Most agendas are a list of topics, which is why most meetings drift. A topic tells you what will be talked about and nothing about when it is finished. Every item needs:
The type field does the most work. A decision item needs the decision-maker present and the options prepared. A discussion item needs the people with context and no expectation of resolution. An information item needs no meeting at all in most cases — it needs a document.
| Item | Type | Owner | Min | Outcome |
|---|---|---|---|---|
| Silent read of the proposal | Info | All | 8 | Everyone has read it |
| Pick launch date: 14 or 28 Oct | Decision | Priya | 10 | Date chosen and recorded |
| Pricing objections from beta users | Discussion | Sam | 12 | Objections listed and ranked |
| Owner for the migration risk | Decision | Priya | 5 | One name assigned |
| Actions and owners read back | Info | Chair | 5 | Written actions confirmed |
| Total | 40 | In a 45-minute slot | ||
Two deliberate choices there. The total is 40 minutes in a 45-minute booking, leaving genuine slack — Parkinson law is real, and an agenda that fills the slot exactly always overruns. And the meeting opens with silent reading rather than a presentation, which is the practice popularised at Amazon: circulating a written document and reading it in the room guarantees a prepared audience without depending on anyone having done homework, and it surfaces sharper questions than a slide walkthrough.
Put the purpose at the top in one sentence, phrased as what will be different afterwards. Not discuss Q4 planning but choose between the two Q4 scopes so engineering can start on Monday.
If you cannot write that sentence, the meeting is not ready. That test alone cancels a surprising share of recurring meetings, and cancelling them is worth more than any improvement to how they run.
An agenda is necessary and nowhere near sufficient. Research on meeting effectiveness has found that simply having an agenda is a weak predictor of whether a meeting goes well — what matters is whether it is used, whether the chair actually enforces the timeboxes, and whether people were prepared. A beautifully structured agenda in a meeting with the wrong attendees, an absent decision-maker, or an unresolvable political disagreement will fail exactly as thoroughly as no agenda at all.
Timeboxes also cut both ways. Cutting off a discussion that is genuinely converging on something important, because the clock says twelve minutes, is worse than overrunning. The chair judgement is the mechanism; the agenda is only the instrument.
And an agenda cannot make a meeting necessary. If every item is type information, send the document. If a decision has already been made elsewhere, do not stage a discussion about it — people notice, and it costs more trust than the meeting saves time.
25 or 50 minutes rather than 30 or 60. It leaves transition time between back-to-back meetings, which is otherwise unpaid and comes out of focus. Shorter defaults also force sharper agendas.
Yes, with the pre-read attached, at least a day ahead. Sent five minutes before, it is documentation rather than preparation.
Park it visibly — a named list at the end — and either assign it a follow-up or a slot next time. Parking works because it acknowledges the point without surrendering the schedule.
No. Everything you type stays in your browser and nothing is transmitted or saved.
Every tool comes with a written guide, and every category is one click away.