Open your phone right now. Somewhere on it, a red circle with a number in it is telling you how bad your day is. 342. 1,204. 51. That number is the single most influential metric in email, and it measures almost nothing useful.
Unread count answers one question: how many messages have you not yet clicked on. It does not answer whether any of them matter. It has never once told you what to do next. And yet it's the number every email client puts front and center, badge on the icon, bold in the tab title, impossible to ignore.
What unread count actually tracks
Unread is a binary flag flipped the moment you open a message. That's it. A newsletter you'll never read and a wire transfer confirmation you need to act on today both count as 1 unread, identically weighted, identically urgent-looking.
The number goes up when a mailing list posts. It goes up when a colleague CCs you on something that doesn't need a reply. It goes up when spam slips the filter. None of those events changed what you need to do today, but the badge doesn't know that. It just counts.
The badge as anxiety generator
A high unread count produces a specific, useless feeling: generalized dread with no target. You know something in there might matter, but the number doesn't tell you what or where. So you either ignore the badge (and risk missing something real) or you open the inbox and scroll (and burn the morning finding the needle).
Inbox zero culture tried to fix this by making 0 the goal. But 0 unread doesn't mean 0 things need you, it just means you've clicked on all of them. You can hit inbox zero and still miss the one email that actually mattered, because you were too busy clearing the badge to notice it.
Unread count measures whether you clicked. It has nothing to say about whether you should have.
What urgency-based measurement looks like instead
The right question isn't "how many have I not opened." It's "how many need me today, and in what order." That's a completely different number, built from completely different signals:
- Who sent it, and have they emailed you before.
- Whether it contains a direct question or a deadline.
- Whether it's part of a thread you already replied in.
- Whether it's transactional (receipt, confirmation) versus conversational (someone waiting on you).
- How much time has passed since it landed.
None of that is visible in a read/unread flag. All of it is visible if you actually look at the content and the pattern, which is exactly the kind of thing a fully on-device classifier is good at and a badge counter is not.
What replaces the badge
Instead of "347 unread," imagine your inbox opening with a single sentence: 7 threads need you today. That's the whole headline. Not a count of everything you haven't clicked, a count of what's actually on you right now. The rest exists, searchable, but it isn't shouting at you from the app icon.
This is the design decision behind STAMP. Classification runs entirely on your Mac, no server round trip, and it sorts by urgency, not by whether you've opened the message. The badge you'd normally dread becomes a number you can trust, because it's been filtered down to only the threads worth the anxiety.
A test you can run today
Look at your current unread count. Now ask yourself, honestly, how many of those messages need a reply from you in the next 24 hours. For most people the real number is somewhere between 3 and 10. The gap between that number and the badge is the size of the lie unread count has been telling you.
If your client can't answer that second question, it's optimizing for the wrong metric. See how STAMP counts differently on the changelog, or read about the 7-email rule for the version of this idea we ship by default.
What to do this week
Turn off the badge notification for one day. Instead, ask yourself each morning: what actually needs me today. Write that number down. Compare it to what your unread count said. The difference is the tax the wrong metric has been charging you.
Count what matters, not what's clicked. hello@stamp.email