team
The daily standup was fifteen minutes when it started. It is now twenty-eight. Six people give an update, two of them talk for four minutes each, and somewhere around the third person a design question opens up that only concerns two of the six. Everyone else waits. The people who have nothing to report have learned to invent something, because saying nothing feels like admitting to a slow day.
The usual response is to ask people to be brief. That response fails every time, because the length is not caused by people being long-winded. It is caused by a format that has no mechanism for stopping. This article is about the mechanisms: what the meeting is actually for, which formats hold the clock, where the conversations that overflow should go, and how to tell when a daily standup is no longer worth holding at all.
A standup is a synchronisation, not a status report. The distinction decides everything else about the format.
Synchronisation means each person leaves knowing two things they did not know when they arrived: where the work has moved since yesterday, and whether anything they are about to do collides with what someone else is about to do. That is the entire job. It is finished in a few minutes when the meeting is built for it.
A status report is a different activity with a different audience. Its purpose is to inform someone who is not doing the work, usually a manager or a stakeholder, of progress. It has to be complete, it has to be defensible, and it is naturally long. When a standup quietly becomes a status report, three things follow. People start addressing the most senior person in the room rather than each other. Updates become justifications of how the day was spent. And the meeting stretches, because completeness is the goal and completeness has no natural end.
The clearest test is where people look while they talk. If everyone reports to one person, the meeting has become a status report, whatever it is called on the calendar. The fix is not a shorter time limit. It is to move the status reporting somewhere it belongs, usually a written weekly summary, so the daily meeting can go back to doing the small job it is good at.
A second test is whether anything changes as a result. If nobody has ever reordered their day because of something heard at standup, the synchronisation is not happening, and the meeting is being held out of habit.
The familiar structure asks what was done yesterday, what is planned today, and what is blocking. It is a reasonable starting point and it degrades in a predictable way.
The question about yesterday invites retrospective narration. It produces the longest answers and the least useful information, because what happened yesterday is visible on the board already for anyone who cares to look. The question about today produces a list of intentions that nobody will check. The question about blockers is the only one that reliably matters, and it arrives last, after attention has already drained out of the room.
There is a structural problem underneath the ordering. The three questions organise the meeting by person, which means the meeting length scales with headcount rather than with the amount of work in flight. Ten people with a quiet week still produce ten updates. The board, meanwhile, might have three cards that genuinely need discussion.
Teams that keep the format often improve it by inverting the order: blockers first, then today, and yesterday only when it explains a blocker. That single change recovers several minutes and puts the useful part of the meeting where people are still listening.
There is no single correct format. The choice depends on team size, how much the work overlaps, and how many time zones are involved.
| Format | How it runs | Holds time when | Main cost |
|---|---|---|---|
| Round the room | Each person answers in turn | Fewer than about six people | Length scales with headcount, not with work |
| Walk the board | Move through cards in progress, right to left | Work is visible and cards are current | Requires the board to be genuinely up to date |
| Written async | Each person posts an update in a thread before a cutoff | Across time zones, or with deep focus work | Loses the spontaneous collision catching |
| Written plus short call | Updates posted in writing, a ten minute call only for flagged items | Most distributed teams | Two habits to maintain instead of one |
Walking the board is the change that most often fixes a standup without anyone having to be told to talk less. Instead of going person by person, the meeting goes card by card through the work that is in progress, starting from the column closest to done. Cards nobody has anything to say about pass in seconds. People with no cards in flight say nothing and that is fine, which removes the pressure to invent an update.
Walking the board also changes the subject of the meeting from people to work. Nobody has to account for their day. The question is whether each piece of work is moving, and if not, what would move it. The side effect is that the board has to be accurate, which is usually an improvement in itself.
A board walk only works if the board is current at the moment the meeting starts. That is a real requirement, and it is where most attempts quietly fail.
Three conventions make it hold.
If cards are dragged into position five minutes before standup, the board is a performance rather than a record. The habit worth building is to move a card at the moment its state actually changes, which takes a few seconds and makes every view of the board trustworthy for the rest of the day.
A card with two owners will be discussed twice and started by neither. A person with nine cards in progress has nine cards none of which are moving. Most small teams find that a limit of two or three cards in progress per person makes the standup shorter without anyone shortening a sentence, because there is simply less in the column being walked.
Work blocked on someone outside the team, a review, an approval, a vendor, or a client, is not in progress and should not sit in the same column as work being actively done. A separate waiting column turns the longest recurring standup conversation into something visible at a glance. The count of cards sitting there is the most useful single number in the meeting. Tools differ in whether such a column is a simple stage, a status field, or a filter, and the features worth checking before committing are whether the same cards can also be seen as a timeline and whether chat sits next to the board rather than in another application.
