Skip to content

Push notification sponsorship pricing has to price the annoyance as well

Exhibitor analyticsUpdated 2026-08-237 min read

In short

Push notification sponsorship pricing sets a price per send against the users who have notifications enabled, then caps the number of sponsored sends per edition. A 5,000 dollar push delivered to 3,010 enabled users is a CPM of 1,661. The binding constraint is opt-out, which shrinks every later send.

A sponsor asks for a push on the Wednesday morning. Sales says yes, because it is five thousand dollars and the app can send it in ninety seconds. Three other sponsors have already been promised one, the marketing team has two of their own booked, and operations needs the channel free on Thursday for a hall changeover.

Nobody is tracking the total. By the end of the show the app has sent eleven notifications in three days and the number of people who will receive the twelfth is smaller than anybody in the building realises.

Push notification sponsorship pricing is the one line on the card where selling more units destroys the asset, and the destruction is invisible until you go looking for it.

Why the supply is short

The constraint is a switch the attendee controls, and both platforms have made it harder to hold.

Apple's App Store Review Guidelines set the terms directly at 4.5.4: "Push Notifications should not be used for promotions or direct marketing purposes unless customers have explicitly opted in to receive them via consent language displayed in your app's UI, and you provide a method in your app for a user to opt out from receiving such messages." A sponsored notification is promotional traffic, so an app carrying them is required to have obtained consent for that purpose and to offer a way out.

On Android the permission itself became a prompt. Google's Android developer documentation on the notification runtime permission, in its revision of August 2026, records that Android 13, API level 33, introduced a runtime permission called POST_NOTIFICATIONS, and that an app must wait to send notifications until it has requested that permission and the user has granted it. Deny it and all notification channels are blocked apart from a few specific roles.

So the reachable base is people who installed the app, saw a prompt, said yes, and have not since changed their mind. Every sponsored send is a fresh chance for the last part of that sentence to change, and the switch they flip is the same switch your operational messages travel through.

The price of one send

Take a 14,000 attendee show with 4,300 app installs. Of those, 3,010 have notifications enabled, which is 70.0 per cent of installs and 21.5 per cent of attendance.

Price a sponsored push at 5,000 dollars. The CPM is 5,000 divided by 3.010, which is 1,661.13 dollars per thousand delivered notifications.

That is an enormous CPM by any comparison, and it should be. A push is the only unit on the card that interrupts. It arrives on a lock screen, it carries the show's name rather than the sponsor's, and it borrows credibility from a channel the attendee accepted for logistics. A banner competes for attention. A push takes it.

Tap-through is where the case has to be made, because the delivered count alone flatters the unit. Suppose 412 people tap, which is 13.7 per cent of delivered. The cost per tap is 5,000 over 412, or 12.14 dollars, and now the sponsor has a figure they can put next to a paid click. Report both, and let the sponsor decide which one they are buying.

What does an opt-out actually cost?

This is the arithmetic almost nobody does, and it is the reason the cap exists.

Instrument the app to record notification permission changes and attribute them to the sixty minutes after each send. Suppose across one edition the pattern comes back like this. The first sponsored push goes to all 3,010 enabled users and 21 of them disable notifications, a rate of 0.7 per cent. The second goes to 2,989 and loses 36, at 1.2 per cent. The third goes to 2,953 and loses 62, at 2.1 per cent. The fourth goes to 2,891 and loses 98, at 3.4 per cent.

Total attrition is 217 users, 7.2 per cent of the base, and the rate roughly quintuples from the first send to the fourth. Revenue across the four is 20,000 dollars.

Now value what you destroyed. At 1.66 dollars per delivered notification, those 217 lost users represent 868 undeliverable notifications next edition if you sell four again, worth about 1,442 dollars. Set that against 20,000 dollars earned and the cash case for a fifth push is comfortably positive, and so is the case for a sixth.

That is the finding, and it is uncomfortable. On revenue alone the model says keep going. The reason to stop is the other channel: the 217 people who switched off will also miss the message about the keynote moving halls, the shuttle timetable changing and the fire door at the north end being closed. Those messages have no price, which is exactly why they lose every argument with a price attached.

