Key takeaways
- →A full calendar is a symptom - the causes are usually unclear decision rights, low trust in written updates, or the same information being distributed by mouth.
- →Audit two real weeks before changing anything; opinions about the calendar are consistently wrong in both directions.
- →Most of the load sits in recurring meetings, and most recurring meetings were designed for a problem that has since been solved.
- →Blanket rules like no-meeting Fridays relieve pressure without removing the cause, and the load reappears on the other four days.
- →The right question for each meeting is not 'is it useful' but 'what would break within a month if it stopped'.
- →Cutting a meeting only works if you replace the job it was doing - a written update, an asynchronous decision, or a shorter meeting with fewer people.
Meeting overload is a symptom, not a diagnosis, and treating it as a diagnosis is why most attempts to fix it fail. Teams declare a meeting amnesty, delete a dozen recurring invites, enjoy about three weeks of space, and then watch the calendar refill - because the thing that generated those meetings in the first place was never identified, let alone changed.
Underneath a crowded calendar there are usually only a handful of mechanisms at work: decisions that nobody is authorised to make alone, information that people do not trust unless they hear it said, work that is coordinated by conversation because it is not written down anywhere, and status reporting that exists to reassure someone rather than to inform anyone.
This guide shows you how to audit two real weeks of calendar, how to map each meeting to the mechanism generating it, which meetings to kill, shrink or keep, what has to replace the ones you cut, and why the popular blanket fixes tend to move the load around rather than remove it.
Why 'too many meetings' is the wrong problem statement
Nobody objects to meetings in the abstract. What people object to is fragmentation - a day chopped into forty-minute gaps between commitments, none of them long enough for work that requires holding a problem in your head. Two three-hour blocks of meetings and four hours of clear time is a better day than five scattered half-hours, even though the meeting total is far higher.
That distinction matters because it changes the fix. If the complaint is volume, you cut meetings. If the complaint is fragmentation, you consolidate them - same total hours, arranged so that clear time exists at all. Most teams that measure it discover their real problem is the second, and cutting a few invites at random does nothing for it.
The second thing hidden inside 'too many meetings' is that the load is not evenly distributed. Managers, anyone in a coordinating role and anyone recently joined typically carry two or three times the meeting hours of the median person on the team. A blanket policy applied evenly across everyone will barely touch the people who are actually drowning.
The mechanisms that generate meetings
When you look at a calendar meeting by meeting and ask why it exists, the answers collapse into a small number of underlying causes. The useful move is to name the cause rather than to judge the meeting, because the same meeting can be entirely reasonable in one team and pure overhead in another.
| What people say about it | Usual underlying cause | What actually removes the load |
|---|---|---|
| 'We need everyone in the room to decide' | Decision rights are unclear, so consensus is used as cover | Name a single decider per decision and a consult list; the meeting shrinks to three people |
| 'This is our weekly sync' | Coordination happens by conversation because nothing is written | One shared written status that is actually maintained, plus a fortnightly meeting for exceptions |
| 'People miss things if we only post it' | Low trust in written channels, often earned by unreadable posts | Fix the writing and the channel discipline before adding a meeting to compensate |
| 'Leadership wants visibility' | Status reporting upwards dressed as a team meeting | A short written report to the person who needs it, not an hour for twelve people |
| 'It is the only time we talk' | A genuine relational need, badly served by a status agenda | Keep the time, drop the status content, make it a real conversation |
| 'We always have it' | A meeting designed for a problem that has since been solved | Cancel it for a month and see whether anything breaks |
| 'Everyone needs to hear this at once' | One-way broadcast that could be read in four minutes | Send the update; use the live slot for questions only |
Run a two-week audit before you change anything
Do not skip to the fixing. Every team believes it knows where its meeting time goes and every team is wrong, usually in the same direction: they overestimate the cost of ad-hoc invites and dramatically underestimate the cost of the recurring meetings that have been in the calendar so long that nobody sees them any more.
The audit takes about ninety minutes for a team of twenty and it produces the one thing you need for the conversation afterwards - specifics. 'We have too many meetings' generates a discussion. 'Our four recurring meetings cost the team eleven person-days a month, and three of them are tagged inform' generates a decision.
Export two real weeks, not a typical one
Pick the last two completed weeks, including the messy one. Teams always want to exclude the atypical week and the atypical week is usually where the truth is - it shows what the calendar does under pressure.
Count hours per person, not meetings per team
Total the hours for each individual and sort the list. The spread is the finding. If your busiest person carries three times the median, you have a distribution problem that no team-wide rule will touch.
Tag every meeting with one of four purposes
Decide, inform, coordinate, or connect. Anything that cannot be tagged with exactly one is a candidate for splitting, because meetings that try to do all four do none of them well and always overrun.
Separate recurring from one-off
Total each bucket. In most teams the recurring block is between half and three quarters of all meeting time, which means the one-off invites everyone complains about are not where the hours are.
