Myth: Your DSP frequency cap reliably limits user exposure

Myth: Your DSP frequency cap reliably limits user exposure

Buyers set a frequency cap — say 3 impressions per user per day — and trust the user genuinely saw the ad three times. The cap is treated as a hard physical limit.

The cap is only as good as identity resolution, which is leaky by construction.

1. The cap keys off a user identifier — a cookie, mobile ad ID, or a graph-resolved ID. Every cap is really "3 per identifier per day," not per human.

2. The same person appears under many identifiers: different browsers, cleared cookies, a phone plus a laptop, Safari's restrictions truncating cookie lifespan. Each fragment carries its own fresh cap.

3. Across supply paths, the same impression opportunity may arrive with different SSP-side user IDs. Without a stable cross-path identity, your cap counts each path's view separately or fails to dedupe wins.

4. In cookieless environments the cap may fall back to IP or contextual signals, which over- or under-count wildly — a shared office IP caps a whole floor as one user.

The realistic read: a 3/day cap on a fragmented identity can deliver 8-10 actual exposures to a heavy multi-device user, and far fewer to a privacy-hardened one.

Why it matters: frequency capping limits exposures per resolvable ID, not per person. Identity fragmentation means your real-world frequency distribution has a long, uncapped tail you should measure, not assume away.