Why Do We Keep So Many Browser Tabs Open?

There is a genre of internet writing about browser tabs that treats them as a personality flaw. You have too many. You are disorganised. Close them and reclaim your focus.

The research says something less moralising and more interesting: tabs are being used as a memory system, they have been used that way for as long as anyone has studied them, and the reason people will not close them is that closing a tab deletes a thought.

It also says the human cost of this, as opposed to the cost to your laptop’s RAM, has barely been measured at all. This article is mostly about what the evidence shows. It is partly about how much of what you have read on this subject was made up.

How many tabs do people actually have?

Fewer than the memes suggest, and the number you have seen is probably a mean dragged upward by a handful of extremists.

The best logged data (not self-report, actual browser logging) comes from Rutishauser and Fritz (2026), who instrumented the browsers of 29 knowledge workers for around eight workdays each. Participants had “an average of 52.62 (±154.54) open tabs in their browsers at any point in time.” But the median was 11.48.

Look at that standard deviation: 154.54, three times the mean. The distribution is extremely skewed. Most people have around a dozen tabs open. A few have hundreds, and they are what makes the average meaningless.

Earlier logging found the same shape. Dubroy and Balakrishnan (2010) logged 21 Firefox users for 13 to 21 days each and reported that “the most common number of tabs to have open is one, with a steady descent down to 9.” One participant had a median of 17; nobody else exceeded 6; the highest count anyone reached at once was 42.

If you have seen “the average person has 50 tabs open,” you have seen a mean quoted without its median.

The real number to know is 8, and it is not what you think either

Chang and colleagues (2021) asked 103 participants at what point they would “start feeling overwhelmed and experiencing difficulty managing” their tabs. The median answer was eight (first quartile 5, third quartile 12).

This figure circulates as “science says 8 tabs is the limit.” It is nothing of the kind. It is a self-reported discomfort threshold, not a performance ceiling, not a cognitive capacity, and not a measurement of anything. Nobody’s comprehension, error rate or task speed was tested. It is the number at which people say they start to feel bad, and it sits below the logged median of 11.48, which tells you something real about the gap between how many tabs people keep and how many they are comfortable keeping, and nothing at all about an optimum.

What tabs are actually for

Chang et al.’s study is the closest thing this field has to a definitive account. It combined a preliminary survey of 64 people, in-person interviews with 10 information workers who walked researchers through every open tab on four occasions across two weeks, and a survey of 103 participants who labelled their own tabs against a taxonomy.

They found six pressures pushing people to keep tabs open. Measured as a percentage of the 633 tabs participants labelled:

(The figures exceed 100% because tabs were assigned an average of 1.39 reasons each.)

The two bolded rows are the same phenomenon at different scales. A reminder tab is a single intention parked in the world. An external mental model is a whole task laid out across a tab bar: this is what I am doing, in the order I am doing it.

In the interview study, this was universal: “All participants in Study 1 pointed to tabs that were intentionally kept around as reminders for unfinished tasks.”

And it is not new. Dubroy and Balakrishnan found exactly the same thing eleven years earlier: “17 participants reported that they often use tabs as a reminder to do something,” 17 of 21, in 2010, before anybody was writing think-pieces about tab overload. The behaviour is not a symptom of modern digital chaos. It is what a tab bar is for, whether or not it was designed for it.

Most of them are never looked at again

Here is the part that should give you pause.

Dubroy and Balakrishnan, logging actual behaviour: “45% of tabs were selected exactly once” and “25% of tabs were never selected.” More than two-thirds of tabs were used once or less.

Rutishauser and Fritz, sixteen years later on different software with different people: for participants with more than ten open tabs, “a substantial share of tabs (37.88%) was not revisited throughout workdays and remained idle.”

Two independent logging studies, a decade and a half apart, both find that a large fraction of open tabs are simply never returned to. The tab bar is a to-do list where roughly a third of the items are never done and never crossed off.

There is a related detail in the newer study that says something precise about what is going on. When participants were given a tool to save pages instead of keeping them open, 88.43% of saved web pages were subsequently closed, and the majority of these (80.20%) were closed within 2 minutes of saving. The moment the information was safely somewhere else, the tab went. People were not attached to the tabs. They were attached to not losing what was in them.

What it feels like, in the participants’ own numbers

From the preliminary survey of 64 people in Chang et al.:

• 59% agreed that if they looked at their tabs right then, they would find some that should be closed.
• 55% agreed they “feel like they can’t let go” of their tabs.
• Only 19% agreed that laziness had anything to do with not closing them.
• 30% agreed they “have a tab hoarding problem.”
• 28% said they often struggle to find the tab they need.
• 25% reported that their browser or computer had crashed because of too many open tabs.

And from the 103-person survey, 67% said they feel overwhelmed by their tabs at least once a week; 17% daily.

That 19% figure is the one worth sitting with. People do not think they are being lazy. They think they are holding onto something. The 55% who “can’t let go” and the 19% who blame laziness are describing the same behaviour and disagreeing entirely about what it is.

