
Straight answer: Meta CPMs rose about 20% year over year in 2026, from an $11.82 average to $14.19, and cost per result climbed even faster for many brands. You cannot reverse the auction price, but you can stop paying for it. The fix is a combination of creative testing, audience discipline, and funnel-level diagnosis. I have run $2.7M+ in client ad spend, and in this guide I will show you the exact system I use when a client's costs start climbing.
Real example from a client account I manage: spend went from £61.9K in March to £77.2K in the last 30 days, but purchases stayed flat at ~7,000. CPA climbed from £8.75 to £10.90. CPM actually fell slightly. The real damage was add-to-cart rate dropping 28% and CTR dropping 13%. The auction was not the problem. The account had a conversion problem at the product page level and a creative problem at the ad level.
First, the data: what is actually happening in 2026
- Average Meta CPM rose ~20% YoY, from $11.82 to $14.19 across all industries
- Cost per lead climbed 21% YoY in 2025, and the trend continued into 2026
- Meta's ad revenue grew 24% YoY while impressions grew only 6%. More money chasing the same eyeballs.
- Ecommerce CPMs spiked up to 66% during the 2025 holiday period as competition peaked
- DTC advertisers shifted spend from Google to Meta: Meta share of DTC ad spend grew from 52% to 63%
The two problems are not the same
- The market got more expensive: your CTR, CVR, and CPA are stable, but CPM went up. The fix is efficiency: creative that earns a cheaper click, and a funnel that converts more of it.
- Your account got worse: CTR dropped, CVR dropped, or frequency climbed. The fix is diagnosis and rebuilding.
The diagnosis: 5 numbers that tell you everything
| Metric | If it moved... | The problem is... |
|---|---|---|
| CPM | Up, everything else flat | Market pressure, need efficiency |
| CTR | Down 15%+ | Creative fatigue or audience saturation |
| Frequency | Above 2.5 on an adset | Same people seeing the ad too often |
| CVR | Down, CTR stable | Landing page or funnel, not the ads |
| CPA | Up, all of the above | Combination, fix the root cause first |
Fix 1: Creative testing with a system, not vibes
- Always have tests running: 3-5 new creative angles in flight at all times.
- Judge creatives on a fair sample: at least $80-100 of spend or ~10 conversions before judging.
- Judge at the ad level, not the adset level: adsets blend multiple creatives and hide the real story.
- Kill on defined thresholds, not feelings: for a £10 CPA account, a creative at £19 CPA after a fair sample is done.
- Watch the hook, not just the click: if less than ~30% of impressions watch to 25% of the video, the hook is dead.