Most Decisions Never Reach the Meeting
Meeting-shaped capture records the calendar event and misses the moment a decision actually becomes final, which is usually a thread, a doc reply, or a hallway two days later. This argues for building work memory around the decision itself, not the meeting it was supposed to happen in.
The meeting is the container we can see, not the one the decision lives in
Almost every work memory tool starts from the same assumption: a decision happens in a meeting, so if you capture the meeting you capture the decision. It is an understandable assumption because a meeting is easy to detect. It has a start time, an end time, a list of attendees, and an audio stream a notetaker can join. The moment a decision becomes final has none of those handles. It does not appear on a calendar, it does not send a recorder an invite, and it rarely announces itself.
So capture fires on the thing that is legible rather than the thing that matters. The result is an archive that is dense around scheduled events and thin around the moments where commitments actually formed. That is backwards from what you want three months later, when you are trying to reconstruct not what was discussed but what was decided and when it stopped being reversible.
This is not a new observation about how work runs. When Henry Mintzberg watched what managers actually did rather than what they said they did, the tidy picture of planning in scheduled blocks fell apart. The real pattern was brief, varied, fragmented activity with a heavy preference for verbal and informal contact. Decisions in that world are not tidy events you can point a camera at. They are threaded through many short exchanges, and only some of those exchanges are meetings.
Where the commitment actually forms
Watch a real decision and you see the pattern quickly. The meeting ends with something soft, let us think about it, let us circle back. The actual yes happens two days later in a direct message: fine, ship it. The pricing call surfaces three options and resolves nothing, and the choice arrives the next morning as a two line reply on the shared doc. The hiring debrief is inconclusive in the room, and the offer gets approved in a hallway on the way to lunch.
In every one of these, the meeting has a full record and the decision has none. The transcript captures the deliberation, which was the easy part, and misses the commitment, which was the part you needed. The notes are not wrong about what they contain. They are just pointed at the wrong moment.
Ask anyone to name the last five decisions that actually changed what their team did, then ask where each one became final. A surprising share will land outside a scheduled meeting entirely. That gap is not an edge case. For most teams it is the majority of the decisions worth remembering.
Why meeting-shaped memory is densest where it matters least
There is a quiet irony in a meeting-triggered archive. It fills up around the events that are easiest to capture, and those events are also the most deliberative, the most attended, and therefore usually the least ambiguous when you look back. The hour long planning session leaves pages of notes and is rarely the thing anyone struggles to reconstruct, because ten people were there and the context was spelled out.
The moment that is genuinely hard to reconstruct is the quick one. A reversal typed into a thread. A quiet no that never got a meeting because saying it out loud felt heavier than letting it happen in text. An approval given in passing. None of these looked like a meeting, so none of them got captured, and they are exactly the moments a work record exists to hold.
Krackhardt and Hanson made a version of this point about companies in general: much of the real work happens despite the formal organization, through the informal networks people build across functions to get things done fast. The formal meeting schedule is the org chart of your calendar. Real decisions route around it for the same reasons real work routes around the chart, because the fast path is rarely the official one.
The record that contradicts reality
A gap in memory is a problem you can at least feel. Worse is a record that is confidently wrong. When the decision moves out of the meeting and the meeting notes stay behind, the notes do not go blank. They keep saying what was true at the end of the call: deferred, revisit next week, no owner yet.
Then the decision gets made on Thursday and the team moves on without touching the notes. Anyone who reads the archive later learns the opposite of what happened, and they learn it with the meeting record's full apparent authority. They act on a plan that was abandoned days after the meeting ended.
A missing record makes you go and ask someone. A wrong record lets you skip that step and be confidently mistaken. The second failure is more expensive precisely because it does not feel like a failure until the plan built on it breaks.
A decision is not an event, so stop trying to pin it to one
The deeper mistake underneath all of this is treating a decision as a moment. We do it because a moment is easy to file. But a decision is a small process. An option gets raised. Objections get worked through. Someone commits. Others ratify it by acting as though it is settled. Those steps are spread across channels and across days, and the meeting is often just the middle of the sequence rather than the end.
Capturing only the meeting is capturing one frame of a longer scene and then labeling that frame the whole story. It is why a perfect transcript can still leave you unsure what was decided. The transcript is complete and the decision is still somewhere else, in the reply that came after the room emptied.
Once you see a decision as a process rather than a point, the meeting stops being the natural unit of memory. It becomes one input among several, and not reliably the most important one.
What capturing decisions outside the meeting actually requires
If the commitment forms in a thread, a doc, or a message, then being present for it means seeing work where it happens rather than only where a meeting bot can be invited. That is a different capability from recording calls. It is screen and work context: the reply that said approved, the message that reversed a plan, the doc comment where the real objection finally landed.
This is where botless, always available capture stops being a convenience and becomes the point. A tool that can only join scheduled meetings is structurally blind to the decisions that never had a meeting to attach to. It will always produce a record that is complete about deliberation and empty about commitment.
