The Email That Saved My Career Was Four Lines Long

I nearly didn't send it. It felt fussy, the kind of thing a person sends when they don't trust anyone, and I sat looking at it for a good ten minutes on a Thursday evening wondering if it made me look like I was covering myself. I sent it anyway. Seven months later it was the only reason a very expensive argument lasted twenty minutes instead of six weeks.

Context: a data centre migration for a housing association. About 180 workloads, a hard exit date on the old facility because the lease was ending, and a client architect who wanted the legacy repairs reporting stack moved as a lift and shift rather than rebuilt.

We'd said on the call that lift and shift meant the reporting performance would be roughly what it was today, possibly worse, because the storage tier in the new site was configured differently. He said fine, that's understood, we'll take it and tune later.

That evening I wrote:

Subject: Reporting stack, decision from today's call (12 June)

Confirming what we agreed this afternoon. Reporting stack moves as a lift and shift rather than a rebuild, to hold the December exit date.

Accepted consequence: report run times after migration will be at least as long as today and may be longer on the new storage tier. Tuning is out of scope for this phase.

Decision made by Martin, agreed by me. Alternative considered was a rebuild, rejected on timeline.

If any of that doesn't match your understanding, tell me by Friday and I'll correct it.

Four lines of substance. Two minutes to write.

In January the repairs reporting runs were about forty minutes slower, the estates team escalated, and somebody senior asked, reasonably, why we'd migrated a system in a state that made it worse. The answer took one search and one forward. No investigation, no defensive meeting, no six weeks of two organisations reconstructing a conversation from memory. Martin, to his credit, replied to the forward with "yes, that was my call and it was the right one at the time, let's fund the tuning now."

That last sentence is what a good record actually buys you. Not exoneration. A conversation that can move forward.

Length is the enemy here, and new leads get this wrong in the same direction every time. They write the full narrative: what was discussed, who said what, the options explored, the surrounding context. It reads like minutes. It takes twenty-five minutes to write, so they stop doing it after the third one, and it takes four minutes to read, so nobody corrects it.

A confirming email has exactly one job: to record the decision in a form the other party can disagree with cheaply. Everything that doesn't serve that is weight.

Four elements, and you can write them in the lift on the way out:

What was decided, in plain terms, dated. Not "we discussed options around the reporting stack". Decisions are verbs with objects.

What it costs, stated as an accepted consequence rather than a warning. This is the part that stops being obvious in month seven. Everyone knew the trade-off at the time. Nobody remembers it afterwards, including you.

Who decided, by name. Not "it was agreed". Agreement has no owner and that is precisely why people write it that way.

An invitation to correct it, with a date. This is what converts your version into a shared one. Silence against an explicit deadline is a very different thing from silence against nothing.

Two failure modes on recipients, opposite directions, both common.

The first is sending it to one person. You and the client architect both have it, and when he moves to another role in September, your only corroboration leaves the building with him. Send it to the people whose work changes because of the decision. On the migration, that meant the client architect, his delivery manager, our account lead, and the engineer from our eight-person tooling team who'd be doing the work.

The second is the defensive copy list. Adding the client's CTO to a routine decision note is not thoroughness, it's a signal, and everyone reads it as one. You've told the room you expect to need protection from these people. The relationship cost is real and it arrives immediately.

The test I use: would I be comfortable if the recipient asked me, face to face, why I copied each of these people? If the honest answer for any name is "so you'd be on the hook too", take them off and have a direct conversation instead.

Tone does most of the work. The same four facts can read as a collaborative record or as a legal filing, and the difference is almost entirely in how you frame the ask at the end.

"Please confirm your acceptance of the above" reads like you're building a case. "If any of that doesn't match your understanding, tell me by Friday and I'll correct it" reads like you're trying to get it right. They achieve the identical thing.

Two other habits worth stealing. Say it in the room first: "I'll write this up tonight so we've both got the same version" takes three seconds and removes any element of surprise. And when someone does come back with a correction, thank them and send the corrected version, because that exchange is worth more than the original email. It proves the record is a shared artefact rather than your account of events.

One more thing, and this is where the sender's motive shows. Write the confirming email when the decision goes your way too. If your notes only exist for decisions you disagreed with, the pattern is visible from space, and the next time you send one everybody knows exactly what it means.

Think back to your last meeting where something actually got decided. Not a status call, a real one.

Write the four lines. What was decided and when, what it costs, who decided it, and the invitation to correct it with a date on it. Send it to the people whose work changes because of it. Keep it under a hundred words and do not attach anything.

Then do it again after your next decision meeting, and the one after that. In eleven months you won't remember writing any of them, which is the entire point.