So the cap is a policy decision made once, in advance, by somebody who is not carrying a sales target. That is a governance answer to a pricing question, and I have not seen a version of this that works any other way.

Who owns the notification channel?

Give it one owner and one calendar, and put every send on it: sponsored, marketing and operational.

The rule that holds up is that operational messages are never sold and never counted against the sponsored cap, and that the sponsored cap is small enough that the total across all three categories stays in single figures for a three day show. Four sponsored sends across three days, at most one per sponsor, none within three hours of another, none before eight in the morning or after seven in the evening local time, and none at all on the final afternoon when the operational load is highest.

Write those constraints on the rate card rather than in an internal document. A sponsor who is told the show sells four of these per edition understands they are buying scarce inventory, which supports the price. A sponsor who finds out afterwards that eleven went out understands something different.

There is demand-side evidence that this unit deserves the care. CEIR published the first report in its Exhibit and Sponsorship Sales Approaches series on 9 October 2023, asking organisers which offerings they found effective at generating revenue, and push text messaging came back rated effective by 80 per cent of them. That puts it above digital signage on premises at 78 per cent and above mobile app advertising at 72 per cent. Organisers rate it highly because it works, and it works because it is rare. Selling it as though it were abundant removes the property that made it valuable.

What to measure on every send

Five fields, recorded per send, and the whole thing becomes manageable.

Enabled base at send time. The denominator, captured before the send rather than reconstructed later.

Delivered. What the push service accepted and confirmed. Tokens go stale, devices are off, and delivered is always below the enabled base.

Taps. The only engagement signal available, and the one the sponsor's cost per tap divides by.

Permission disables in the following hour. Attributed to the send. This is the cost line and it is the field nobody builds by default.

Uninstalls in the following hour. A rarer and more severe version of the same signal, worth separating because an uninstall also removes the person from every other unit in the app inventory.

With those five you can produce the attrition curve above from your own data instead of assuming it. My guess is that most shows will find the first send costs almost nothing and the fourth costs several times as much, but the shape is an empirical question and your audience is not mine.

Where this stops

Three limits, and the first one is a measurement gap rather than a method problem.

Platform reporting is coarse. You can see delivery outcomes and token invalidation, and on iOS you can read the authorisation status when the app next launches, which means a person who disabled notifications and never opened the app again may sit in your enabled count until the next edition. Attribution of an opt-out to a specific send is therefore approximate, and a sixty minute window is a convention worth stating in the report.

Time zones break the schedule at an international show. A send timed for ten in the morning local reaches a delegate still in their home time zone at three in the morning, and the opt-out cost of that send is concentrated in exactly the attendees who travelled furthest to be there. Segment by registration country or accept the cost.

The last limit is that opt-out understates the damage. People who keep notifications enabled and start ignoring them have cost you the channel just as thoroughly, and nothing in the log distinguishes them from an engaged user. Falling tap rates across an edition are the closest proxy, and they belong in the same review as the count of units on sale, which is the general question inventory sizing works through, and alongside how the equivalent send to your list is priced in sponsored email. It is one exhibitor analytics picture of a finite audience being spent.

Count the notifications your app sent during the last edition, split them into sponsored, marketing and operational, and put the dates on a single line. If the total surprises you, you have found the cap conversation, and it costs nothing to have it before next year's card is printed.

Questions people ask about push notification sponsorship pricing

How many sponsored push notifications should a show sell?
Set a hard cap and print it on the rate card. Four across a three day show, one per sponsor, none within three hours of another and none outside show hours, is a defensible starting position. The cap exists because the operational channel and the sponsored channel share one opt-out switch, and only the operational one is irreplaceable.
What audience does a sponsored push actually price against?
Users who installed the app and left notifications enabled. On iOS the person must accept a permission prompt, and since Android 13 the app has to request the POST_NOTIFICATIONS runtime permission as well. Registrations and installs both overstate the reachable base, sometimes by a factor of four.
How do you measure the cost of a sponsored push?
Record delivered count, taps, and disabled notifications in the sixty minutes after each send. The opt-out figure is the one that matters, because it is permanent for the edition and it reduces every later send, including the ones you need for a hall closure or a schedule change.

Related reading

All sponsorship articles