Meeting minutes: why direction is so important
Francisco González
System Architect • @janobourian
Janobourian
Being there does not imply you know why, and even knowing why does not imply you know where you have to go.
Wait, do we need meeting minutes?
Organizing anything sounds complex, and it is. It demands respect. Yet in software engineering organizations that boast about microsecond latency and zero-downtime clusters, human coordination is treated with casual carelessness, as if human time can be wasted without consequence.
Managers drop recurring calendar invites onto dozens of calendars with empty descriptions. Six engineers log into a video call, listen to forty minutes of unstructured chatter, nod in vague agreement, and disconnect. The second that call drops, the entire conversation vaporizes into thin air because nobody documented anything, and all the time spent in the meeting was wasted. I say again: "wasted." Don't get me wrong—I don't hate meetings. They can be productive when they have clear direction and proper organization.
A meeting is not an informal fireside chat; it is an expensive synchronous operation that pauses company delivery while the payroll clock keeps ticking. The mental design behind any meeting must be a disciplined matrix of parallel reasoning: Why are we pulling these minds together? Who actually owns the problem? What concrete output must exist before we leave? Every "W" question acts as a rudder. Without that mental design, the meeting drifts aimlessly, burns cognitive stamina, and disrespects every participant's finite working hours.
Let's be honest: your lead never says, "Hey, this meeting modifies our deadline, so you'll have more time to finish your work." They never say, "Hey, this meeting just blew up my schedule." Instead, they say, "This meeting is important," and you feel obligated to attend and participate—only to end up wasting your time.
The Same Person With and Without Meeting Minutes
To understand the visceral impact of meeting minutes, look at the exact same senior engineer in two parallel universes.
Alex sits through a 50-minute architectural debate about database replication and API rate limiting. Three people talk over each other. Someone says, "Maybe we can cache user tokens in Redis." People nod. The call ends abruptly at 3:00 PM.
Three days later, Alex submits a pull request using Redis. The database lead comments: "Wait, we never agreed to add Redis; we said we'd use DynamoDB DAX." A 45-message passive-aggressive Slack war erupts. Nobody remembers who said what. Alex spends his entire Friday rewriting code, second-guessing his role, and battling friction. Delivery stalls.
Alex attends the same discussion. As the meeting closes, the facilitator writes down the consensus in real time. Ten minutes after the call, a four-line record hits the team channel: Decision: Redis Cluster approved over DAX due to latency budget; Owner: Alex; Target: PR by Thursday.
When an engineer later asks, "Why aren't we using DAX?", Alex drops the link to the minute. The debate dies in five seconds. Alex builds with total clarity, hits his sprint target, and sleeps peacefully on Friday night. The organization moves forward with zero memory decay.
The Real Trade-Offs: Pros and Cons of Meeting Minutes
Writing meeting minutes requires deliberate effort. It is not free. Acknowledging the trade-offs is essential to doing it right without drowning in administrative rot.
The Costs & Friction ("Cons")
- Active Mental Overhead: You cannot zone out or multitask. Capturing minutes requires continuous active listening, synthesizing messy human debate into crisp conclusions.
- Immediate Time Tax: It demands 10 to 15 focused minutes immediately after the call to clean, verify, and post the notes while the context is fresh.
- Loss of Ambiguity: For political actors who love vague commitments and zero personal liability, written minutes eliminate deniability. You are on the record.
- Risk of Bureaucratic Bloat: If done by an untrained amateur, it can turn into an unreadable, multi-page transcript that nobody opens.
The Exponential Leverage ("Pros")
- Asynchronous Liberation: People who only needed informational updates can skip the call completely. A 60-second read saves an hour of their lives.
- Immunity to Memory Decay: Human memory starts degrading within hours. A written artifact locks decisions in amber.
- Single-Owner Accountability: Every task has exactly one name and one firm deadline attached. No diffusion of responsibility.
- Instant Dispute Termination: Kills the dreaded "zombie meetings" where teams repeatedly debate the exact same problem every three weeks.
- Architectural Audit Trail: Onboards new engineers months later by explaining not just how the system is built, but why alternative paths were rejected.
The Mindset of Meeting Minutes
Most developers hate meeting minutes because they confuse them with court transcripts. They imagine an executive assistant writing down every sneeze, sigh, and digression. That is useless garbage.
Meeting minutes are not transcription. They are an architectural steering mechanism. Their sole purpose is to convert chaotic human conversation into directional vectors. If a sentence does not change team behavior, clarify an assumption, or define an action, it does not belong in the document.
