Capturing Decisions and Following Up

Running a Meeting ended on a specific note: a good meeting closes by naming what was decided and who does what next. This page is about what happens to those words after everyone leaves the room. Because here is the uncomfortable truth about meetings — the best-run, best-facilitated, most decisive meeting in the world produces nothing if the outcome evaporates the moment the call ends.

You have all felt this. A great discussion, real energy, a clear decision in the room — and two weeks later nobody can quite agree on what was decided, nothing moved, and the same meeting gets scheduled again. The problem wasn’t the meeting. It was that the meeting had no memory and no follow-through. This page gives you the two mechanics that convert talk into change: capture (writing decisions and actions down properly) and follow-up (making sure they actually happen). They are unglamorous. They are also the difference between a team that gets things done and a team that just meets a lot.

The principle: an outcome that isn’t written down didn’t happen

Start from how memory actually works. A room of eight people leaves a meeting with eight slightly different recollections of what was decided. Each one is confident. Each one is partly wrong. There is no shared record, so there is no shared reality — just eight private versions that quietly drift apart until they collide a week later.

The written record fixes this by creating one version everyone can point to. It is not bureaucracy; it is the team’s shared memory, held outside anyone’s head so it can’t be quietly rewritten by wishful thinking or a bad night’s sleep. Two things must survive the meeting for it to have mattered:

• Decisions — what the group agreed to, and enough of why that someone who wasn’t there understands it.
• Action items — the specific next steps, each with a single owner and a due date.

If those two things are captured well and followed up on, the meeting produced change. If they aren’t, the meeting was, functionally, a conversation people enjoyed. Nothing more.

  MEETING  ──►  CAPTURE  ──►  DISTRIBUTE  ──►  FOLLOW UP  ──►  CHANGE
   (talk)      (write it     (send it to     (revisit at
               down while    everyone         the next
               it's fresh)   affected)        meeting)

Miss any link in that chain and the meeting leaks out the gap. Most teams are fine at the first box and terrible at the rest.

Recording decisions so an absent person understands

A decision is the most valuable thing a meeting produces, and it is astonishing how often it is recorded as a fragment nobody can reconstruct later. “Decided to go with option B.” Which option B? Why not A? Who has to know? Six weeks on, when someone questions the choice, that note is useless.

The test for a well-recorded decision is simple: could someone who was not in the room read this and understand what was decided and why, without asking anyone? That means capturing three things.

• The decision itself, stated plainly. Not “discussed pricing” — that’s a topic, not a decision. “We will raise the standard plan from £12 to £15 starting in September.”
• The key reason, in a sentence. This is the part people skip and regret most. Without the why, the decision becomes a naked instruction that invites re-litigation the moment someone forgets the context.
• What was ruled out, if it matters. Recording that you considered and rejected an alternative saves the team from re-opening it in three months.

DECISION LOG — Ward staffing review, 8 July

Decision: Night shift moves from 2 nurses to 3, effective 1 August.
Why:      Incident review showed the two busiest hours (2–4am) were
          consistently understaffed; near-misses clustered there.
Ruled out: Hiring agency staff nightly — too costly and inconsistent.
Decided by: Ward manager, after input from the night team.

Notice this is four short lines. Capturing the why does not mean writing an essay. It means one honest sentence that lets a future reader stand where you stood.

The common trap — recording topics instead of decisions

The most common note-taking failure is writing down what was talked about rather than what was decided. “Discussed the launch timeline.” “Talked through the budget.” A month later these notes are worthless — they prove a conversation happened but tell you nothing about its outcome. Before you write a note, ask: “Did we actually decide something here, or just discuss it?” If you decided, record the decision. If you only discussed, record that it’s still open and who owns bringing it to a close — that’s a decision too.

Action items: owner + action + due date, or it won’t happen

If decisions are the meeting’s memory, action items are its hands. And there is a precise formula that separates an action item that gets done from one that quietly dies:

  ACTION ITEM  =  OWNER  +  ACTION  +  DUE DATE

Drop any one of the three and the item is a wish, not a commitment. Look at what each part does:

• Owner — exactly one person. Not “the team,” not “marketing,” not “we.” When something is everyone’s job, it is no one’s job — each person assumes another will pick it up, and it falls through the floor. One named human is accountable for it moving. That person may delegate the work, but the accountability stays with them. “Shared ownership” is the polite name for an orphan.
• Action — a concrete verb, not a vague intention. “Look into supplier options” will not happen, because there is no visible finish line. “Get written quotes from three suppliers” will, because you can tell at a glance whether it’s done.
• Due date — an actual date. “Soon,” “next sprint,” and “when I get a chance” are not dates; they are permission to defer forever. A date creates a moment where the item is either done or visibly not done, and that visibility is the entire mechanism.

