Setting a cap that keeps push notification ads profitable past week one

Last updated: 8 September 2026

Push notification ads share one device with every other campaign currently targeting the same subscriber, and none of the competing buyers can see each other's send volume. A subscriber hit four times in a day by four unrelated advertisers experiences it as one advertiser sending four times, and the response curve for whichever campaign lands third or fourth collapses regardless of how good that creative is. Getting the cap right therefore matters more here than on almost any other paid format, and the mistake is rarely visible until a full billing cycle has passed.

Why a cap matters more on push notification ads than elsewhere

Email frequency capping protects a sender's own reputation inside one inbox provider's rules. Push notification ads capping protects something less visible: the subscriber's tolerance for the format itself, shared blindly across every network with access to that device. A publisher selling the same subscriber to three separate ad networks has no way to coordinate total daily volume across them, so the effective cap a subscriber experiences is the sum of every buyer's individual setting, not any single one of them.

That structural gap is why a buyer capping at once per day can still watch response decay on a subscriber who is, from the device's perspective, receiving six or seven notifications daily from other sources entirely outside that buyer's visibility.

The problem compounds on devices that opted into multiple publisher lists at once, which happens more often than most buyers assume since the same accidental-tap mechanism that builds one list tends to repeat across several sites a person visits in a single browsing session. A heavy torrent-index user can end up subscribed to five or six separate lists within a week without ever realising it, and every one of those lists counts that device as a fresh, uncapped subscriber.

No dashboard anywhere aggregates this cross-list exposure, because doing so would require every network to share subscriber data with competitors, which none of them have any incentive to do. The practical result is that a buyer's own cap setting is always an upper bound on that buyer's contribution to a subscriber's daily load, never a statement about the total load the device is actually carrying.

How response decays as push notification ads volume rises

CTR on this format drops fastest between the first and second send of a day, then continues declining at a slower rate through each additional message, following a curve closer to diminishing returns than to a hard cliff at any single number.

The gap between reported and real fatigue

A buyer's own dashboard shows fatigue only for sends that buyer controls, which understates the true decline because it cannot see the other advertisers also hitting the same list that day. Two campaigns with identical creative and identical caps can show meaningfully different CTR purely because one happens to be reaching a list with lower total daily volume from other buyers.

Comparing notes with other buyers on shared lists, or checking aggregate delivery guidance published on push ads documentation, gives a rougher but more honest picture of total exposure than any single account's dashboard can.

A workaround some agencies use is tracking response by hour rather than by day, since a sudden drop confined to a specific two-hour window often points to a competing sender active only in that slot, information that a daily aggregate would hide entirely inside an average that looks merely mediocre rather than sharply split.

Building this hourly view takes nothing more than tagging each send with a timestamp and pulling CTR against that field rather than against the daily total, a change most tracking platforms support without any additional cost, yet one that a surprising share of accounts never bother setting up until fatigue has already cost them several weeks of underperforming spend.

Cap settings for push notification ads and what they actually cost

Raising a cap increases delivered volume immediately and predictably, while the corresponding drop in conversion rate takes longer to show up in a blended report, since the first send of the day keeps performing normally and hides the decline in the additional sends layered on top of it.

The blended cost per action at a raised cap frequently looks acceptable for the first week or two, right up until the list's baseline response has permanently shifted and the campaign needs a fresh source to recover the numbers it started with. A comparison of cap settings against measured decay is laid out in more detail on push notification ads delivery pages, ahead of any single account's dashboard catching up to the same conclusion.

The safest default for a new campaign on an unfamiliar list is starting at the lowest cap setting and raising it only after a full week of stable baseline data, rather than starting aggressive and pulling back once numbers already look weak. Recovering from an aggressive start costs more time than starting conservatively ever costs in delayed volume.

Reading the unsubscribe wall before push notification ads hit it

Subscribers rarely unsubscribe through a formal opt-out link, since most panels do not surface one prominently; they revoke the browser permission instead, which is a silent action the sending network only detects on the next failed delivery attempt rather than in real time.

Signals that precede a wave of revocations

A sustained drop in delivery rate over several consecutive days, arriving before any visible CTR change, is the earliest available signal that a cohort is approaching mass revocation. Buyers who watch delivery rate as a leading indicator rather than waiting for CTR to confirm the damage typically catch the shift a week or two before it becomes visible in cost-per-action reporting.

The revocation wave itself is rarely gradual once it starts; a cohort that has tolerated daily sends for months can lose a meaningful share of active devices within a single bad week if a competing sender pushes an aggressive uncapped campaign against the same list, dragging every other advertiser's numbers down alongside their own.

Recovering from a revocation wave is slower than triggering one. A list that loses a third of its active devices in a bad week does not refill from new opt-ins at anywhere near that pace, since acquisition runs on its own separate timeline tied to publisher traffic rather than to how quickly a buyer would like the damage undone. Budgeting for that asymmetry, rather than assuming a bad week simply corrects itself the following month, avoids a second round of overspend chasing numbers that were never coming back that fast.

Setting a push notification ads cap that survives shared-list competition

Since total subscriber exposure is invisible and shared, the more reliable lever is not the cap itself but dayparting: concentrating sends into windows when competing buyers are least active, which recovers some of the response lost to invisible competition without needing any single network to change its own cap policy.

A dayparting approach that works around invisible competition

The delivery-window notes published on push-ads.io break down typical competing volume by hour for several major GEOs, which is useful groundwork before setting a dayparting schedule rather than guessing at windows from a single account's own limited view.

A cap on this format only protects what one buyer can see; the schedule built around it is what protects the rest, and the campaigns that hold a stable cost per action for a full quarter are almost always the ones treating dayparting as a standing decision rather than a launch-day setting nobody revisits.

None of this requires sophisticated tooling to start. A simple hourly breakdown pulled from existing tracking data, checked once a week against the delivery-window notes for the target GEO, catches most of the avoidable overlap before it compounds into a full month of underperforming spend, and it costs nothing beyond the time it takes to look.