CPM (cost per thousand impressions)

CPM is the cost of one thousand impressions, and it is the native pricing unit of social and display auctions. On Meta you are buying attention and converting it into outcomes through creative, which makes CPM the cost of the raw material rather than a measure of how well you used it.

The chain from CPM to CPA is short enough to compute in your head, and worth computing because it shows where results actually come from. A $25 CPM at a 1.5% click-through rate produces a $1.67 CPC; at a 2.5% conversion rate that is a $66.67 CPA. Now hold CPM fixed and lift CTR to 2.2%: CPC falls to $1.14 and CPA to $45.45, a 32% improvement with no change in what you paid for the impressions. Better creative did not buy cheaper media. It extracted more from the same media.

Run the comparison the other way to see why CPM is a poor target. Cut CPM from $25 to $18 by moving into cheaper placements, and suppose CTR drops from 1.5% to 0.9%: CPC rises to $2.00 and CPA to $80. The media got 28% cheaper and the customer got 20% more expensive. Media buyers who optimize CPM in isolation reliably end up cheap and invisible.

Where CPM earns its place is diagnosis. Read it against CTR and frequency and it tells you which of three different problems you have. Rising CPM with stable CTR usually means auction conditions changed — more competition, a seasonal peak, or a narrower audience definition. Stable CPM with falling CTR is creative fatigue: the same people are seeing the same work and responding less. Rising CPM with rising frequency means the audience is too small for the budget, and the platform is paying more to show the same people additional impressions.

Those three readings call for different actions — wait or rebid, refresh creative, broaden targeting — and the CPM number alone cannot distinguish them, which is the argument for never reading it alone.

Two practical notes. CPM varies enormously by placement, country, and season, so the useful comparison is an account against its own recent history rather than against a published benchmark. And treat the fourth quarter as its own regime: retail bidding lifts CPMs across the board, so a stable CTR at a higher CPM in November is not a performance problem, it is the calendar. Comparing November to October and concluding the creative broke is one of the most reliable ways to retire work that was still performing.