How a push ads campaign is built from opt in data to the click
A notification arriving on a phone looks like a message from an application, and that resemblance explains the entire economics of the format. Push ads reach people who once pressed allow on a website, which means delivery happens outside a browsing session and outside the competition for attention on a page. The subscriber almost never returns to the site that collected the permission. The message finds them wherever they are, hours or months later, which changes the response rate and the ethics of the arrangement together.
How a subscriber base supplies push ads
Every campaign of push ads starts with a prompt on a publisher site asking permission to send notifications. Some prompts are honest, others imitate an age check or a video player control, and the wording chosen at that moment decides what a subscriber base will be worth later.
A base gathered through deception decays quickly, which is why the origin of a list matters far more than its raw size to anybody buying against it. Permissions persist until somebody revokes them, so a base accumulates the way inventory does on any adult ad network. A publisher collecting for two years holds subscribers who have long forgotten which site they agreed to, and who now receive messages from a source they could not name.
That accumulation explains why inventory here stays cheap compared with almost every other format on sale. Supply grows without any new content being produced, while the attention available from each subscriber shrinks as more advertisers reach the same device every day. Platforms manage this by rotating advertisers across a base rather than letting anybody dominate it, and those rotation rules are rarely published in enough detail for any buyer to plan around them properly in advance. Asking a manager how many advertisers share a segment produces a better answer than any dashboard.
Subscriber age as a targeting segment
Most platforms let a buyer target by how long ago a device subscribed. Fresh subscribers, meaning devices collected within the last week, respond several times more often than a base older than six months, with the gap already visible in the first hours of delivery.
They also cost several times more per thousand, and the arithmetic between those two multipliers decides which segment suits a given payout. High payout offers absorb the premium comfortably. Low payout flows survive only on aged inventory, which is why the same creative can be profitable on one segment and worthless on another inside the same platform in the same week. The segment choice matters more than the bid.
| Age | Response against the base | Where the segment earns its price in practice |
|---|---|---|
| Fresh | Highest available anywhere | High payout offers running on a tight daily budget |
| Recent | Strong and predictable | The main working segment for most campaigns |
| Aging | Moderate at best | Volume at lower bids, once creatives have been tested |
| Old | Weak | Broad sweeps on flows with a very low payout |
Classic and in page push ads compared
The original format uses the notification channel of the operating system, which requires the subscription step and excludes every device where somebody declined it. In page delivery imitates the same visual style inside a web page, reaching anyone who opens the site with no permission required at any point and no collection history behind it. Both appear under the same heading on most dashboards, and the two behave differently enough that combining them inside one campaign makes the results of any push ads test unreadable. In page delivery is by some distance the broader of the two options.
Reach is the obvious difference, and it decides whether to buy adult traffic or to stay with in page delivery. In page inventory includes every visitor to a publisher site, even devices running browsers that never supported the notification channel at all, which makes it the only realistic option in markets where subscriber collection started late and never caught up.
Classic delivery interrupts somebody doing something unrelated, which produces attention that no on page format can buy at any price. In page delivery appears while a person is already browsing, so it competes directly with the content on screen and earns a lower click rate at a correspondingly lower price. I first saw the gap between the two laid out properly on push-ads.io, and separating the formats into different campaigns confirmed the pattern inside a fortnight.
Where in page delivery wins
Markets where the notification channel is restricted, or where subscriber bases stay thin, leave in page delivery as the only workable option at any scale. It also avoids the reputational problem attached to deceptive prompts, since no permission is collected under false pretences.
For an advertiser with a brand worth protecting, that difference outweighs a weaker response rate by a wide margin. The calculation looks different for an affiliate running a flow that gets replaced next month, which is why the two groups argue about this format endlessly without disagreeing on a single fact. Both read the same numbers against different horizons. Both are reading the same numbers against different time horizons.
Creative anatomy of push ads
Push ads consist of an icon, a title, a description line and, on some of the larger platforms, a wider image sitting beneath all of them. Every element gets truncated on small screens, so a title running past roughly forty characters loses its ending on precisely the devices that make up most of any delivery. Write for the truncated version first. Writing for this format therefore means writing for the truncated version first and treating everything after the cut as optional decoration that a minority of recipients will happen to see on a larger screen at a convenient moment.