Every standup produces conversations that concern two people and bore the rest. The meeting does not run long because those conversations happen. It runs long because there is nowhere for them to go.
The mechanism is a parking lot, and it needs three parts to work. First, anyone may name a topic and it is written down immediately rather than discussed. Second, the names of the people who need to be in that conversation are written next to it. Third, the conversation happens straight after the standup, with only those people, while everyone else leaves. The third part is the one that gets skipped, and skipping it teaches people that the parking lot is where topics go to be forgotten.
A parking lot that lives in the meeting chat thread rather than in someone's notebook has an additional benefit: the two people involved can often resolve the item in writing without another meeting at all. Teams whose chat sits alongside the board tend to do this more, because the card and the conversation about it are in the same place.
One person should be responsible for saying the words that stop a conversation. Rotating that role weekly works better than assigning it permanently to a lead, because it stops the standup from feeling like a meeting run by management, and it spreads the mildly uncomfortable job of interrupting.
Across more than about three hours of time zone difference, a synchronous daily meeting costs someone their morning or their evening. The written standup solves that, and it introduces its own failure modes.
The format that holds up is a single thread per day, one post per person, with a posting cutoff rather than a meeting time. The cutoff matters more than the format of the post: without one, updates arrive throughout the day and nobody reads them.
Written updates should be shorter than spoken ones, not longer. Three lines is right. The temptation is to write a paragraph per task, which nobody reads, which makes the whole thing ceremonial. Linking to the card rather than describing its contents keeps posts short and keeps the detail in one place.
The thing genuinely lost in a written standup is the accidental collision catch, the moment when someone says a file name and another person says they are editing it right now. Teams that go fully asynchronous usually recover some of that with a short call once or twice a week focused only on the week ahead.
There is also a hybrid that suits teams spread across two regions: everyone posts in writing before a cutoff, and a ten minute call happens only when someone flags an item that needs voices. Most days the call does not happen.
A daily standup is not free. Fifteen minutes for six people is an hour and a half of working time every day, and the interruption cost around it is real for anyone doing deep work.
Four signals suggest the format is wrong for the team as it now is.
Nobody ever reorders their day because of it. The synchronisation is not happening. Either the work does not overlap enough to need daily coordination, or the meeting is not surfacing the overlaps.
The same blocker is reported on three consecutive days. The standup is recording an unresolved problem rather than resolving it. Something outside the meeting needs to change, and repeating the report daily creates the feeling of action without the fact of it.
Updates are written for an audience that is not in the room. The meeting has become a status report and the reporting should move to a written weekly summary.
Half the team works alone on unrelated things. Daily coordination is most valuable where work overlaps. A team of specialists on separate projects often does better with a written daily post and one real meeting a week.
Dropping a standup is a legitimate outcome. Two conditions make it safe: the board must be current enough that anyone can see the state of the work without asking, and there must be a channel where blockers get raised the moment they appear rather than waiting for a meeting.
Change the shape before changing the duration. Walk the board instead of going round the room, put blockers first, and start a parking lot that is genuinely used the minute the meeting ends. If the meeting still runs long after that, the cause is almost always a board that is not current, and the next step is to make updating a card part of doing the work rather than part of preparing for the meeting. A tool where kanban, timeline and chat sit together makes that habit cheaper to keep, and Pinateca is free for up to five people and ten boards.
Fifteen minutes is the usual ceiling, and most teams of five or six can finish in under ten once they walk the board instead of going round the room. The length is a symptom rather than a target. A meeting that consistently runs over has a format problem, and shortening the calendar slot without changing the format simply moves the overflow into the corridor afterwards.
No. Daily coordination pays off when the work overlaps, so a team of four people on one product usually benefits and a team of specialists on unrelated projects often does not. Three days a week, or a written daily post with one meeting a week, are both reasonable. The test is whether anyone ever changes what they were going to do because of something they heard.
Attendance is less important than behaviour. A manager who listens and asks questions afterwards changes little, while one who receives the updates turns the meeting into a status report and roughly doubles its length. If attending cannot be separated from receiving reports, it is usually better to read the board and set up a separate weekly summary.
Change the format rather than speaking to the person. Walking the board removes the open-ended slot that long updates fill, because the unit is the card rather than the person. If a single item genuinely needs discussion, the parking lot exists for it, and rotating the role of the person who invokes it stops the interruption from feeling personal.
It is better for time zones and for deep focus work, and worse at catching accidental collisions between two people about to touch the same thing. The usual compromise is written updates before a daily cutoff plus one short call a week. Whichever is used, posts should be around three lines and should link to cards rather than describing what is on them.