A template can't remember the conversation. That's the part that mattered.

It was great meeting you at the conference

You've gotten this email. Probably this week. Maybe twice. The subject line is your name. The first sentence is a variant of "It was great meeting you at the conference", and you can't remember which conference, or which meeting, or who is sending it. You scan the rest of the message for a clue. There usually isn't one.

This is the follow-up template doing its job. It is, by every measure the category claims to optimize for, a successful artifact: it's polite, it's grammatical, it arrived on time, it has the right structural beats. And it's about to get archived in three seconds, exactly the way every other template-shaped follow-up that recipient gets gets archived.

The template wasn't supposed to fail. The category, the listicles, the SaaS canned-message features, the "50 follow-up templates that close deals" PDFs every sales-tech vendor publishes, has been iterating on this pattern for fifteen years. The prose has gotten cleaner. The CTAs have gotten softer. The personalization tokens have gotten more numerous. The recipient still archives in three seconds. Something structural is wrong, and it's not solvable by writing better templates.

The thing a template can't carry

Here's the structural problem: a follow-up has exactly one job that matters, and it's the one job a template can't do.

That job is to prove you remember the conversation. Not the event. Not the venue. The conversation. One sentence, "You mentioned your team was figuring out the FY27 procurement timing," does the work that fifteen template variations can't. The recipient reads that sentence and thinks: this person was actually there. This person was actually paying attention. The relationship resumes from where it left off Wednesday afternoon, not from cold-outreach baseline.

This is the part templates were built to skip. Templates exist precisely because writing the personal sentence is hard, slow, and doesn't scale. The pitch of the entire template category is "send the same well-written message to twenty people." The math of the relationship is: you can send the same well-written message to twenty people, and twenty people will archive it in three seconds, because none of them got the sentence that proves you were there.

The category solved a problem nobody actually had, the prose-quality problem, and skipped the problem that mattered.

Four ways the category fails at this

Across the follow-up-template SERP and inside every canned-message feature in the major contact apps, four failure shapes recur. Each one is recognizable enough that you'll map your own most-recent template to one. None of them solve the prove-memory problem, because none of them can.

1. Generic-opener templates. "It was great meeting you at the event." The opener is the highest-volume slot in the category, every listicle leads with one. The structural problem is that "great meeting you" is a phrase the recipient has read a dozen times this month, in identical syntax, from senders they don't remember. Generic openers actively flag the message as one of many. The recipient's pattern-matcher is faster than the prose; the message lands as cold outreach in formal clothes.
2. Slot-fill templates. "Hey [name], it was great chatting about [topic] at [event]." The category's response to the generic-opener problem was to insert variables. The slot-fill template is structurally identical to the generic-opener template, same skeleton, same beats, same archive rate. The recipient doesn't read the [topic] slot as a sign you remember; they read it as a sign you found the merge field. The personalization tokens make the template feel less personal, not more.
3. Asking templates. "I'd love to schedule fifteen minutes this week to discuss next steps." The asking template skips the proving-memory step entirely and goes straight to the request. From the recipient's seat, this lands as a stranger asking for time, because the message hasn't earned a reason to feel non-stranger. Conversion rate from cold ask is approximately the conversion rate from cold outreach, which is the floor.
4. Recap templates. "As we discussed, your team is exploring options around [project], and I think our solution can help by [pitch]." The recap is the most-suggested fix in the category, "include a summary of what you talked about," and it's the closest of the four to working. It also fails, because the recap reads as transcript, not memory. A summary written from notes feels different than a single specific reference written from the moment. The recipient can tell which one they're looking at.

If you've sent a follow-up in the last quarter, you sent one of these four. The reason none of them work is structural: all of them are formats that survive the absence of the conversation. The thing that lands is the format that requires the conversation to write at all.

What the working ones look like

The follow-ups that work all look roughly the same. Two paragraphs. Three or four sentences total. The first sentence is the proof, one specific thing the recipient said, paraphrased loosely enough that it sounds like memory rather than transcript. The second sentence is small. Not "let's schedule a call to explore partnership." Something the recipient can answer in twenty seconds: "That made me think, does your team actually use [specific tool / process / framework] in practice, or is it more aspirational right now?"

A follow-up shaped like this does several things templates can't. It demonstrates the conversation happened, in a form that can't be merge-fielded. It gives the recipient something to push against, a question, not an ask. It compresses the response surface from "reply-or-don't" to "answer the question or don't," which is a smaller decision and lands more replies. It reads as a person, because it could only have been written by someone who was there.

The length isn't an accident. Long follow-ups are themselves a category failure, the recipient pattern-matches a long message as effort-spent-on-format, not effort-spent-on-them. The working format is short, specific, and asks a small question. Templates can produce all three of those traits structurally. None of them can produce the specific.

Why the category keeps producing templates anyway

If templates don't work, why does the category keep producing them, and why do they keep getting downloaded? Both questions have honest answers.

Templates keep getting produced because they're easy to make and easy to sell. A 50-template PDF is one weekend of writing for the marketer; the SEO graph around "follow-up email template" is enormous and durable; every sales-tech vendor wants a piece of it. The economics of producing follow-up content rewards quantity-of-templates over quality-of-outcomes. A vendor whose templates lead to 3% reply rate and one whose templates lead to 1% reply rate look identical in the SERP; the metric the recipient cares about is invisible to the buyer at template-selection time.

Templates keep getting downloaded because the alternative, actually writing the specific sentence for sixty contacts, looks impossible. From inside the post-conference Tuesday, you have a stack of cards, no notes, no memory of which conversation was which, and a closing warm-window. Templates feel like the only way to even ship the queue. The math is brutal: at sixty contacts, three minutes of personalization each, you're three hours in before you're done, and the warm-window closed eight contacts ago.

The category isn't producing templates because nobody noticed they don't work. It's producing them because the alternative, capturing the specific sentence at meeting-time and using it at follow-up-time, is hard to do without an app built around exactly that flow. The structural fix isn't a better template. It's a different shape of tool.

What it looks like when capture and follow-up are the same flow

If the working follow-up requires one specific sentence the recipient said, then the follow-up's quality is decided at capture-time, not at compose-time. By the time you sit down on Tuesday morning to write, the moment to capture the sentence has passed. The card in front of you has a name and a title; the conversation is gone.

The shape of tool that solves this isn't a faster template engine. It's a capture flow that takes the sentence at meeting-time, when the conversation is still in your head, when the right question to ask in three days is obvious, and stores it next to the contact, not in a separate notes app. By Tuesday, when the queue needs to ship, the sentence is already attached. Drafting becomes editing; editing fits in the warm-window.

This is the design choice Met is built around. The capture step takes a tag, a line of context, and a follow-up question while the moment is fresh. By Tuesday, the draft is one tap away, referencing what was said, asking the small question, and the queue actually ships before the warm-window closes. It's not a template feature with better merge fields. It's the absence of the template stage entirely.

If the alternative to templates is "have actually captured the conversation," the design problem is no longer how to write better follow-ups. It's how to make the right capture frictionless enough to happen during a thirty-second conversation between booth visits.

Try the framework on one contact

If this lands and you want to test the framework against your own most recent follow-up, the cheapest way is the 72-Hour Follow-Up Generator, a free, no-signup tool that takes one contact at a time. Tell it who, tell it the one thing they said, tell it what you'd want next. It drafts a follow-up shaped like the working ones above. It will not look like a template. That's the point.

If the generator works for one contact, the app version is what handles all sixty without losing the specific.