Against Notification Culture

The notification is a piece of design. Like all design, it embodies decisions made by people with particular interests and constraints. The people who designed the notification architecture you live inside were solving specific problems — problems of user retention, engagement measurement, and platform stickiness. They were not solving the problem of how you might best organise your working day. These are not the same problem. In most cases, they are opposite problems.

This sounds obvious, and it is obvious, once stated. The reason it is worth restating is that most people's relationship to notifications is conducted as though the system were neutral — as though the question of whether to have notifications on or off were simply a matter of personal preference and practical convenience. It is not. It is a question about whose interests your attention serves, and whether you have any meaningful say in that.

Design for whose benefit

Notifications are not primarily designed to help you receive information more efficiently. They are designed to ensure that a platform, or an application, or another person's message, can interrupt whatever you are doing and redirect your attention. The word "interrupt" is doing significant work here. An interruption is not simply a delivery mechanism. It is a forced change of context — a demand, embedded in the design, that you stop what you're doing and attend to something else.

The platform's interest in this is clear. An app that cannot interrupt you cannot compete for your time against the current activity. Notifications are the mechanism through which apps defend their position in your attention economy. A platform that gets you to check it fifty times a day has fifty opportunities to show you something that increases engagement. A platform you check intentionally, twice a day, at times you choose, has two. The design pressure toward persistent notification is therefore not accidental. It is structural.

None of this is a hidden or controversial observation. Most people in the technology industry understand it perfectly well. What is less often examined is how completely this design pressure has been normalised — to the point where turning off notifications for a work communication tool is treated, in many organisations, as a statement of antisocial intent rather than a reasonable management of one's own cognitive resources.

The cost model

The psychological literature on attention is fairly consistent on the subject of interruption costs. Interruptions do not simply consume the time it takes to process them. They impose a switching cost — a period during which attention is not fully recovered to the original task. The length of this recovery period is context-dependent, but the literature suggests that for complex cognitive work, it can extend for minutes rather than seconds. This means that a notification that takes fifteen seconds to read and dismiss may cost several minutes of productive engagement with the work it interrupts.

The more interesting point, which is less often discussed, is that this cost is invisible to the person experiencing it. There is no phenomenological signal that you are in recovery mode after an interruption. You do not feel less capable of the work. You simply are less capable of it, in a way that cannot be introspected. The common experience of sitting down to work and finding oneself unable to make progress without any obvious cause is frequently a cascade of invisible recovery periods from interruptions accumulated throughout the day.

This has a direct implication for the question of notification management: you cannot calibrate your notification settings based on how interruptions feel, because they don't feel like what they cost. You have to reason from evidence rather than experience, and the evidence is fairly clear about what frequent interruption does to the quality of complex work.

Ambient availability

The specific social phenomenon that makes notification management genuinely difficult is what I have come to think of as the norm of ambient availability. In many organisations and social contexts, the default expectation is not just that you will respond to communications within a reasonable time frame. It is that you are, at any given moment, likely to be aware that a communication has arrived. This is a different standard. It requires persistent monitoring, which requires persistent interruption.

The norm has been constructed gradually and largely without deliberate decision. It emerged from the convergence of always-on devices, platform notification defaults, and a set of social scripts about responsiveness that were imported from the phone call — a medium in which real-time synchronisation was technically mandatory — into asynchronous communication, where it is not. The expectation of near-immediate response to a message that has no technical requirement of immediacy is a historical accident that has become a social norm, and it is one that has significant costs for anyone trying to do work that requires sustained attention.

The difficulty with any individual attempting to change this is that the costs of non-compliance fall entirely on them. If your organisation expects near-immediate responses and you begin responding in batches, the cost of that decision — in social friction, in missed windows, in the particular kind of professional anxiety that comes from being seen as unresponsive — is yours to carry. This is true even if the change is rational, well-explained, and demonstrably better for the quality of your work. Social norms do not yield easily to rational argument, particularly when the norm serves the interests of some participants at the expense of others.

What refusal looks like

Despite these constraints, most people have more room to manage their notification environment than they exercise. The gap between the notification settings people have and the ones that would serve their work better is, in my observation, usually a gap created by inertia and social anxiety rather than by genuine organisational necessity.

The practical question is where to start. The clearest case is turning off all notifications for any application that does not involve direct human communication — software update prompts, news apps, any kind of automated alert that can be reviewed at a scheduled time rather than absorbed in real time. These produce interruption costs with essentially no offsetting benefit from immediacy. They can be eliminated without social consequence in almost every case.

For direct communication tools, the question is more contextual. But the starting point is distinguishing between what actually requires immediate response and what has been allowed to accumulate the expectation of immediate response by default. Most of what arrives in a working day falls into the second category. The test is simple: if you responded to this two hours later, would anything of actual consequence happen? For the majority of workplace communications, the honest answer is no. For those where the answer is yes, scheduled checking periods are still usually sufficient — the genuinely urgent message can be accommodated within a structure of deliberate rather than ambient monitoring.

The goal is not silence. It is sovereignty over when you choose to listen, as distinct from having that choice made for you by platforms optimised for other interests than your own.