whatiscpm.com

Benchmarks

What is a good CPM?

By Oliver Wakefield-Smith · updated October 2026

Benchmarks reviewed October 2026 · next review January 2027

The short answer

There is no universal good CPM. There is only your CPM against your channel's typical level for your region, objective and month. The dated reference points we can cite: Meta averaged $10.02 in December 2025 and $8.06 in June 2025, and TikTok $2.84 in December 2025 (Gupta Media's public tracker table); podcast 30-second host-reads average $18 to $22 per 1,000 downloads (Libsyn Ads).

Good relative to what

Four comparisons matter, in order: your channel's typical level (the same CPM can be high on one platform and ordinary on another), your month (Meta's tracker average moved from $8.06 in June to $10.02 in December 2025), your region (blended averages hide country mix, so a US-only audience is not comparable to a blended figure), and your objective (conversion-optimized campaigns chase a smaller, more contested slice of the auction and pay more per impression by design).

Meta, Dec 2025$10.02Meta (Facebook and Instagram combined) average CPM, December 2025, the latest month in the public tableGupta Media Social CPM Tracker, public monthly tableretrieved October 2026
Meta, Jun 2025$8.06Meta (Facebook and Instagram combined) average CPM, June 2025, same tableGupta Media Social CPM Tracker, public monthly tableretrieved October 2026
TikTok, Dec 2025$2.84TikTok average CPM, December 2025, the latest month in the public tableGupta Media Social CPM Tracker, public monthly tableretrieved October 2026
LinkedInno published rateLinkedIn describes auction pricing and budget controls; it publishes no average CPMLinkedIn ads pricing pageretrieved July 2026
Podcast 30s$18-22Podcast 30-second baked-in host-read ad, overall average CPM range per 1,000 downloads (tools use the $18 low end)Libsyn Ads (formerly AdvertiseCast) podcast advertising ratesretrieved October 2026

The verdict logic here is the same as the calculator's DeltaPip: more than 5% below the loaded preset reads as below, within 5% reads as at, more than 5% above reads as above. Run yours through the calculator with a preset loaded and the verdict renders automatically.

Why objective changes the answer

An awareness campaign buys the cheapest eligible impressions; a conversion campaign buys impressions Meta predicts will convert, which are scarcer. Meta's own auction documentation describes price as bid times estimated action rate times quality: optimize for action and you are bidding into a thinner pool. A higher CPM with a lower cost per result is a win, not a problem. The funnel math is worked at CPM vs CPA.

When a high CPM is fine

  • Narrow B2B audiences.A LinkedIn CPM several times Meta's can still be cheap per qualified decision-maker reached; the break-even is worked at LinkedIn vs Facebook.
  • High-LTV niches. If a customer is worth $5,000, the difference between a $10 and $25 CPM disappears three decimal places into your unit economics.
  • Premium formats. CTV and podcast host-reads price attention, not just delivery. A podcast 30s at $18 to $22 per 1,000 downloads buys a listener mid-episode, not a pixel at the bottom of a page.

At Meta's December 2025 average, $1,000 buys about 99,800 impressions; at TikTok's, about 352,113. Whether the bigger number is better depends entirely on who those people are and what they do next.

If your number is above median and you did not choose that on purpose, start at why is my CPM so high, then how to lower CPM. Full channel table: average CPM rates.