Most inbox systems fail on the day you need them most. This one has four outcomes, one pass, and no folder hierarchy to maintain.
Key takeaways. Triage is deciding, not doing. Separating the two is what makes ten minutes enough. Every message gets exactly one of four outcomes: reply now, becomes a task, waiting on someone, or gone. Two minutes is the cutoff. Under it, reply in the pass; over it, it is a task with a date. The routine has to survive being skipped. A system that needs a perfect streak is a system that ends on day twelve.
Why inbox zero keeps failing. The usual advice is a filing system: folders, labels, rules, and an empty inbox as the finish line. It works beautifully for about eleven days, and then one genuinely bad week puts four hundred messages in front of you and the system asks you to file all of them before you are allowed to feel finished. The flaw is not discipline. It is that the system measures the wrong thing. An empty inbox is a state you can reach by archiving everything, and a full one can be perfectly under control if you know what is in it. What you actually need to know is: is there anything in here that needs a decision from me, and have I made it? So the routine below optimises for decisions made, not messages filed. It takes about ten minutes on a normal day, it produces four possible outcomes, and it is designed to be resumable after you skip it for a week.
The one rule: triage is deciding, not doing. The single change that makes this work is refusing to do the work during the pass. The moment you start writing a careful reply to message four, the pass is over — you will surface forty minutes later with the other ninety messages untouched and the sense that email is an unwinnable game. During the pass you are allowed to decide, and to execute only the decisions that take under two minutes. Everything else becomes something with a name and a date, and you leave. If you find yourself opening a document during triage, you have stopped triaging. Note the task, close it, keep going.
The four outcomes. Every message gets exactly one. There is no fifth, and there is no "I will decide later", because that is what the inbox already was. Reply now — only if it takes under two minutes. A confirmation, a yes, a date, a thank-you. Send it and move on. Becomes a task — it needs real work. Write the task as a verb with an object and give it a day. "Reply to Priya about the Q3 numbers, Thursday" is a task. "Priya" is not. Waiting on someone — you have done your part and the ball is elsewhere. Label it and set a date to check, so it is not living in your head. Gone — archive or delete. Most newsletters, most notifications, most threads you were copied on for visibility. The two-minute cutoff is doing more work than it looks like. It is not about the two minutes; it is that having a bright line removes the per-message negotiation about whether this one is worth handling now, and that negotiation is where the ten minutes actually goes.
Making the pass shorter. Three habits compress the routine more than any tool will. Do it at a fixed time, once or twice a day. Continuous partial triage all day long is the most expensive way to read email, because each interruption costs the context you were holding. Sort by thread, not by message. A thread is one decision, however many messages are in it. Unsubscribe during the pass rather than after it. It is the only maintenance worth doing inline, because it reduces every future pass. What you should not do is build a folder hierarchy. Search is better than you are at remembering which of nine folders you chose eighteen months ago, and the hierarchy is a second system that also has to be maintained.
Where an agent helps, and where it does not. The part of this that a person is uniquely good at is the decision — whether this thread matters, whether the relationship needs a warmer reply, whether the deadline is real. The parts a person is bad at are the sorting, the remembering, and the retyping of a message into a task in a different app. That is the split worth automating. Drift reads across Gmail, Google Calendar and Google Tasks in one pass, so asking what needs attention returns a short list with the reason each message is on it, rather than a folder. Turning one of those into a task with a date is a sentence, not a copy-and-paste, and it can check the calendar before it proposes the date. It does not send anything for you unattended. Drafting a reply is free and happens in the conversation; sending it stops on a confirmation card showing the recipients and the full body, because a reply that goes to the wrong person is not a mistake you can take back. That is a deliberate trade — see what Drift asks before it acts for the exact list.
Ten minutes for a normal day, once or twice a day at a fixed time. If it is consistently running longer, the usual cause is doing work during the pass rather than deciding — the two-minute cutoff exists to stop exactly that.
Not as a goal in itself. An empty inbox can be reached by archiving everything, and a full one is fine if every message in it has already had a decision made about it. Aim for zero undecided messages instead, which is a state you can actually hold.
It can do the assembling: reading across your mail, calendar and tasks, surfacing what needs attention with the plain reason it is there, drafting replies, and turning a thread into a dated task. The decision stays yours, and sending anything always stops for your approval first.