SuckButtBecause some things just suck

Tech That Sucks

Every app now wants to send you notifications about itself

The permission prompt appears before you have used the thing, because asking later would mean asking someone who has an opinion.

Two broken smartphone screens on an orange background, illustrating device fragility.
Photograph by Towfiqu barbhuiya via Pexels
Editorial note. Independent reporting and analysis. Nothing here is sponsored or paid for. How we work.

The theory of notification permissions is well covered elsewhere. This is about the version you meet in practice.

What holds up in practice

  • Permission prompts shown at first launch convert far better than contextual ones.
  • Notification volume is tied to engagement targets rather than to events.
  • Per-app and per-category controls exist on most platforms and are rarely used.

The prompt arrives before you have any information

A permission request shown at first launch asks you to decide about a service you have not yet experienced, which is the point rather than a flaw. Asking at the moment a notification would actually be useful produces a much better-informed answer and a much lower acceptance rate.

Platforms have pushed toward contextual prompting for exactly this reason, and the first-launch prompt keeps reappearing because it works commercially. Some apps also show a fake pre-prompt first, so that a refusal there avoids spending the single real system prompt they are permitted. Recognising the pre-prompt is easy once you know it exists, because it uses the app's own styling rather than the operating system's.

Notifications are an engagement target, not a message queue

Product teams are frequently measured on daily active users, and a notification is the cheapest available lever on that number. That structure means notification volume is set by what moves the metric rather than by whether anything happened worth telling you about. It explains the entire genre of notifications that summarise activity you did not ask about, or that report on other people's use of the app.

Once the introductory rate lapses, re-engagement notifications aimed at users who have gone quiet are a specific category, and they are timed against your usual usage window. None of this requires anyone to be cynical, only for a team to optimise honestly toward the number they were given.

The cost is measured in attention, not in seconds

An interruption costs more than the time it occupies, because returning to a task requires reconstructing what you were holding in mind. Research on task switching consistently finds a recovery period after an interruption, and it is considerably longer than the interruption itself.

Notifications that arrive during focused work therefore charge a fee that is invisible on any device usage statistic. Batching, where notifications are delivered at set times rather than immediately, addresses this directly and is supported on most platforms. It is the single highest-value setting on a modern phone and almost nobody has ever opened it.

Marketing hides inside functional categories

Many apps combine transactional notifications with promotional ones under a single permission, so refusing the marketing means losing the delivery updates. Where a platform supports notification categories, splitting them is possible and depends entirely on the developer implementing it. Apps that do implement categories tend to be the ones confident that their useful notifications are genuinely useful.

The bit they bury in the confirmation email: where categories are absent, the practical choice is between everything and nothing, which is a design decision presented as a technical limitation.

Email follows the same pattern, which is why unsubscribing from a newsletter sometimes stops your receipts as well.

Web notifications are the same trick with fewer manners

Browser notification prompts on first visit are widely disliked, and browsers have progressively restricted them because the abuse rate was so high. Some browsers now block the prompt entirely by default or hide it behind an icon, which is an unusual admission of a broken pattern. Sites that ask before you have read a paragraph are asking whether you would like more of something you have not had yet.

Read the terms and there it is: a global setting to block notification requests exists in every major browser and removes the entire category permanently. It is a two-minute change that improves the internet more than most things you can do in two minutes.

One bad unit is bad luck; a pattern across thousands of buyers is a design decision.

Taking the phone back

Go through the notification settings once and deny everything except messages from people, then re-enable individually as you miss something. Almost nothing gets re-enabled, which is the most persuasive available argument about what the permission was worth. Turn off badges as well as banners, since a red dot is a notification that never expires and never has to justify itself.

On hold for the fourth time, use scheduled delivery for anything that is useful but not urgent, which covers most news, social and commerce apps entirely. Keep alerts for the small set of things where a delay causes a real problem, and be honest about how short that list is.

The takeaway

Deny everything, then re-enable only what you actually missed.

The fix is usually trivial, which is the most annoying part.

Questions readers ask

Does turning off notifications break the app?

Almost never for core functionality. The exceptions are messaging and two-factor prompts, where the notification is the feature.

Why do notifications come back after an update?

Some updates reset preferences or introduce new categories that default to on. Re-checking after a major version is a reasonable habit.

Tech That Sucksappsnotificationspermissionsattention
More in Tech That Sucks
Debashis Panda
Editor, SuckButt

Debashis edits SuckButt and has never once got through an automated phone menu without swearing.

Also by Debashis Panda