A caution about the 25% crash figure, which is the single most-recycled statistic in this literature: it is 25% of a preliminary survey of 64 people, self-reported, lifetime, in a paper whose headline samples are 10 and 103. That is roughly sixteen people. Carnegie Mellon’s own press release hedges it as “about 25% of the participants in one aspect of the study,” and almost every retelling drops the hedge.

The thing everyone says about this that is not true

“Open tabs nag at you: it’s the Zeigarnik effect.” You have read this. It is the standard explanation, and it did not survive being tested.

Ghibellini and Meier (2025) meta-analysed 59 publications on the Zeigarnik effect (better memory for unfinished tasks) and the related Ovsiankina effect (a tendency to resume unfinished tasks). Their finding on the memory half: the weighted ratio of interrupted-to-completed tasks recalled was 0.99, interrupted tasks accounted for 49.16% of recall (chance) and the pooled effect was Cohen’s d_z = 0.15. Their conclusion: “the Zeigarnik effect lacks universal validity.”

The resumption half held up. The weighted rate of returning to an interrupted task was 67.00% across 21 publications.

That is actually the better fit for tabs anyway. An open tab is not an intrusive memory that nags at you ; the evidence says unfinished tasks do not do that. It is a resumption cue: something in the world that makes going back easier when you pass it. Using the environment to hold an intention is the same move examined in the cognitive offloading guide, and the wider point that cues and defaults do more work than willpower is in your phone isn’t the problem, it’s defaults. That is a more modest mechanism and a more accurate one.

“Digital hoarding research explains browser tabs.” It does not. We checked the primary sources. Neave, Briggs, McKellar and Sillence (2019) developed and validated the standard Digital Hoarding Questionnaire across samples of 424 and 203 adults; the words “tab” and “browser” do not appear anywhere in the paper, and there is no tab item in the instrument. Sweeten, Sillence and Neave (2018), the qualitative study of 45 people that founded this literature, describes the hoarded material as “emails, photographs, files and software.” A 2023 follow-up on personal digital data (N = 167) also does not mention tabs.

The only people calling tabs hoarding are tab researchers borrowing the word, and their evidence for it is that 30% of 64 survey respondents agreed with a statement containing it. The clinical construct has never been validated on tabs.

What it costs you: and the honest gap

Your device: measurable, and measured by an interested party. Microsoft reported that its Sleeping Tabs feature “typically reduces memory usage by 32% on average” and that “a sleeping tab uses 37% less CPU on average than a non-sleeping tab,” based on telemetry aggregated across roughly 40,000 devices. That is a real measurement. It was also published on a product blog to promote the feature, and nobody independent has replicated it.

You: here the evidence runs out, and we are going to say so rather than fill the gap.

There is no controlled experiment measuring the effect of tab count on attention, task performance, error rate or stress. We searched for one. What exists is: a qualitative taxonomy of pressures derived from ten interviews; self-reported harms from a 64-person survey; and logged behaviour showing that people with many tabs switch between them a great deal (Rutishauser and Fritz logged 63.66 tab switches per hour on average, though with a standard deviation of 91.63 that number is as skewed as the tab counts).

Everything you have read asserting that tabs damage your focus is either self-report, inference from the general task-switching literature (which was measured on things that were not tabs) or invention. The claim may well be true. It has not been tested.

What to do about it

The research does not license a lecture, so here is what it actually supports.

Recognise what you are protecting. The 88% figure is the useful one: when people had somewhere reliable to put a page, they closed it within two minutes. The tab is not the thing you want. The retrievability is. Any system that makes retrieval trustworthy (a real bookmarking habit, a reading list, a notes file, a task manager) takes the pressure off, and the evidence says the tabs then go on their own.

Expect a third of them to be dead. Two logging studies fifteen years apart agree that roughly a third to a half of tabs are never revisited. If you are hesitating over a tab because you might need it, the base rate says you probably will not.

Do not treat the discomfort as a character problem. 55% of people say they cannot let go; 19% think it is laziness. The gap between those numbers is the gap between what this behaviour is and what people assume it is. Keeping a tab open is an attempt to remember something, and it is a reasonable attempt in an interface that offers nothing better.

And be sceptical of the numbers. Eight is not a limit. Fifty is not an average. Tabs are not clinically hoarded. Zeigarnik did not replicate. This is a small literature with a very large folklore attached to it.

A note on this article’s keyword

The obvious term here is open tabs (2,400 US searches a month). We are not targeting it. Its search results are dominated by Google Chrome and Mozilla Firefox help documentation and a Chrome Web Store extension, and its question cluster is “how to open an incognito tab” and “how to close all open tabs.” The people searching it want to operate their browser, not understand their behaviour.

too many tabs open (210 searches, and a results page of forum threads, essays and community discussion) is the query that matches this article. It is a tenth of the volume. It is the honest one.