This field record asks whether marketers pay more to deliver video ads than image ads in live Meta auctions (CPM), and whether that premium also appears in cost per click (CPC). The short name used in the repository directory is the video vs image CPM and CPC study.
Sample
The asset set is dataset EMR-2026-07B: 4,952 Meta ads with measurable spend in June 2026 (1,199 video; 3,753 image), drawn from a random sample of advertiser accounts that each spent at least $100,000 in that month. Clients sampled of various industries. Metrics are spend-weighted across the pooled sample.
Contrast
Two creative classes are compared. A video ad is classified when the ad creative carries a video identifier. An image ad is classified when the creative carries an image hash or photo/link object type without a video identifier. Outcomes are spend-weighted CPM (spend ÷ impressions × 1,000) and spend-weighted CPC (spend ÷ link clicks).
Indexed to image ads at 100, video ads land at 159 on CPM and 85 on CPC. That is a 59% higher delivery cost for video and a 16% lower cost per click for video. These are relative differences on the cost indices, not percentage-point changes on a fixed 0–100 rate.
Findings
In this June 2026 Meta window, video creatives clear the auction at a meaningfully higher CPM than image creatives. The same pooled spend shows a lower CPC for video, so the delivery premium does not appear as a click-cost premium in this observational contrast.
Video costs more to show, but less per click
Spend-weighted cost per 1,000 impressions and per link click, Meta auctions, June 2026
Source: Darwin IQ Labs, dataset EMR-2026-07B · 4,952 Meta ads (1,199 video, 3,753 image) · Clients sampled of various industries
What $1,000 bought in the same auctions
Implied from the spend-weighted CPM and CPC above
Source: Darwin IQ Labs, dataset EMR-2026-07B · Figures are $1,000 ÷ CPM × 1,000 and $1,000 ÷ CPC
What this record does not claim
- No creative-level file, account list, or brand identifiers are released with EMR-2026-07B.
- No confidence interval or client-level subgroup split is part of this release.
- The comparison is an observational field contrast in live Meta auctions, reported as observed. It is not presented as a laboratory randomization or as a causal estimate of format on auction price after controlling for objective, placement, or audience.