Compare a wish with a commitment:

  A WISH:        "We should improve the onboarding docs."
  A COMMITMENT:  "Priya rewrites the onboarding checklist by Fri 18 Jul."

The second one has a name, a verb, and a date. It can be checked. It can be chased. It has somewhere to live and someone to answer for it. The first is a nice sentiment that will still be a nice sentiment next quarter.

In practice — the cafe order that fixed itself

A cafe manager kept ending team huddles with “we really need to speed up the morning rush.” Nothing changed for months — it was a wish, owned by everyone and no one. She switched to a single action item written on the board: “Sam sets up a pre-batched pastry tray each night before close, starting Monday.” Owner, action, date. It got done the first Monday, because Sam knew it was his and the whole team could see whether the tray was there. The vague complaint had been true for a year; the specific action item took four days.

Distribute promptly — to attendees and to the affected absent

A record that lives only in your notebook is a private diary, not a shared memory. The notes have to travel. Two rules make distribution work.

First, send them promptly — same day if you can. Memory decays fast; a note that arrives while the meeting is still fresh gets corrections (“actually the date was the 20th”) while people still remember. A note that arrives four days later gets ignored, and by then the private versions have already hardened. Prompt distribution is also how you catch your own mistakes — the attendees are your proofreaders.

Second, send them to people who weren’t in the room but are affected by what was decided. This is the step teams forget, and it’s where trust is won or lost. The warehouse supervisor who wasn’t at the planning meeting still needs to know the shift pattern changed. The engineer on holiday still needs to know the architecture decision that reshapes her work. A decision that affects someone who never hears about it isn’t a decision — it’s an ambush waiting to happen. Ask at the close of every meeting: who isn’t here who needs to know this?

Keep the format boring and consistent. People should know exactly where to look:

NOTES — Weekly planning, 8 Jul  ·  Present: A, B, C  ·  FYI: D (on leave)

DECISIONS
  1. Launch moves to 5 Aug (need extra QA week). Ruled out: shipping 29 Jul.

ACTIONS
  1. [Ravi]  Book the extra QA week with the vendor  — by 11 Jul
  2. [Ana]   Tell the sales team the new launch date  — by 10 Jul

STILL OPEN
  1. Pricing for the enterprise tier — Ana to bring a proposal next week.

Three headings — decisions, actions, still open — and anyone can scan it in thirty seconds. Fancy tooling is optional; consistency is not.

The follow-up loop: revisit open items first, every time

Here is the mechanism that makes the whole thing real. The first agenda item of the next relevant meeting is: review the open action items from last time. Not the last item, when everyone’s tired and leaving — the first item, when attention is fresh.

This one habit changes behavior more than any amount of exhortation. When people know that every action item will be read out loud next week, with their name attached, in front of their peers, items get done. Not because of shame, but because commitments now have consequences — the quiet accountability of someone will ask. Without the loop, an action item is a promise made into the void; with it, it’s a promise made to a room that remembers.

The loop is short:

  1. Read each open item aloud: owner, action, due date.
  2. Mark it: DONE  /  IN PROGRESS  /  BLOCKED  /  DROPPED.
  3. Blocked? Name the blocker and who clears it — a fresh action item.
  4. Dropped on purpose? Say so out loud so it's not silently abandoned.
  5. Carry the rest forward. Repeatedly carried-forward items get a hard look.

Point five matters. An item that rolls over three weeks running is telling you something — it’s not important enough to prioritize, or the owner is overloaded, or it’s genuinely stuck. The loop surfaces that instead of letting the item die quietly and dishonestly. “Dropped on purpose” is a fine outcome; “forgotten” is not.

Who should capture, and how lightly

Capture works best when one person owns it for the meeting — a designated note-taker — rather than everyone scribbling their own version. That person isn’t a court stenographer; their job is not to transcribe the discussion but to catch the outputs: decisions, actions, open items. Everything else is noise they can safely let go. In many teams the role rotates, which has a useful side effect — when everyone occasionally has to write the notes, everyone gets better at making decisions crisp enough to be written.

A common worry is that all this recording slows the meeting down. It does the opposite. The discipline of “we can’t leave this item until I can write the decision” forces the group to actually reach a decision rather than drift on. The note-taker’s question — “so what did we just decide, and who owns the next step?” — is one of the most valuable interventions in any meeting. It converts a warm, vague consensus into something concrete before the moment passes. If nobody can answer it, you haven’t finished the item, and it’s better to learn that now than next week.

The evidence — what gets measured moves

A durable finding across management research and practice is that visible, owned commitments with deadlines are completed far more reliably than vague intentions — and that the act of reviewing progress publicly raises completion further still. The exact size of the effect varies by study and setting, so treat specific percentages with caution. But the direction is not in doubt: an action item that will be checked gets done at a much higher rate than one that won’t. The follow-up loop isn’t overhead; it’s the part that makes the meeting pay off.

