attribution rebuild (our client results). The gain came from finding which layer was actually broken — not from spending more.

When is rising CAC not fixable?

Honesty requires this section. If cohort analysis shows CPMs rising uniformly across your category, you are watching auction inflation — the 40–60% structural rise (Alexander Group, June 2026) — and the fix is not tactical. The levers become strategic: better monetization of existing traffic, retention economics, channel diversification, and brand strength that lowers your dependence on auctions. A vendor promising to "fix" structural CAC inflation with media tactics is selling optimism.