The Monday inbox triage.

Monday because the weekend is when queues quietly rot — Friday's "I'll get to it" threads plus two days of new arrivals. Twenty minutes because the ritual must be cheap enough that you never skip it. A triage habit you skip under load is worthless; load is the whole point.

One person runs it — rotate weekly if you like — with the open-conversations view in front of them. This is a hygiene pass, not a work session: the triager's job is to put every thread into a correct state, not to answer them. Answering happens after, by the right people, precisely because triage made "right" visible.

It also matters that the ritual happens before the week's reactive work starts. Triage done at 4pm competes with everything that arrived since lunch, and loses; triage done before standup sets the table instead. If your team is distributed, "Monday" means the start of whoever opens the rotation — the point is a fixed slot at the boundary between weeks, not a particular timezone's morning.

And resist the temptation to automate the pass away. Auto-close timers and assignment rules exist precisely so nobody has to look at the queue — and nobody looking at the queue is how it rotted in the first place. The twenty minutes isn't overhead on the system; it is the system, the one moment a human reads the whole board.

Run the four passes in this exact order, because each one makes the next cheaper: closing shrinks the field before you sort it, sorting puts threads in front of the right owners, and ownership clears the board before you study it.

1. 1

   Minutes 0–5: close what's actually done

   Walk the oldest open threads first. Resolved-but-never-closed conversations are queue fog — every one you clear makes the real workload legible. If a thread is waiting on the customer, say so in the thread and leave it; never close it just to tidy the count. A closed conversation that reopens is fine. A customer who has to start over is not.

2. 2

   Minutes 5–10: fix the categories

   Scan for miscategorized threads — the "general" conversation that's actually a billing dispute, the support question that's really a sales lead. Categories drive routing, so a wrong category means the right people never saw it. Flip them now and the destinations catch up on their own.

3. 3

   Minutes 10–15: assign everything that's open

   By the end of this pass, every open conversation has exactly one assignee. Unassigned threads are where queues go to compound — everyone assumes someone else has it. Once it's done, each teammate's "Assigned to you" filter is their true Monday worklist, and the inbox stops being a shared anxiety.

4. 4

   Minutes 15–20: read the week's shape

   Look at what just passed through your hands. Three threads about the same confusion? That's a macro to write — or a docs fix to request. A macro nobody used in a month? Delete it. This is the five minutes where triage stops being maintenance and starts being intelligence.

Some Mondays the pass will blow through its box, and the overrun is information. If closing ate the time, your team is answering but not resolving — replies go out, threads never get marked done, and the count inflates all week. If categories ate the time, your sources have the wrong defaults; fix the default once instead of re-sorting every Monday. If assignment ate the time, the load is genuinely unbalanced, and no inbox feature fixes a staffing problem — but at least you found it on Monday instead of in a resignation letter.

The one thing the overrun should never become is a longer ritual. Hold the twenty minutes, write down what made it tight, and fix the upstream cause during the week. A triage pass that bloats into a full morning of admin will get skipped by February, and a skipped ritual protects nothing.

The goal isn't inbox zero. It's inbox honest: every open thread real, owned, and routed right.

Four Mondays in, the queue's character changes. The open count becomes trustworthy — when it says fourteen, fourteen customers are genuinely waiting. Stand-up questions like "who has the refund thread?" disappear, because the answer is on the thread. And your macro library starts evolving weekly instead of fossilizing, because pruning has a scheduled home.

The deeper effect is cultural. A queue that someone demonstrably tends is a queue people trust, and a trusted queue gets answered faster — the compounding starts running in your favor. Twenty minutes is a small price for interest that accrues in the right direction. If your current tool makes any of these four passes hard — categories, assignment, suggested replies — the fix costs nothing to try.

The ritual scales, too. The pass that takes twenty minutes at five teammates takes barely longer at fifteen, because every step works on exceptions rather than volume — categories are mostly right, threads are mostly owned, and Monday is mostly confirmation. Queues compound. So, it turns out, do habits.