Why reliable capture builds trust and kills repeat meetings

Step back and see what capture-and-follow-up buys you beyond tidy notes.

It builds trust. When people learn that decisions made in your meetings are recorded accurately, distributed to everyone affected, and actually followed up, they start to believe the meeting. They stop hedging, stop keeping private side-agreements, stop assuming they’ll have to fight the same battle again. A team that trusts its own record can move fast, because it isn’t constantly re-checking what was agreed.

It kills repeat meetings. The single biggest source of pointless meetings is re-deciding things that were already decided but never captured. When there’s no record, every decision is provisional — it can be quietly reopened by anyone who forgot or disagreed, so you meet again to settle it, and again after that. A clean decision log ends this. “We settled this on the 8th — here’s the note and the reason” closes a reopened debate in ten seconds instead of an hour. Reliable capture is, paradoxically, the thing that lets you meet less.

The common trap — the log nobody trusts

The whole system depends on the record being believed. It stops being believed the moment it’s wrong, stale, or incomplete — an action item marked done that wasn’t, a decision recorded inaccurately, half the meetings captured and half not. Once people learn the log can’t be trusted, they go back to their private versions and you’re worse off than before, because now there’s a false shared memory nobody double-checks. The fix isn’t fancier tooling; it’s small and consistent. Capture every relevant meeting the same way, correct errors the moment someone flags them, and never mark something done that isn’t. A modest record people trust beats an elaborate one they’ve quietly stopped reading.

Case study — Amazon’s written record culture

Amazon is well known for a meeting culture built around writing things down — famously replacing slide decks with dense written memos read silently at the start of a meeting (a practice publicly described by Jeff Bezos around 2018). The underlying principle generalizes far beyond that one company: putting the substance in durable writing forces clarity, creates a shared record, and lets people who weren’t in the room understand what happened. You don’t need Amazon’s exact ritual. You need the discipline behind it — that decisions and next steps live in writing, not just in the memories of whoever attended. Treat the specific format as optional and the principle as not.

How it connects

Capture is where a meeting connects to everything outside it. The action items you record here feed straight into how work actually gets tracked and prioritized — the concern of the project-management and ways-of-working parts later in the book. The follow-up loop is also the backbone of the recurring meetings in the next page, Making Standups, Planning, Retros, and Reviews Useful — each of those rituals lives or dies on whether last time’s commitments get revisited. And clean decision records are what make disagree-and-commit enforceable: you can only hold people to a decision the group can actually point to.

Try this

For your next meeting, do just the capture-and-follow-up part deliberately. In the last five minutes, read back every decision and every action item out loud in the form owner + action + due date, and get a nod on each. Send the notes the same day, using the three headings — decisions, actions, still open — and copy in anyone affected who wasn’t there. Then open your next meeting with that list as item one. Do this for three meetings in a row and watch how many things that used to slip now actually land.

Reflect

1. Think of a decision your team made recently. If someone who wasn’t there read the notes, would they understand what was decided and why — or is it just a topic label?
2. Look at your last set of action items. How many had a single named owner, a concrete action, and a real date? What happened to the ones that didn’t?
3. When you distribute notes, who reliably gets left off — and has a decision ever “ambushed” someone because they never heard about it?
4. Does your team start meetings by reviewing last time’s open items, or does follow-up depend on people remembering? What changes when the loop is missing?
5. Can you name a meeting that keeps recurring because the same thing gets re-decided each time? What would a clean, pointed-to record have prevented?

Show reflections

1. The tell is whether the note captures a decision (a plain statement of what will happen) plus one sentence of why. If your notes are topic labels — “discussed X” — an absent reader learns nothing, and neither will you in six weeks. Aim to make every decision reconstructable without asking anyone.
2. Sort them honestly into “had all three” versus “missing one.” You’ll usually find the ones that stalled were missing an owner or a date. That’s not a motivation problem; it’s a capture problem, and it’s fixable with a formula rather than more nagging.
3. The people left off are almost always the affected-but-absent — the person on leave, the adjacent team, the shift that wasn’t in the room. Naming a real ambush makes the cost concrete and gives you a habit: ask “who isn’t here who needs to know?” at every close.
4. If follow-up depends on memory, it’s not a loop — it’s luck, and it decays. When the loop is missing, items quietly die and trust in the meeting erodes, because commitments turn out not to be real. Putting “review open items” as agenda item one is the cheapest fix in this whole part.
5. Repeat meetings almost always trace back to a decision that was never captured, so it stayed reopen-able. A record you can point to (“settled on the 8th, here’s the reason”) ends the loop. If you can name such a meeting, you’ve just found where a decision log would buy back real hours.