What the data actually shows
Psychologists have long observed that estimating elapsed time depends on attention. When a task fully occupies your attention, you have less spare capacity to monitor duration, and time tends to feel shorter in the moment — the more engrossed you are, the more time seems to vanish. This is the same mechanism behind "time flies when you're having fun," and it applies to absorbing online activity as much as anything else.
There is a meaningful distinction researchers draw between genuine flow — Mihaly Csikszentmihalyi's term for deep, skilled, challenging absorption — and what some have called "pseudo-flow": the passive, low-effort absorption of scrolling or autoplay. Both can make time disappear, but they are not the same experience. Flow tends to follow effortful engagement with a clear goal and leaves people feeling restored; passive feed-absorption more often leaves people feeling that the time was lost rather than spent.
The interface design matters too. Features like infinite scroll and autoplay deliberately remove the "stopping cues" that older media had — the end of a page, the end of an episode, the bottom of a list. Without a natural endpoint, the default is to continue, and continuation requires no decision while stopping does. The result is that sessions tend to run past the point most people would have chosen to stop, which is part of why estimates of online time are so often well below the actual figure.
Why this feels different from how it actually is
It feels like the time vanished rather than passed because you were genuinely absorbed for part of it — and absorption is exactly the state in which your sense of duration is least reliable. You are not misremembering out of carelessness; the monitoring system that would have flagged "this has been an hour" was busy.
It also feels different because the activity rarely produces a clear memory marker. Effortful, varied experiences lay down more distinct memories, which makes a stretch of time feel longer in retrospect. A long, uniform scroll produces few such markers, so when you look back, there is little to anchor the duration to — the time feels both short while it happened and short when you recall it, even though the clock disagrees.
And because there was no endpoint, there was no moment of completion to register. With a book or an episode, finishing gives you a natural checkpoint. A feed gives you none, so the experience has no edges — which is part of why it can feel like you simply surfaced from it rather than chose to leave.
A book ends, a TV episode ends, a phone call ends; a feed never does.
What the research says to do about it
The most reliable lever is reinstating the stopping cues that the interface removed. External endpoints — a timer, an app limit, a different room for the device, turning off autoplay — work because they replace a decision that absorption makes hard with a signal that arrives from outside your attention. The research on attention and self-control consistently favours changing the environment over relying on in-the-moment willpower.
Distinguishing the activities that leave you restored from the ones that leave you depleted is also useful. Genuine flow — a skill, a craft, an effortful game or project — tends to make time disappear and feel worth it. If the goal is not to eliminate absorption but to spend it well, the practical move is to notice which online time produces the flow feeling and which produces the "where did that go" feeling, and to shift accordingly.
Periodically checking the clock or your own actual usage figures helps recalibrate. Because absorbed time is systematically underestimated, seeing the real number — rather than your felt sense of it — is one of the few things that reliably corrects the gap between intended and actual use.
What the research says does not help
Resolving to "just have more willpower" against an interface designed to remove decision points tends to fail, because the design works precisely by not requiring a decision to continue. Asking willpower to fight a default it never gets prompted to override is a losing setup.
Treating all time-disappearance as a problem is also misleading. Losing track of time in genuine flow is associated with some of the most satisfying human experiences; the issue is not absorption itself but passive, cue-less absorption you didn't choose. Trying to stay constantly aware of the clock would also undercut the flow states worth having.
Relying on memory to estimate how long you were online does not work, because absorbed time is exactly the kind your sense of duration handles worst. Self-reported screen time tends to diverge substantially from measured screen time, so "it only felt like a few minutes" is not evidence about how long it actually was.
'It only felt like a few minutes' is not evidence about how long it actually was.
What this looks like in real life
'Just five minutes' that became an hour
Opening a feed to fill a moment, then surfacing an hour later genuinely surprised at the clock, is the ordinary case. Absorption left little capacity to track duration, and the uniform scroll laid down few distinct memories to anchor the time to — so it felt short both while it happened and looking back, even though the logged figure disagrees.
Two kinds of lost time
An evening lost to an absorbing craft or game can leave you feeling restored — that is closer to genuine flow. An evening lost to passive scrolling more often leaves the 'where did that go' feeling. Noticing which is which, rather than trying to eliminate absorption altogether, is the practical move the research points to.
Real numbers in context
There is no single clean statistic for "how much time disappears," but a consistent finding is worth knowing: self-reported screen and internet time tends to diverge meaningfully from objectively logged time, and people frequently underestimate. Reviews of studies comparing the two have found that self-estimates correlate only modestly with actual measured use — a reminder that the felt duration of online time is not a reliable guide to the real one.
The design point is concrete even where the numbers are soft: infinite scroll and autoplay are relatively recent defaults, and they replaced interfaces that had natural endpoints. The mechanism — removing stopping cues so that continuing requires no decision — is well documented in how these products are built, even if the exact minutes lost vary enormously from person to person.