Driffle was built around this gap. It sees work context and turns screens, meetings, decisions, and follow ups into searchable memory, without a bot sitting in the call. The reason that matters is not that meetings are unimportant. It is that the decision you are trying to remember was usually made in the space between them.
Trying harder inside the meeting does not fix an outside problem
The common response to weak decision memory is to tighten meeting hygiene. Sharper agendas. A decision log filled in during the last five minutes of every call. A rule that no meeting ends without a written outcome. All of that is good practice, and all of it helps the decisions that do get made in the room.
It does nothing for the majority that do not. A decision log written at the end of a meeting inherits the meeting's blind spot exactly, because it is anchored to the same event. The decision that arrives on Thursday is not in Tuesday's log, and no one goes back to add it, because by Thursday the meeting is over and the log is closed.
You cannot close an out of meeting gap with an in meeting ritual. The ritual runs when the meeting runs, and the decisions you are missing are the ones that waited until after.
Make the decision the unit, not the meeting
The change that actually helps is to move the anchor. Instead of organizing memory around meetings and hoping decisions fall out of them, organize it around decisions and treat meetings as one of the places decisions get informed. Anchor a record to the decision itself: what was decided, who committed, when it actually became final, what it reversed, and what it is still waiting on.
In that model the meeting is an input to the decision object, sometimes a major one and sometimes a footnote. The Thursday reply that made it final is another input, and it belongs on the same record. Retrieval then answers the question people actually ask, which is when did we decide this and where, rather than the question meeting notes answer, which is what happened in a particular call.
The difference shows up the first time someone asks why a plan changed. With meeting-anchored memory you get a transcript that says the plan was still open. With decision-anchored memory you get the moment it closed, the reasoning that closed it, and the person who owned it.
A privacy line worth drawing early
Capturing decisions that happen in messages and threads is more intimate than recording a scheduled meeting that everyone knew was being recorded. That is a real difference, and pretending it is not is how good intentions turn into surveillance. The control that matters here is scope and destination, not coverage.
Capture the decision and the reasoning behind it, not a running feed of everyone's private conversation. Keep it with the people who are accountable for the work rather than piping it into a dashboard that managers scroll for its own sake. A record of a decision sits close to intent, and intent deserves a narrower audience than raw activity does.
The goal is memory, not monitoring. A team can tell the difference within a week of using either one, and so can the tool's design. If the output is a searchable answer to when did we decide this, it is memory. If the output is a timeline of who typed what, it drifted.
A check you can run this week
Take the last five decisions that genuinely changed what your team is doing. For each one, trace where it became final. Not where it was first discussed, not where it appeared on an agenda, but the specific moment it stopped being reversible. Count how many of those moments happened inside a scheduled meeting.
If most of them landed somewhere else, in a thread, a doc reply, a direct message, or a hallway, then your meeting notes are not your decision memory, and no improvement to your summaries will make them so. You are polishing a record of the wrong moments.
The point is not to record less in meetings. It is to stop assuming the meeting is where the decision lives. Build memory around the decision, meet it wherever it becomes final, and the archive finally lines up with what actually happened.
Sources
- For the finding that observed managerial work does not match the tidy picture of planning in scheduled blocks, and is instead marked by brevity, variety, and fragmentation with a strong preference for verbal and informal contact rather than written reports - Henry Mintzberg, The Manager's Job: Folklore and Fact (Harvard Business Review, 1990)
- For the point that much of the real work of a company happens despite the formal organization, through the informal networks of relationships that employees form across functions and divisions to accomplish tasks fast - David Krackhardt and Jeffrey R. Hanson, Informal Networks: The Company Behind the Chart (Harvard Business Review, 1993)
FAQ
If decisions usually happen later, are meeting notes still worth capturing?
Yes. The meeting is where the option space and the objections get laid out, and that context is real and worth keeping. The mistake is treating the meeting summary as the decision record when it is only the setup for a decision that will often be made somewhere else.
How is this different from keeping a decision log?
A decision log is a good habit, but it is usually filled in during or right after a meeting, so it inherits the same blind spot. It captures the decisions made in the room and misses the ones that become final in a thread two days later. Anchoring to the decision means recording it wherever and whenever it actually closes.
Doesn't capturing threads and messages feel invasive?
It can, which is exactly why scope and destination matter more than raw coverage. Capture the decision and the reasoning behind it, keep it with the people accountable for the work, and do not turn it into a feed of everyone's private messages. The output should be a searchable answer to when something was decided, not a monitoring timeline.
Why not just ask people to write down decisions themselves?
Because the moment a decision becomes final is the moment people are least likely to stop and document it. They have just committed and moved on to the next thing. A record that depends on that discipline will always be missing the decisions that mattered most. The memory has to form from the work rather than from a habit nobody sustains under pressure.
What single change helps the most?
Move the unit of memory from the meeting to the decision. Ask your tools and your team to answer when and where something was decided, what it reversed, and who owns it, rather than only what happened in a given call. Once the decision is the anchor, the meeting becomes one input instead of the whole record.