To build this mindset, run the five fundamental questions in parallel:
- Why: What specific dilemma provoked this gathering? If you cannot formulate the "why" in one crisp sentence, the meeting should have been an email.
- Who: Who attended, who was absent, and most importantly, who owns the resulting action items? If three people share an action item, nobody owns it.
- What: What was resolved? Not "we discussed auth tokens," but "we decided to revoke tokens via Redis denylists instead of database lookups."
- When: Hard dates. "Next week" or "soon" is corporate code for never. An action item without an ISO date (YYYY-MM-DD) or sprint milestone is merely a wish.
- Where: Where is the system or project moving next? What is the downstream blast radius on the architecture, the deployment pipeline, or upcoming epics?
Blueprint: Guide to Creating Sharp Meeting Minutes
Follow this direct, repeatable operational framework to write minutes that engineers actually respect and read:
# [SYNC] Auth Gateway Architecture Refactor
Date: 2026-10-10 | Facilitator: @janobourian | Duration: 25 mins
Attendees: Alex (Backend), Elena (SecOps), Carlos (DevOps)
## 1. Context & Objective
Resolve P99 latency spikes on token validation under peak traffic.
## 2. Key Decisions Made
- ADOPT: Redis cluster for in-memory session cache (Rejects DAX due to AWS cost model).
- DEPRECATE: Synchronous database validation on public endpoints.
- ACCEPT: Eventual consistency boundary of 250ms during token revocation.
## 3. Action Items (Owner + Task + Deadline)
- [ ] @carlos: Provision Terraform module for Redis cache cluster. [Due: Oct 14]
- [ ] @alex: Update Auth middleware to read from cache layer. [Due: Oct 16]
- [ ] @elena: Security audit on Redis encryption-in-transit setup. [Due: Oct 18]
## 4. Open Questions / Blockers
- Awaiting confirmation from Finance regarding reserved node pricing.
Notice the discipline: zero filler, zero transcript bloat. Anyone reading this in thirty seconds knows the state of the world, who is on the hook, and what to test next.
Best Practices: How to Evaluate Good vs. Bad Minutes
How do you evaluate whether a meeting minute succeeded or failed? Use this objective score card:
| Dimension | Bad Meeting Minute | Good Meeting Minute |
|---|---|---|
| Length & Depth | Five pages of raw dialogue; chronological play-by-play of who said what. | One page or fewer. Read time is under 90 seconds. Focuses purely on outputs. |
| Action Items | "The team will check on logs." Vague, communal, unassigned, and untimed. | "[Name] will write Terraform script by [Date]." Singular owner, hard deadline. |
| Decisions | "Discussed caching options." Zero commitments recorded. | "Chosen Redis; rejected Memcached due to clustering requirements." |
| Dispatch Speed | Sent three days later as a PDF attachment. Nobody opens it. | Posted directly in the primary project channel within 15 minutes of call end. |
| The Ultimate Test | An engineer absent from the call has to ping someone to ask: "So what happened?" | An absent engineer reads it in 45 seconds and immediately knows how their work is affected. |
But should I prefer bad meeting minutes over no minutes at all?
There is no simple answer to this question; it depends on your standards. If you want to be an effective leader, the truth is that bad meeting minutes can be worse than no minutes at all—they create a false sense of alignment while steering the team in the wrong direction.
Amateurs measure work by movement; masters measure work by direction. Anyone can call a meeting, fill an hour of calendar time, and generate the soothing illusion of activity. That is easy. It is cheap. It requires no courage and no rigor.
Real leadership begins the moment the talking stops. A meeting without minutes is just a conversation that evaporated into the ceiling. Meeting minutes are not bureaucratic paperwork for secretaries; they are the contract of execution. They define the vector: magnitude and direction.
Before you join or host your next call, remember: being there does not mean you know why you came, and knowing why does not mean you know where you are going. Demand direction. Write down the decisions. Assign the owners. If you cannot capture the direction in writing, cancel the meeting and give your engineers their time back.
References
- Grove, Andrew S. 1983. High Output Management. Random House.
- Drucker, Peter F. 1967. The Effective Executive: The Definitive Guide to Getting the Right Things Done. Harper & Row.
- Winters, Titus, Manshreck, Tom, and Wright, Hyrum. 2020. Software Engineering at Google: Lessons Learned from Programming Over Time. O'Reilly Media.
- Newport, Cal. 2016. Deep Work: Rules for Focused Success in a Distracted World. Grand Central Publishing.