We Built a World Full of Notifications and Somehow Made Information Harder to Find

A notification is meant to be a small courtesy: something happened, and you may want to know. Somewhere along the way it became a tiny electronic person tapping us on the shoulder every twelve seconds, increasingly convinced that its concern is the only concern in the room.

The odd result is that we have more access to information than any generation before us, while routinely failing to find the one thing we actually need. The parcel update is lost among supermarket offers. The message from a colleague sits underneath a sports score, a password prompt, and an app announcing that we have not opened it lately. Which, frankly, was beginning to feel like a sensible boundary.

This is not merely a complaint about modern life, although I am British and therefore professionally qualified to complain about queues, weather and badly designed interfaces. It is a design problem. We have confused delivery of information with access to information, and treated interruption as proof of importance.

An alert is not the same thing as information

In engineering, an alarm should mean something. A temperature-monitoring system, for example, cannot simply shout every time a sensor value moves a fraction of a degree. Sensors have tolerances, systems have noise, doors open, compressors cycle, and readings need context. If every minor fluctuation triggers an alarm, people quickly learn the most dangerous lesson in any safety-critical environment: ignore the noise.

That is alarm fatigue. It is not a mysterious failure of human attention. It is a predictable response to a system that repeatedly asks for attention without earning it.

Consumer software has adopted the same failure mode on an industrial scale. A notification saying someone liked a post and an alert warning that a bank transaction may be fraudulent are delivered through the same narrow channel: a badge, a vibration, a banner, a red dot. The interface may make slight cosmetic distinctions, but our nervous systems are not a well-indexed database. Repeated low-value urgency degrades the meaning of genuine urgency.

The trouble is compounded by duplication. The same event may generate an email, a push notification, an in-app badge and, if the product team is feeling particularly enthusiastic, a reminder about the notification you did not read. Information is no longer communicated. It is pursued down the street wearing a fluorescent tabard.

Finding requires structure, not more pings

The most useful information often is not urgent. It is something you need at the point of action: the address before leaving home, the detail of an appointment, the name of the person who sent a useful document, the family message you meant to answer properly rather than with a thumbs-up while holding a kettle.

Yet many digital products are designed around feeds, recency and engagement rather than retrieval. They are very good at telling us what is new, less good at helping us find what matters. Searching can be weak, messages disappear into endless threads, and important details are trapped inside formats that cannot be sensibly indexed.

Anyone building a system that produces events should ask a dull but vital question: what will the person need to do with this information later? If the answer is “find it again”, then notification delivery is only the beginning. The information needs a stable home, sensible names, searchable content and a clear history.

This is why a well-designed audit trail beats a flurry of alerts. In technical systems, an event record should say what happened, when, where, how serious it was, and whether anyone acted on it. It should also remain available after the initial alert has gone. A notification can invite attention; it cannot substitute for a record.

Urgency needs a budget

We should be much more ruthless about what is allowed to interrupt us. Every alert spends a little of the user’s attention and trust. Neither is an infinite resource, despite what some product roadmaps appear to believe.

A useful approach is to classify events by consequence:

• Immediate action required: security warnings, genuine safety risks, failures that will worsen quickly.
• Action required soon: a request needing a reply, a payment due, a task blocking other people.
• Useful awareness: delivery updates, routine status changes, ordinary reminders.
• Pleasant but entirely optional: likes, promotional nudges, and apps feeling lonely.

Only the first category normally deserves a disruptive push notification by default. The second may warrant a carefully configured alert. The rest should be available when the person chooses to look, perhaps in a daily digest or an activity view that does not behave like a smoke alarm because somebody uploaded a new profile picture.

For system designers, this means supporting notification preferences that are meaningful rather than decorative. Let people choose channels, quiet periods, severity thresholds and summaries. Avoid dark patterns that make opting out labyrinthine. If it takes six screens to silence an app, the app has already told you what it thinks of your time.

The human cost is easy to underestimate

Notification overload is not equally irritating for everyone. For some disabled people, frequent interruptions can carry a heavier cognitive, sensory or emotional cost. A device that buzzes unpredictably can be distracting, stressful or simply inaccessible. “Just turn them off” is not always an answer when a phone also carries medical, care, work, family or accessibility-related communications.

The better question is how systems can make priority legible. Can a person distinguish a critical message from a routine update without reading a screen? Can they reliably defer something without losing it? Can they return later and understand what happened, rather than facing a digital kitchen drawer full of uncategorised odds and ends?

These are not luxurious features. They are part of respecting users as people with finite attention, varied needs and lives that do not exist solely to service an app’s engagement graph.

A quieter digital life does not mean knowing less. It means building tools that let important information arrive clearly, routine information wait politely, and everything remain findable when it is actually needed. That is a more modest ambition than constant engagement, perhaps, but it is also considerably more useful. And usefulness is a rather good place for technology to start.