What happens between a subscription and a push ads impression

How a browser turns a click into a push ads subscriber

Every subscription starts with a native browser prompt, triggered by JavaScript on a publisher's page and rendered by the operating system itself, not by the site. Once granted, the permission is stored against the origin, and the site receives an endpoint URL it can call at any time, from any server, without the visitor being present. That endpoint is the entire mechanism behind a push ads send: no cookie, no pixel, no re-targeting tag, only a subscription record tied to a domain and a device.

Chrome and Firefox render the request as a small dialog anchored to the address bar, while Safari on macOS routes the same decision through the system Notification Center and expects a distinct web app manifest plus a registered service worker before any prompt appears at all. Android's build of Chrome additionally ties the permission to the device's own notification settings rather than to the browser alone, so a subscriber who mutes notifications system-wide can stay technically subscribed while never seeing another impression.

The prompt itself is a single decision point, and most publishers only get one attempt before a browser suppresses repeat requests on the same domain. A soft prompt placed inside the page content, asking for interest before the native dialog appears, raises opt-in rates measurably because it filters out visitors who would otherwise dismiss the hard prompt out of habit. A domain that burns its one native attempt on an uninterested visitor rarely gets a second chance with that browser profile.

Where subscribers actually come from

Three sources dominate subscriber growth in practice: content publishers monetising their own traffic, redirect and URL-shortener chains that route a visitor through a subscription page before the destination loads, and download portals where a prompt sits ahead of the file. Each source produces a different profile, and a push ad network segments inventory by source precisely because a redirect-sourced list clicks differently from one built on a technology blog.

What an advertiser is actually paying for in a push ads campaign

A push notification creative carries less than a display banner: an icon, a short title, one line of body text and an optional larger image, all rendered by the operating system's own tray rather than by a browser page. That limit forces the copy to do the entire persuasion job in roughly ninety characters, which is why creative testing in a push ads programme moves faster than in formats that tolerate a paragraph of supporting text.

A closer breakdown of how the tray renders each field on different platforms sits on the push notification ads page, since a title that fits on Windows can still be cut on an older Android build.

Delivery itself is near-instant once a send is triggered, arriving whether or not the subscriber has a browser tab open, which is the property advertisers are actually buying. Frequency capping sits on top of that reach: without a cap, a subscriber list burns out inside days as the same segment receives competing offers from every buyer active on the network, so serious campaigns cap sends per subscriber per day and rotate creative to slow the decay.

Pricing models that decide who carries the risk in push ads

Cost per click remains the default entry point in push ads because it lets an advertiser pay only for engagement rather than for raw delivery, and it matches how the format is actually consumed: a notification either gets tapped within its first minutes or it never does. Cost per mille shifts the risk the other way, rewarding networks with large verified subscriber pools and punishing advertisers who cannot judge audience quality in advance, which is why CPM buys concentrate among larger accounts running heavy volume.

Reputable networks run their lists through duplicate-detection and device-fingerprint checks before selling access to them, because a list padded with emulator traffic or recycled device identifiers produces clicks that never convert and eventually gets a buyer's own domain flagged by ad-quality vendors. Advertisers who skip this check usually learn about it only after a payout dispute, which is why experienced buyers ask for a sample delivery log before committing a full month of budget.

CPC against CPM in practice

A campaign manager choosing between the two is really choosing who absorbs a weak list: on CPC, a network with dead subscribers simply generates fewer clicks and the advertiser pays the same rate for less delivery, while on CPM the advertiser pays for every send regardless of whether the subscriber base still exists. Hybrid models with a bid floor plus a CPC ceiling have grown for exactly this reason, splitting the exposure between both sides of the trade.

Consent rules that shape every push ads campaign in the EU

The ePrivacy Directive treats a browser subscription the same way it treats a cookie: as information stored on a terminal device that requires informed, freely given consent before it is set. A single native browser prompt satisfies the technical consent step, but it does not on its own satisfy the information duty under the GDPR, which is why compliant publishers disclose what the subscription will be used for in the soft prompt that precedes it, not only in a buried privacy policy.

A closer look at how regulators judge that disclosure, including the wording several data protection authorities have accepted, is set out on the push notification ads reference page.

The double opt-in question

Regulators in several EU member states have pushed publishers toward a second confirmation step after the browser permission, arguing that a single click on an ambiguous banner does not meet the bar for specific consent when the subscription will later carry third-party advertising rather than the publisher's own content. Networks buying EU traffic increasingly require proof of this second step from their publisher partners before accepting a subscriber list, since a regulator complaint traces back to whichever party collected the consent.

Withdrawal has to be as simple as the original grant, which in practice means the unsubscribe path sits inside the notification itself, not three menus deep in browser settings, because a subscriber who has to hunt for the exit tends to block the entire domain instead of opting out cleanly, taking every future send on that domain down with it.

Where push ads earn a place next to display and native buys

Display media relies on a page being open at the moment the ad loads, and native placements depend on a feed being scrolled; a push send depends on neither, which makes push ads the format that reaches a device during moments no other channel can reach, such as a phone screen lighting up hours after the browser was last used. Media planners rarely treat it as a replacement for either format; instead it fills a re-engagement slot that display and native structurally cannot cover, aimed at a list the advertiser already partly owns rather than a fresh audience.

This comparison first grew out of research compiled for Dijon Japonais, an editorial project that documents public disclosure practices for a restaurant in eastern France, before the same attention to what a prompt actually discloses was turned toward browser permission requests in advertising.

Reading the numbers a network reports

A subscriber count on its own says little, since a list built two years ago with no cleaning reports the same headline size as one refreshed last month while performing at a fraction of the rate. The more useful figures are recent delivery-to-click ratios by geography and by device, because those numbers reflect the list as it behaves today rather than as it was acquired, and a full breakdown of current benchmarks by country is documented on push-ads.io, which is worth checking before a first test budget is committed to any push ads plan.