Too many browser tabs: what the habit is actually doing

The counter says forty something. The favicons have shrunk to slivers. Every one of those tabs was opened for a reason, and the reasons have long since stopped being visible. The usual explanation offered online is a personality one: curiosity, perfectionism, fear of missing something, an inability to decide. Those descriptions are not wrong, but they are not actionable either. Knowing that a habit comes from anxiety does not tell anyone which tab to close. What helps more is noticing that the strip contains at least four different kinds of object, and that only one of them was ever supposed to be a tab.

The research points at the tool, not the temperament

The most cited study of this behaviour is a 2021 paper from Carnegie Mellon University, "When the Tab Comes Due: Challenges in the Cost Structure of Browser Tab Usage," presented at the CHI conference. The team interviewed information workers about their tab habits, walked through every open tab on their machines repeatedly over two weeks, then surveyed a larger group to see how common each pressure was. The university's summary of the work reports that about 25% of the participants in one part of the study said their browser or computer had crashed because too many tabs were open.

The interesting finding was not the crash rate. It was the reason tabs stay open even when the person hates having them open.

People feared that as soon as something went out of sight, it was gone. Fear of this black hole effect was so strong that it compelled people to keep tabs open even as the number became unmanageable. Source: hcii.cmu.edu

That reframes the whole thing. A tab is not being kept because the person is untidy. It is being kept because the browser offers no other place where the item stays alive. Bookmarks, reading lists and history all exist, and all of them feel like a black hole, because nothing in the interface ever brings the item back unprompted. Keeping the tab open is a rational response to a tool that has no reliable resurfacing mechanism. The same paper notes that tabs are being used as an external form of memory while having none of the structure that memory needs.

Personality still plays a part in how far the habit runs. Someone who researches every purchase across a dozen review sites will accumulate faster than someone who buys the first acceptable option. But the accumulation happens to almost everyone, and it happens because of what the tab strip is, not because of who is sitting at the keyboard.

Four different objects, one strip

Look at any overloaded window and the tabs sort into four groups with almost no overlap.

Application tabs are the ones open every working day. Mail, chat, calendar, a task tool, a document workspace. They are not being read. They are being inhabited. They hold a logged in session, they carry a notification badge, and they are the destinations for most switches during the day.

Reminder tabs stand in for a task with no other home. An expense form to fill in. A reply to write. A booking to confirm. The tab is doing the job of a to-do list entry, badly, because it cannot hold a date or a note.

Queue tabs are things to read later. An article, a documentation page, a video. They have no deadline and no owner. They accumulate silently and are the largest group in most windows.

Decision tabs are the six hotel listings, the four laptop reviews, the three job descriptions. They exist as a set. Closing any one of them destroys part of the comparison, which is why they resist closing individually and tend to be abandoned all at once.

Kind of tab	What it is really holding	What closing it costs	Where it belongs
Application	A live session and a notification stream	Signing back in, losing the badge	Its own persistent window
Reminder	A task with no due date	Forgetting the task entirely	A task list with a date
Queue	Something to read later	Nothing, if it can be found again	A reading list checked on a schedule
Decision	The working state of an unfinished choice	The comparison itself	A written shortlist

The four groups need four different answers. Applying one answer to all of them is why tab cleanups never last. Closing everything on a Friday afternoon deletes the reminders along with the queue, the reminders come back as missed work on Monday, and the lesson learned is that closing tabs is dangerous.

The application tabs are the expensive ones

They are usually the minority by count and the majority by cost. Ten of forty tabs might be applications, and those ten absorb almost every switch made during the day.

Sharing a strip with thirty research tabs does specific damage to them. The keyboard shortcuts that jump to a numbered tab stop being reliable, because opening one search result shifts every position after it. The favicon is often the only thing left to aim at, and two accounts of the same service have the same favicon. Clicking a link from chat opens it in whichever profile the browser considers current, which for Google accounts is normally the one signed in first during that session rather than the one being looked at. Notification badges appear in a strip where thirty other titles are competing for the same few pixels.

