Passage
The uses of friction
“Frictionless” has become a term of praise for services that remove pauses, forms and visible decisions. A payment happens with one tap; a subscription renews without a request; a traveller walks through a gate while a camera confirms identity. These systems can save time and reduce repetitive effort.
The difficulty is that friction does more than delay action. It can also reveal that an action is taking place, give people a chance to reconsider it and show where responsibility lies. Consider automatic renewal.
For a customer who uses a service continuously, renewal avoids a pointless annual ritual. For someone who has forgotten the subscription, the same convenience depends on inattention. A reminder is technically a small obstacle, but it restores the transaction to awareness.
Companies sometimes argue that cancellation is always available in account settings. Availability, however, is not equivalent to a decision being visible at the moment when money is taken. The distinction matters in public services as well.
Online forms can remove queues and allow applications at any hour. Yet a form designed to prevent mistakes may refuse an answer that does not fit expected categories. In a staffed office, an unusual case can become a conversation; in a frictionless system, it may become an error message.
The absence of human negotiation is efficient only when the categories already describe the person accurately. Designers often try to solve this problem by adding help pages or chat assistants. Such tools are useful, but they can leave the underlying process untouched.
If the system asks the wrong question, explaining that question more clearly does not create a suitable answer. Assistance becomes a layer around the decision rather than an opportunity to revise it. The user receives more guidance while the institution learns less about where its categories fail.
Some friction is plainly harmful. Repeatedly entering the same information, waiting for approval that no one meaningfully reviews, or navigating a building through unnecessary barriers does not produce reflection; it distributes fatigue. Defenders of friction sometimes romanticise inconvenience as though effort automatically created seriousness.
The relevant issue is not whether a process is easy, but whether the points at which it becomes difficult correspond to decisions that deserve attention. This is why confirmation screens vary in value. A warning before deleting a photograph that can be recovered from a folder is mostly noise.
A clear pause before publishing private information to thousands of people may be essential. When every minor action triggers an alert, users learn to dismiss alerts; when no action does, the system withholds distinctions that people need in order to govern themselves. A well-designed service therefore moves friction instead of simply removing it.
Routine steps become easier, while consequential transitions become more legible. The user can see what is happening, what can be reversed and who will receive the result. Smoothness remains valuable, but only after the service has decided which moments should not disappear into smoothness.
Friction also marks relationships among organisations. A consent process that feels repetitive to one institution may be the only moment at which a person can see that information is moving to another. Combining screens can make the journey smoother while hiding the change in responsibility.
Conversely, asking for confirmation at every internal transfer produces fatigue without additional control, so the design question concerns where accountability changes, not how many clicks a process contains. The placement of friction also determines whether refusal is meaningful, because an option technically available after ten screens and several transfers does not provide the same control as a choice presented before commitment, even when both routes satisfy a formal requirement. Designers often count the existence of an exit while users experience the probability of finding and understanding it under pressure.
A process should therefore be assessed through the sequence in which knowledge, commitment and reversibility change, not through an inventory of buttons that can be cited after the fact.