Note attendee counts against purpose
A decision meeting with fourteen people is not a decision meeting. Flag every 'decide' meeting above eight attendees and every 'inform' meeting of any size - those are your two richest seams.
Ask three people what they would cancel
Do it privately and in writing, not in a meeting about meetings. You will get honest answers you cannot get in a group, because cancelling someone's meeting in front of them is socially expensive.
Measure the gaps, not just the blocks
Count how many clear stretches of ninety minutes or more each person had across the fortnight. For some people the answer is zero, and that number will change the conversation faster than any total.
The numbers worth putting on a slide
These are working defaults rather than research findings - thresholds that make the audit produce action instead of discussion. Set them before you look at the data so you are not choosing the line after you have seen where it falls.
The kill, shrink or keep test
Run every recurring meeting through the same short test. The critical question is not whether the meeting is useful - almost everything is useful to somebody - but what would visibly break if it stopped for a month. If nobody can name a consequence, you have your answer.
- ✓Somebody can name a specific thing that breaks within 30 days if this stops
- ✓It has one purpose: decide, inform, coordinate or connect - not a blend of all four
- ✓There is a named owner who prepares it, not a rota of whoever remembers
- ✓The agenda exists before the invite goes out, every time
- ✓Everyone invited either contributes or has a decision at stake
- ✓It ends with something written down that outlives the call
- ○The information in it could not be read faster than it can be said
- ○It has been reviewed in the last six months rather than inherited
What has to replace the meeting you cut
This is where most meeting reduction quietly fails. Every recurring meeting is doing some real job, even a badly performing one, and deleting it without replacing that job creates a gap that gets filled within weeks - usually by a new meeting with a different name and the same attendee list.
Decide what replaces it before you cancel it, and say so in the cancellation message. 'We are stopping the Tuesday sync; status now lives in the shared doc, updated by Monday lunchtime, and anything blocked gets flagged in the channel with a name attached.' That is a replacement. 'We are stopping the Tuesday sync' is a vacuum.
Written status with a deadline
Replaces the weekly round-the-table. Works only if it has an owner, a fixed slot in the week and a length limit - three bullets, not an essay.
Asynchronous decision with a clock
Post the options, the recommendation and a deadline. Silence counts as assent. Reserve the meeting for the cases where people actually disagree.
Same meeting, half the people
Cheapest fix available and almost always available. Cut a twelve-person decision meeting to five and the hour is genuinely productive for the five.
Broadcast plus a live Q&A
Send the update in writing, then hold twenty minutes for questions instead of an hour of presentation that people could have read faster.
Fortnightly with an exceptions agenda
Halve the frequency and change the agenda to 'what is blocked or off-track'. Most weekly coordination meetings survive this comfortably.
Keep it, drop the status content
For the meeting that exists because it is the only time the team talks. Keep the slot, remove the reporting, and let it be the conversation it already was.
Why the popular blanket fixes disappoint
No-meeting Fridays, thirty-minute maximums and a company-wide invite freeze all share the same shape: they constrain the symptom without touching the mechanism. The work that generated the meetings still has to happen, so it moves. A team with genuinely unclear decision rights and a no-meeting Friday will hold the same meeting on Thursday afternoon, and the Thursday will be worse than it was.
That does not make them worthless. A protected day is a real improvement in fragmentation for people who had none, and a shorter default length does force agendas to tighten. Treat them as relief while you fix the cause, not as the fix, and say that out loud - otherwise the policy gets judged on whether the calendar looks emptier in March, which it will not.
The blanket fix that does tend to hold is a structural one: making it impossible to send an invite without an agenda and a stated purpose. It is mildly annoying, it applies to everyone equally, and it kills a specific category of meeting - the one somebody booked because they were not sure how else to make progress.
Be sceptical too of cutting meetings as a headline number. A team that halves its meeting hours and doubles the volume of messages, documents and one-to-one clarifications has not reduced its coordination cost; it has moved it somewhere that is harder to see and harder to measure. The point is not fewer meetings. The point is fewer hours spent in rooms where nothing that needed the room happened.
The meetings worth defending
Top Tips
- Decision meetings with a named decider, a real disagreement and eight people or fewer - these are the highest-value hours on any calendar and they are often the first casualty of a blunt cull.
- Sessions where people build something together in the room: a plan, a ranked list, a set of risks. If the artefact would not exist without the meeting, the meeting is doing work no document can replace.
- Difficult conversations of any kind. Compressing conflict, feedback or bad news into writing to save an hour reliably costs three hours later, plus a relationship.
- Genuine connection time for distributed teams, provided you stop pretending it is a status meeting. The failure mode is not that it exists - it is that it carries an agenda that crowds out the thing it is for.
- Anything where the group's collective judgement is the point: estimating, prioritising, sense-checking a risky assumption. Independent input gathered and compared in the room beats a serial email chain by a distance.
- Onboarding and induction sessions in the first fortnight. New joiners have no written context to fall back on, and the meeting hours you save on them are paid back with interest in confusion later.
