Essay

The attention economy in 2026: a field report

9 min read · Updated March 2026

The average adult in 2026 spends 4 hours 41 minutes per day on a connected device, down from a peak of 5h 12min in late 2022 (Data.ai global average, Q4 2025). The decline is small but the first since smartphones existed. Total available attention is no longer growing. Every platform, every brand, every creator is now fighting over a fixed pie that is, if anything, slightly shrinking.

This essay looks at where that fixed pie sits in 2026: which platforms hold it, which formats inside those platforms still capture it, and where there's still leverage for brands and creators who don't have a $100M media budget.

Where the 4h 41min goes

The 2025 split, averaged across the global mobile internet user base:

Two things matter here. First, short-form video is now more of total attention than messaging — that wasn't true even in 2022. Second, web browsing is small (38 minutes) but disproportionately commercial: the average user is closer to making a purchase decision in a Safari tab than in a Reels feed.

What changed since 2022

Three big shifts in three years:

The "scroll session" got shorter

Average TikTok session length dropped from 11 minutes in 2022 to 7 minutes in 2025. Reels and Shorts followed. The platforms blame "competition for attention"; the real driver is mild platform fatigue and the fact that the algorithms have peaked — they cannot get sharper without showing you the same five videos repeatedly.

Messaging stopped growing

WhatsApp/iMessage/Telegram are flat in time-spent, despite continued user growth. Group chats are still active; one-on-one DMs declined ~9% YoY in 2025 (Sensor Tower). Plausibly: people are messaging less in messaging apps and more inside social/short-form (DMs in TikTok and Instagram are up).

The web came back, slightly

For the first time since 2017, time-on-web grew year over year (+4% in 2025). The driver: AI search and AI-curated reading apps (Arc Search, Perplexity, Comet, NotebookLM) are pulling people back into a tab-based reading experience. Small relative to social, but a meaningful directional shift.

What the platforms are doing about it

Every major platform's 2025 product roadmap chases the same thing: longer sessions per visit, because total visits per day cannot grow.

The pattern: keep the user in the app longer, give advertisers fewer levers, raise CPMs. Works for the platforms. Slowly squeezes brands.

Where attention still sits, undermarketed

Three pockets of attention are mispriced in 2026 (i.e. the ROI is high relative to what brands pay):

1. Mid-tier creators (10K–250K followers)

Engagement rates on creators in this band average 4.8% — five times the engagement of 1M+ accounts and twice the engagement of 1–10K accounts. Sponsorship pricing is still anchored to the (much lower) absolute follower count. Mid-tier creators are the highest-leverage paid distribution available in 2026.

2. Niche newsletters

The average operator-run newsletter (5K–50K subscribers) gets a 38% open rate. The same audience reached via Meta would cost $20–60 CPM and convert at 0.8%. Newsletter sponsorship at $40–120 CPM with 1.8% CTR mathematically wins by a factor of 3–5×.

3. Interactive web ad formats

WebGL, scrollytelling, pull-to-read. Total spend in this category in 2025 was under $400M globally — a rounding error against display's $200B. CPMs are 5–15× display, but cost-per-second-of-attention is 5–10× cheaper. The arbitrage is open. See 5 interactive formats that actually convert.

The strategic implication

Every dollar a brand spends in 2026 should be evaluated against one question: am I buying impressions, or am I buying seconds of attention? Impressions are a 30-year-old metric for a 2002 web. Seconds of attention — measured as dwell time, scroll depth, video watch-through, interaction rate — is what the modern audience actually trades to brands.

The marketing teams who internalised this in 2023 are the ones whose CACs are flat-to-down in 2026 while their competitors' CACs are up 30%. The only thing they did differently was switch the success metric. Everything else followed.

Related posts

Buy seconds of attention, not impressions

Pull-to-read ads average 8s dwell. Free pilot for indie brands.

START A PILOT