When your inbox grows faster than you can answer it

A four-digit unread count gets read as a character flaw — proof that someone is disorganised, or not working hard enough. It is far more often arithmetic. If forty messages arrive on an ordinary day and you can properly close twenty-five of them, the backlog grows by fifteen a day no matter how disciplined you are, and no amount of additional effort applied to the same method will close a gap that is structural.

This distinction matters because the two problems have completely different fixes. A discipline problem responds to trying harder. An arithmetic problem responds only to changing what counts as handling a message, reducing what arrives in the first place, or deciding openly that some categories will never be answered. Working out which one you have takes about a week of honest counting: how many messages arrive, how many you genuinely finish, and what the daily gap between those two numbers actually is. Most people have never counted, and are therefore trying to fix the wrong problem with more willpower.

Declaring email bankruptcy means accepting that everything older than a chosen date will not be answered, archiving it in one action, and starting from an empty inbox. It is the right call when the backlog is old enough that most of it is stale — anything genuinely urgent from three months ago has already been resolved, escalated, or has resurfaced through another channel. Nobody is still waiting politely on an email you never answered in April.

Digging out is the right call when the backlog is recent and contains live commitments: unpaid invoices, a client waiting on a quote, a supplier expecting confirmation. Here bankruptcy destroys real obligations rather than stale ones. The usable middle path is a bounded dig: work backwards from newest to oldest for a fixed period, then archive whatever you did not reach. If you declare bankruptcy, send a short note to your most important contacts saying so and inviting anyone still waiting to resend. That single message converts an act of abandonment into a reset people can work with.

The two-minute rule is simple: if a message can be fully handled in about two minutes, handle it the moment you read it, because filing it, remembering it, and returning to it costs more than the reply itself. It works well on acknowledgements, one-line answers, calendar confirmations and forwarding something to the right person.

It breaks in two specific ways worth knowing before you adopt it. First, it misjudges messages that look short but are not: a one-line question that requires checking a figure, opening a file, or confirming something with somebody else is not a two-minute message, and starting it as though it were is how a triage pass turns into an hour. Second, applied without limit it turns your whole day into reactive work — every quick reply is genuinely quick, and thirty of them still consume the morning you needed for something harder. The rule works best inside a bounded triage window rather than as a permanent standing instruction to answer anything easy the instant it arrives.

Good filtering moves predictable, low-decision mail out of the way rather than trying to guess what matters. Newsletters, receipts, platform notifications, delivery updates and automated reports are all identifiable by sender and can be routed to folders you read deliberately instead of reactively. That is safe because these messages are not waiting on you.

The dangerous filter is the one that tries to judge importance. A rule that hides anything not addressed to you directly will eventually hide the one thread where you were CC'd on something that mattered. Two habits keep filtering honest. Keep filters based on facts about the sender, not guesses about urgency — a rule that says 'this came from a billing system' is verifiable, while a rule that says 'this is probably not important' is a prediction. And review the filtered folders on a fixed schedule, because a folder nobody opens is deletion with extra steps. If you find yourself repeatedly rescuing messages from a folder, that filter is miscalibrated and should be narrowed rather than trusted.

The cost of a large backlog is not mainly the unanswered mail. It is that the number is visible constantly, and it functions as a standing reminder of unspecified obligation. You cannot tell from the count whether it represents four hundred newsletters or four hundred angry clients, so the mind treats it as the worse case. That produces a low-grade dread that follows you into work that has nothing to do with email, and it makes opening the inbox feel like a confrontation rather than a task.

This is why the unread count is worth attacking directly even when the underlying mail is genuinely low-value. Getting the number to zero by archiving three hundred newsletters changes nothing about your obligations and a great deal about your relationship with the inbox, because the ambient signal stops being alarming. It also restores the count's usefulness as information: once zero is normal, a count of twelve means twelve real things, and you can read the number at a glance instead of ignoring it. A number you have learned to ignore has stopped being a signal at all.

Anything you have written more than three times in materially the same form is a template candidate: acknowledging receipt, sending directions to your office, answering a pricing enquiry, declining a supplier pitch, confirming an appointment. The point of a template is not to avoid thinking but to avoid re-deciding structure every time.

What keeps a template from sounding automated is leaving a deliberate gap for one specific line. A reply that opens with a sentence referring to something only this person said, followed by templated substance, reads as a personal message; the same body with a generic opener reads as a form. Three openers give the game away immediately — 'Hope this finds you well', 'I wanted to reach out regarding', and anything beginning with an apology for a delay you have apologised for before. Write templates in your own register rather than in business English you would never speak aloud, and keep them short enough that the personal line is a visible proportion of the message. A template that takes thirty seconds to personalise is doing its job; one that takes five minutes has become a first draft.