None of that is fixed by closing the research tabs. It is fixed by taking the applications off the strip entirely and giving each one a window that persists, keeps its own session, and can be reached by name rather than by position. The way a set of those windows can be grouped and switched as a unit is covered in Workspaces, and the behaviour of a window that keeps a service logged in is described in Features.

The test for whether a tab belongs in this group is simple. If it would be reopened tomorrow morning without thinking, it is an application, and it should not be competing with a half-read blog post for space.

The queue is a queue, and it needs a schedule

The reading queue is where the fear identified in the research does the most work. Every tool for storing an article exists already: reading lists, bookmarks, pinned messages, read later services. They are avoided because saving something into them feels identical to deleting it.

The missing piece is not storage. It is a moment when the queue is looked at. A queue with a fixed review slot stops being a black hole, because the item does come back. Anything from a Friday afternoon slot to a recurring calendar entry works, and the specific mechanism matters far less than the fact that one exists. Until that slot exists, moving articles out of tabs into a list is just moving the anxiety somewhere less visible, and the tabs will quietly return.

There is a second, harsher option that some people find easier: accept that most queue items will never be read, and close them without saving. The information is on the open web and can be found again. That option is available specifically because queue tabs, unlike reminder tabs, cost nothing when lost.

Decision tabs close when the decision closes

A set of comparison tabs is not clutter. It is the working state of a task that has not finished, held in the only place available. The research describes exactly this pattern: sense making tasks that require absorbing information from many sources, stitching it together and reaching a conclusion, with tabs standing in for the structure of that thinking.

That means no tab management technique will clear them. The tabs go when the decision goes. Two things move a decision forward when it has stalled. Writing the shortlist down in three lines, with what each option costs and what would settle the question, converts twelve tabs into one note. Setting a deadline for the decision converts an open question into a closed one, even if the answer is to postpone deliberately rather than by drift.

A stalled decision that has been open for weeks is worth a separate look. If a set of tabs has survived three restarts, the decision is probably not blocked by missing information.

What the personality explanations get right

Search results on this topic tend to sort people into types. The curious one who follows every link. The perfectionist who wants the complete picture before acting. The anxious one who cannot risk losing a page. The one who postpones every small decision. Those descriptions land because they are recognisable, and they do explain differences in volume between two people using the same browser on the same day.

Where they stop being useful is at the point of action. A trait is not something that can be changed on a Tuesday afternoon. The tab strip is. Someone who follows every interesting link will keep doing that, and the only question worth asking is whether those links pile up in the same row as the mail client or somewhere separate. Someone who postpones decisions will keep postponing, and the useful intervention is a written shortlist rather than a resolution to be more decisive.

There is also a fairness point buried in the type lists. High tab counts correlate with jobs that involve comparing, researching and reconciling information from many sources. A person who opens eight tabs to answer one question is often doing the work properly. Treating that as a flaw, rather than as a task that has outgrown its tool, gets the diagnosis backwards. The research team behind the CHI paper made the same argument from the other direction, noting that managing this kind of task is one of the most important parts of productivity and that the main tool everyone reaches for is tabs, which do not do the job well.

What browsers changed, and what that fixes

Browsers have already accepted that people keep many tabs open. Chrome's performance settings include a memory saver that deactivates tabs that have not been used, reloading them automatically when they are visited again, and it can be tuned between moderate, balanced and maximum savings. That is a real improvement to how a laptop behaves with sixty tabs open.

It also solves a different problem from the one being discussed here. Memory management makes an overloaded window survivable. It does nothing about the fact that the mail tab and the half-read article are still the same size, in the same row, competing for the same attention. Machine resources and human attention run out separately, and the second one is the reason the tab count feels bad.

What to change first

Count the tabs that would be reopened tomorrow morning without thinking. That number, usually somewhere between six and twelve, is the set that should not live on a tab strip at all. Give those a window each and a name each, and the remaining tabs stop being noise, because what is left is honestly a queue and can be treated as one. A tool built around that separation is SpaceDeck, and what it covers is listed in Supported apps.