Ad Frequency Cap¶

Advertising delivery rule — instantiates Saturation Avoidance

Caps how many times a campaign shows the same person an ad, stopping at the point where extra impressions stop lifting response and start breeding fatigue.

Advertising has a response curve that first rises, then flattens, then bends down: the first few exposures build recognition, later ones add nothing, and past a point they irritate the viewer into actively avoiding the brand. Ad Frequency Cap is the domain-specific rule that stops delivery near the top of that curve — limiting impressions per person so spend lands where it still moves response instead of pouring into wear-out. What makes it this mechanism rather than the general Frequency Cap is that it is tuned to the advertising marginal-response curve specifically: it is set by watching incremental lift per impression and creative-fatigue signals, and it exists to protect campaign effectiveness and brand sentiment, not just to reduce message count. The cap is really a bet about where effective frequency ends and wear-out begins.^[1]

Example¶

A retailer runs a two-week programmatic video campaign. Without a cap, the auction cheerfully re-serves the same twenty-second spot to the same in-market shopper eleven, twelve, thirteen times, because impressions are cheap and the pixel keeps finding them. Ad Frequency Cap watches the marginal signal: measured lift per additional impression climbs through exposures one to three, is flat from four to six, and by exposure eight the brand-favorability survey and the rising "hide this ad" rate say the extra views are now negative. The team sets the cap at roughly six impressions per viewer per week.

The budget that used to buy exposures seven-through-thirteen for saturated viewers is reallocated to reaching fresh ones. Same spend, but it now buys reach where response still exists instead of hammering people already past the point of caring — and the campaign's irritation metrics come back down.

How it works¶

Read the incremental-lift curve. Estimate response per additional impression (via holdout or incrementality testing) rather than total response, so the cap targets where the marginal impression stops paying.
Watch fatigue signals, not just count. Rising ad-hide / skip / negative-sentiment rates are the saturation signal that the curve has bent; the cap responds to those, not to a fixed folklore number.
Cap per viewer, redeploy the surplus. Enforce an impressions-per-person ceiling and route the freed budget toward unsaturated reach.
Refresh, don't just cap. Because wear-out is partly creative, rotating the creative resets the curve — an alternative to lowering the cap.

Tuning parameters¶

Impression ceiling and window — impressions per viewer per period. Too low leaves cheap effective frequency unbought; too high spends into wear-out and irritation.
Fatigue-signal sensitivity — how fast a rising hide/skip rate pulls the cap down. Sensitive caps protect sentiment but may cut a still-working campaign short.
Reach-vs-frequency split — whether freed budget buys new reach or richer frequency on responsive segments. Governs whether the campaign broadens or deepens.
Creative-rotation coupling — whether hitting the cap triggers a creative refresh instead of a hard stop, resetting the wear-out curve rather than ending exposure.

When it helps, and when it misleads¶

Its strength is spending discipline grounded in the ad-response curve: it keeps budget in the zone where impressions still lift response and out of the zone where they damage the brand, and it makes wear-out — normally invisible until sentiment craters — a managed number.

It misleads when the cap is set by convention rather than measurement ("three is the magic number") divorced from the actual lift curve, which varies wildly by product, creative, and audience. It is also gamed by fragmentation: caps enforced per-device or per-exchange let the same human be saturated across a dozen uncapped surfaces, so the reported frequency looks moderate while the person is buried. And a cap tuned only on click response can miss brand-level wear-out that shows up in sentiment, not clicks. The guard is to set the cap from measured incremental lift and fatigue signals, resolved to the person across surfaces, and to treat creative rotation as a first-class alternative to simply capping harder.

How it implements the components¶

Ad Frequency Cap fills the archetype's response-curve-aware rationing components, specialized to advertising:

input_reduction_rule — it caps impressions per viewer, deferring or dropping delivery past the ceiling.
marginal_response_metric — its ceiling is set by incremental lift per impression: the point where one more ad stops producing useful additional response.
saturation_signal — it reads ad-fatigue indicators (hide, skip, negative sentiment) as the sign the response curve has bent down.

It does not model total channel capacity or a cross-sender recovery budget — the general per-recipient rule is Frequency Cap's job — and it does not prioritize among competing messages (Queue Admission Limit) or reroute demand (Alternate Pathway Routing).

Instantiates: Saturation Avoidance — protects advertising response by stopping delivery near the top of the exposure curve.
Consumes: Frequency Cap — the general per-recipient capping rule this specializes for advertising.
Sibling mechanisms: Frequency Cap · Attention Cap Management · Saturation Dashboard · Channel Capacity Management · Queue Admission Limit · Worker Caseload Limit · Alternate Pathway Routing · Overflow Queue · Capacity Expansion Trigger · Graceful Degradation Mode

References¶

[1] Effective frequency is the exposure range where an ad still builds response; wear-out is the point past which additional exposures stop working and can turn viewers against the brand. An ad frequency cap is an operational bet on where the first ends and the second begins. ↩