- Any recurring meeting that consistently ends early. That is a well-designed meeting, not an under-used slot, and someone will eventually try to fill it - do not let them.
Make the remaining meetings worth the hour
Once you have cut the load, the meetings that survive have to justify the space you defended for them. That mostly means making the hour participatory rather than presentational: the load complaint and the boredom complaint have the same root, which is that people are sitting in rooms where their presence changes nothing.
The mechanical fix is to give every attendee something to do in the first three minutes. A one-question poll on where the group stands, a silent round of typed input before the discussion opens, a quick ranking of the options on the table. Session Flo is built for exactly this - polls, ranked votes and typed contributions running alongside the discussion so twenty people can contribute in ninety seconds instead of six of them talking for an hour.
It also gives you evidence for the next audit. When every meeting produces a visible artefact - a vote, a ranked list, a set of contributions and a short recap - you can see which slots are producing anything and which are ritual. Meetings that consistently generate nothing are far easier to cancel when there is a record of the nothing.
And it changes the cancellation conversation from a cultural argument into a practical one. You are no longer telling people that meetings are bad. You are showing them which specific hours produced a decision, an owner and a date, and which produced a calendar entry.
Keep the calendar from refilling
Calendars regrow. Someone joins, a project starts, a crisis passes and its emergency stand-up quietly becomes permanent. Without a scheduled prune, you will be back where you started inside two quarters, and the second cull is harder because the first one has already taken the obvious candidates.
So put the review in the calendar - one hour, quarterly, with the same audit and the same test. Every recurring meeting carries an end date rather than repeating indefinitely, and anything that wants to continue past it has to be renewed by its owner. Renewal takes thirty seconds; the value is that somebody has to consciously choose it.
Give the mechanism to the people carrying the load rather than to the people booking it. Anyone should be able to decline an invite that arrives without a stated purpose and an agenda, without that being read as obstruction, and the norm holds only if senior people decline first and visibly.
The end state is not an empty calendar, and a team that gets there has probably broken something. It is a calendar where every recurring block has an owner who could tell you in one sentence what would break without it - and where the hours you did protect are spent in rooms that genuinely needed everyone in them.
Frequently asked questions
How do I know whether my team really has meeting overload?
Count two things across two completed weeks: total meeting hours per person, and the number of uninterrupted ninety-minute blocks each person had. The second number is the diagnostic one. A team averaging fifteen meeting hours a week with two long clear blocks a day is fine; a team averaging ten hours arranged so that nobody ever gets ninety consecutive minutes has a serious problem that a headline hours figure will hide completely.
Should we just cancel all recurring meetings and see what people miss?
A full amnesty is dramatic and it does surface which meetings people fight to reinstate, but it also breaks coordination for teams that had no written alternative in place. A safer version is to pause each recurring meeting for thirty days on a rolling basis and note what actually breaks. You get the same information, spread out, without a month of everyone rediscovering what the Tuesday sync was for the hard way.
Do no-meeting days actually work?
They reliably improve fragmentation for the people who had no clear time, and they do not reduce the underlying coordination load. The work moves to the other four days, so a protected Friday usually makes Thursday afternoon worse. Use them as relief while you fix the cause - unclear decision rights, missing written status, one-way broadcasts held live - rather than as the intervention itself, and be honest with the team about which one you are doing.
What is the right maximum number of people in a meeting?
It depends entirely on purpose. Decision meetings should be eight or fewer, because above that contribution per person falls away and quieter attendees stop speaking at all. Working sessions can run to twenty or thirty if the format is genuinely participatory and people are contributing in parallel rather than in turn. Broadcasts can be any size, but if it is a broadcast you should be asking why it is a meeting at all.
How do I push back on a meeting my manager runs?
Do not attack the meeting - offer to replace the job it does. 'Would a written status by Monday lunchtime cover what you need from the Tuesday sync, and we keep a fortnightly slot for anything blocked?' is a proposal, and it is much easier to say yes to than a complaint. If the answer is that the meeting exists for visibility, ask specifically who needs the visibility, and solve for that person rather than for the twelve people currently attending on their behalf.
Does making meetings more interactive reduce overload?
It does not reduce the hours, but it changes what those hours produce, and it makes overload much easier to diagnose. When every meeting generates a visible output - a vote, a ranked list, a set of typed contributions, a short recap - the slots that produce nothing become obvious and easy to cancel. It also fixes the related complaint: much of what people experience as too many meetings is really too many meetings in which their presence made no difference.
Keep reading
- killing the status meeting without losing visibility — The single largest recurring block on most calendars, and what replaces it.
- what actually happens on meeting-free days — Why a protected day relieves fragmentation without removing the cause.
- choosing between synchronous and asynchronous formats — Deciding which meetings genuinely need everyone at once.
- building a meeting culture worth having — Keeping the calendar from refilling once you have cut it back.
- more on running sessions people do not resent — Practical guides for facilitators and meeting owners.