You can't judge your follow-up until you've archived one exactly like it

The email that gives you nothing

The email arrives at 8:41 on a Tuesday, three days after the conference. Subject line: "Great connecting last week!" The sender's name means nothing to you. Neither does the company, though it sounds vaguely like four others you saw on badges.

You read the first sentence. It says it was great to meet. You read the second. It says they'd love to grab twenty minutes this week to walk you through what they're building.

Nothing in the email tells you where you met. Nothing tells you what either of you said. No booth, no session, no hallway, no name of the person who introduced you.

So you do the thing everyone does. You open LinkedIn in a second tab, find the profile, study the photo for a beat, and feel nothing click. You close the tab. You archive the email.

You aren't being rude. The email asked you to do the remembering, and you couldn't.

The archive audit

Every warm follow-up gets audited before it gets read. The receiver runs four checks, fast, in the same span they give every other message in a morning inbox.

Do I know this name. Do I know the room. Do I know what we actually talked about. Do I owe this person anything.

Miss one and the email survives on goodwill. Miss two and it's gone, and the sender never learns why, because archiving produces no signal. No bounce, no unsubscribe, no reply. Just silence that reads identically to "busy."

That's the whole asymmetry. The sender experiences a non-reply as a timing problem and schedules a bump. The receiver experienced it as a small failed identification and moved on within seconds. Two people with completely different accounts of the same event.

Archiving produces no signal. The sender learns nothing, so they send the same email again in five days.

Every follow-up guide is written from the wrong chair

Search for follow-up advice and you'll get subject-line length, send-day timing, the personalized opener, the single clear call to action, the three-touch cadence. All of it measured by open rates and reply rates, most of it inherited from cold outreach.

Cold outreach and warm follow-up are different problems wearing the same clothes. A cold prospect has no memory of you to fail. There's nothing to recover, so persuasion is the whole job, and optimizing the subject line makes sense.

A warm follow-up has one job before persuasion is even on the table: re-establish who you are in a mind that has moved on. Advice built for the cold case skips that step entirely, which is why it keeps producing polished emails that die in the audit.

The tooling pushes the same direction. Sequencers, enrichment layers, and most CRMs start their record at "lead," with fields for title, company, and stage. Almost none of them have a field for where you met and what was said, because the software was designed for a pipeline, not a memory.

What the receiver is actually doing, step by step

Watch the sequence closely, because each step is a place the email can die.

First the sender name, read as a string with no face attached. Then the subject line, which in most follow-ups is a generic pleasantry and contributes nothing. Then the first sentence, which is the last real chance to supply an anchor. Then, if and only if the reader has residual curiosity, the LinkedIn check, which almost never rescues a lost memory because a headshot and a job title aren't what the memory was encoded with. People encode context: a room, a moment, a thing that was said.

Then the decision, and the decision is cheap. Archiving costs nothing and feels like inbox hygiene rather than a rejection.

The table below is the audit as a checklist. Read the right-hand column first. That's the one your last follow-up probably tripped.

Run it on a real email

Same sender, same conference, two versions.

Version one: "Hi Dana, great connecting at the conference last week. I'd love to find twenty minutes to show you what we're building at Northgate. Are you free Thursday?" Check one, weak. Check two, fails. Check three, fails. Check four, fails. Archived before the ask is processed.

Version two: "Hi Dana, we talked on the expo floor Wednesday afternoon, right after the demo screen at your booth died and you had to narrate the whole thing from memory. You mentioned your team was rebuilding onboarding in-house and had underestimated it by a quarter. I said I'd send the migration checklist. It's attached. No ask."

Same sender, same product, and the second one survives all four checks before it asks for anything, which is the point: it doesn't ask. The receiver now remembers the broken screen, remembers saying the thing about the quarter, and has a document in hand. The meeting request can come later, from someone whose name now has a room attached to it.

The difference between those two emails has nothing to do with writing skill. Version two contains information version one didn't have.

What changes when you write from that chair

Three things change immediately, and none of them are style choices.

The anchor moves to the first sentence. Most follow-ups bury the where-we-met in paragraph two, behind a pleasantry, which is a paragraph past the archive decision.

The ask shrinks or disappears on the first touch. An ask is a withdrawal against a relationship account that currently has nothing in it.

And the bump gets rewritten instead of resent. If the first email failed the audit, sending it again five days later fails the same audit five days colder.

There's a harder consequence too. You can only write an anchored first line if you still have the anchor, and most people don't. We've written before about the 72-hour memory cliff, the window where the room, the name, and the thing that was said come apart from each other. Past it, the badge photo survives and the conversation doesn't. That's the real reason follow-ups read as generic: not laziness, but an empty file.

The anchor has to be captured, not recalled

By the time you sit down on a Sunday night to write twelve follow-ups, the details you needed are gone. Not fuzzy. Gone, and replaced by a plausible reconstruction that the receiver will read as generic because it is.

Which is why the fix belongs at the moment of the conversation, not the moment of the email. A card scanned with no note attached is a name and a title, the exact two fields that fail checks two and three. The useful capture is the one that takes the room with it: where you were standing, who introduced you, the one sentence they said that you'd never invent.

Met is built around that thirty seconds. Event Mode keeps the capture flow open so you can take a card, add the room and the sentence, and get back to the conversation without leaving the person standing there while you type. Contacts stay in your own iCloud, not on our servers, because the whole point is a private record of what was actually said. When you open it three days later, the anchor is still there, and your first line writes itself.

What we're tracking next

Two open questions worth watching.

First, whether the receiver-side experience shows up as a search behavior at scale. Queries like "someone emailed me after a conference but I can't place them" appear in public forums and get answered badly, usually with advice about how to politely ask. No one seems to be tracking this systematically, so for now it's anecdotal.

Second, whether an anchored first line measurably changes reply rate on warm follow-ups. We don't have that number and won't quote one we haven't run. If you're sending enough follow-ups to A/B it, the test is simple: same ask, same sender, one version with a place and a quoted detail in sentence one.

Until someone runs it properly, the receiver's chair is still the best test available. Forward yourself your own last follow-up, read it cold, and see how far it gets.