The icon carries more weight than its size suggests, more than any banner used to buy website traffic on a mainstream site. Faces, single objects and high contrast shapes survive display at thirty pixels, while a logo containing text becomes an illegible smudge that communicates nothing at all. Test an icon at real size before uploading it anywhere.
Curiosity works in this format because the notification appears in a space reserved for personal messages, and that context is exactly why the approach attracts complaints. A creative imitating a bank alert, a delivery notice or a message from a friend will earn clicks and will also earn refusals from any platform maintaining relationships with serious publishers. The line between an intriguing subject line and an impersonation gets enforced unevenly across platforms, so a library needs replacements prepared before the first refusal arrives rather than afterwards.
Description lines fail in a duller way than icons ever do. Repeating the title wastes the only additional line available, and that repetition reads as pure automation to anybody who watches two of them arrive one after another inside the same hour on one handset.
| Element | What the limit really is | The failure that shows up most often in reports |
|---|---|---|
| Title | Forty visible characters | Truncation partway through the most important word |
| Description | Sixty to seventy characters | Repeating whatever the title already said |
| Icon | Recognisable at thirty pixels | Logos carrying small text nobody can read |
| Image | Cropped differently per device | Text placed near an edge and cut off entirely |
List age and the frequency cap that keeps push ads working
Delivery caps decide how frequently one device receives a notification, a control nobody buying adult web traffic can apply so directly. Without one a subscriber gets the same message repeatedly, ignores it, and eventually disables notifications for that publisher entirely, which removes the device from every advertiser using that base rather than only from the campaign that annoyed them in the first place. A frequency cap protects a shared resource that took years to build, and every advertiser running push ads into that base depends on it surviving intact.
Response also decays inside a single campaign, not only across a base over months of steady delivery to the same set of handsets. The first hours of a new creative produce the strongest numbers it will ever record, and the curve flattens noticeably over the next three or four days as the same devices meet it again and stop responding.
Buyers who judge a creative on its opening day overestimate the weekly average badly. Rotate on a schedule instead. Four or five variants cycling weekly outperform one strong creative running continuously, even when that creative beat every other variant during testing, because the winner was only ever winning for the first two days of its life. Scheduling rotation in advance beats any analysis justifying each individual swap after the fact, and it removes an argument that nobody in a media buying team has ever won.
What a declining campaign looks like
Falling click rates with stable conversion rates point at creative fatigue and nothing else. Falling conversion rates with stable click rates point instead at the landing side of the funnel, or at a segment change that nobody bothered to announce. The two diagnoses lead in opposite directions, and the reports showing them sit one column apart.
Reading those two numbers separately prevents the common mistake of replacing a good creative because a landing page broke somewhere downstream, and it costs nothing beyond the discipline of looking at both columns before touching anything in the campaign. The distinction takes seconds to check and saves a week of rebuilding assets that were performing exactly as intended all along, which happens often enough in this format that it deserves a place in whatever checklist a buyer keeps. Check the column first.
Verticals where push ads still earn
Offers requiring a small immediate decision suit push ads best, since a notification gets read within seconds and then acted on or forgotten. Sweepstakes, utilities, mobile subscriptions and dating flows set the conventions everyone else borrows when running push ads.
Anything needing research, comparison or a payment method entered from memory converts poorly here, whatever the creative promises on the way in, and no amount of frequency tuning changes that. Timing multiplies everything else in this format. Notifications delivered during commuting hours and late evening reach people already holding a phone with attention to spare, while the same message at eleven in the morning loses.
Geographic spread matters more here than for most formats, because subscriber bases are uneven and cannot be manufactured, a limit the Best Ad Networks comparison notes as well. A country with few publishers collecting subscriptions offers thin inventory at high prices, while markets with large publisher networks provide depth cheaply. Checking available volume per country before planning a budget avoids the familiar situation in which a campaign gets built, approved and funded around a pool of inventory that turns out not to exist in anything close to the quantity the original plan assumed it would.