The OOB Social Attention Cost

Headline findings

• UK advertisers paid roughly 25p (24.5p) for every hour Britons spent on social media in 2025, about £207 per social media user identity per year. A market-average measure, not a price any single advertiser pays; carried through the survey-based inputs' uncertainty the honest range is the low-to-mid 20s of pence, which is why we quote it as an order-of-magnitude floor.
• UK social ad spend grew 21% in 2025 while the audience and its hours barely moved (about a 19-percentage-point gap; on 2025 inputs spend growth runs roughly 12 times audience growth, though the gap is the more robust statement). The implied 2024 cost is about 21p an hour (see "Implied 2024 figure" below).
• Companion metric, the Attention Value Gap. Two numbers travel with this release as a deliberate contrast, not two versions of one figure: 25p is the cost to advertisers of an hour of social time; £12.21 is that hour's notional worth at the wage floor. At the 2025 National Living Wage the pool is worth about £571bn a year; advertisers' entire annual spend against it was £11.5bn, about 2p in the pound.

Method

Inputs and sources

Implied 2024 figure (derived, not measured)

IAB UK's 2025 report publishes both the spend level (£11.5bn) and its growth (+21%), which implies a 2024 base of about £9.5bn (11.5 ÷ 1.21). DataReportal's Digital 2025 edition measured 54.8m UK identities on the same weekly-hours method as Digital 2026, giving a prior-year attention pool of about 46.2bn hours. Together:

That puts 2025's 24.5p roughly a fifth higher than 2024's implied 20.6p. We label the 2024 figure implied, not measured: IAB never published a standalone 2024 social spend line (the £9.5bn is back-calculated from IAB's own comparative growth figure), and we cannot verify from outside the report whether the +21% is computed on a fully like-for-like basis under IAB's new methodology partner. Quote it as "an implied ~21p" with that caveat attached.

Limitations (stated up front)

• Identities are not people. Platform ad tools over-count via multi-accounting. This cuts both ways: fewer real people raises per-person spend and shrinks the hour pool, nudging the per-hour price up. Headline per-user figures are therefore quoted per identity; the per-person figure is higher.
• The two factors have different population bases. The 55.5m identity count is an all-ages platform-derived total, while the 16h13m time figure is GWI survey data for internet users aged 16–64. Combining them assumes working-age usage intensity holds across all ages, a known approximation, and a further reason to read 25p as an order-of-magnitude floor rather than a point estimate.
• "Attention" here means time on social platforms. It is self-reported time-on-app, not eyes-on attention as measured by attention vendors (Lumen, TVision). Since not all time on an app is active attention, the true price per attentive hour is higher. Again, a floor. We'd welcome collaboration with measured-attention providers to layer attention quality on top of this market-level price.
• Spend and time definitions differ. IAB's "social" spend line and DataReportal's "social" time measure cover different platform sets (YouTube is the clearest case: much of its ad spend sits in IAB's video line). Treat 25p as an order-of-magnitude floor, not a precise unit price.
• Spend per hour is not a CPM. A 21% rise in spend against flat hours may reflect higher auction prices, more ads per hour, or both. The effect for brands is the same; we do not claim which mechanism dominates.
• The wage valuation is a yardstick. Scrolling is not labour and leisure time has no market wage; the NLW is a deliberately simple public benchmark (the statutory wage floor) chosen for legibility, not because leisure time is worth a wage (valuation studies typically value leisure below the wage). Read the Value Gap as an illustrative scale comparison, not a welfare estimate.
• No multi-year back-series. IAB published no standalone social spend figure before its 2025 report and DataReportal changed its time methodology between editions, so a verified 2023–2025 series cannot be built. 2026 is the declared baseline edition; the single 2024 figure above is derived and labelled implied, and any multi-year UK attention-cost series published elsewhere rests on estimated inputs.

Independent checking and triangulation

Every figure was independently recomputed before publication and all recomputations reproduced exactly. Every limitation above was surfaced by adversarial review and is disclosed rather than smoothed over. Reassuringly, an independent lineage lands in the same range: DataReportal's own UK social ad-spend estimate (~£9.8bn, a different methodology we do not blend with IAB's) against the same hour pool implies about 21p an hour, a different source arriving at the same order of magnitude.

A commitment on the growth multiple: "spend growing ~12× faster than the audience" is a ratio over a small denominator and we will not publish that multiple in editions where audience growth is below about two percentage points; the spend-vs-audience gap and the price change itself are the robust statements and will be reported every year.

Using the data

Licence: CC-BY 4.0 covering Our Own Brand's derived analysis (underlying IAB UK, DataReportal and gov.uk data remain subject to their own terms). Cite as "Our Own Brand analysis of IAB UK and DataReportal data, 2026", including a live link to ourownbrand.co/oob-social-media-attention-cost wherever the medium permits. Headline chart (SVG) and dataset (CSV) are downloadable from that page. Custom cuts by platform, sector or demographic available on request.