The design principle, and what it eliminates
A cadence is the mechanism that turns telemetry into behaviour. A business can have excellent numbers and change nothing, because nothing consumes them — and the consuming is a rhythm rather than a dashboard.
The design starts from decisions rather than from topics. What does this business decide on a recurring basis? Whether to take on more work. Whether an account needs attention. Whether a project is at risk. What to fix next. Each of those has a natural frequency, and the frequency determines the meeting rather than the other way round.
The immediate consequence is that most existing recurring meetings do not survive. If nothing is decided, the meeting is a status broadcast, and status should be written and read asynchronously — every high-performance field treats reflection as part of the job, and reflection is not the same as everybody listening to updates in sequence.
The second thing it eliminates is the meeting that exists because a problem happened once. Most bloated cadences are archaeological — each recurring session was created in response to a specific failure, and none was ever retired when the failure stopped recurring. Ask of every meeting when it started and what it was solving.
The architecture: four rhythms and what each decides
The daily rhythm is asynchronous and is not a meeting. A written thread: what is blocked, what needs a decision today. Its output is unblocking, and its value is that a blocker raised at 9am does not wait for Thursday. Owner: everyone, five minutes each.
The weekly delivery review decides what is at risk and what to do about it. Thirty minutes, evidence prepared in advance: work items past their promised date, projects with no next scheduled step, and current committed load against the capacity threshold. The rule that makes it work is that the evidence is circulated before, not presented during — presenting the numbers consumes the slot the decision was supposed to occupy.
The weekly pipeline review decides which opportunities get attention. Twenty minutes, evidence: deals with no scheduled next step, stage age against the average, and anything that crossed a stale threshold. Not a recital of every deal — only the exceptions, because a review that walks the whole board is a status broadcast wearing a decision's clothes.
The monthly business review decides what to change. Sixty to ninety minutes, evidence: the scoreboard metrics with their trend, the client health movers, the failure record from delivery, and progress on whatever was decided last month. This is the only meeting where structural changes get made, and it needs the longest slot for that reason.
The quarterly review decides direction, and it is the one most likely to be skipped and most expensive to skip. Evidence: everything above at quarter scale, plus the register and estate reviews, plus what was decided three months ago and whether it happened.
What runs between the meetings
The decision record is the artifact that makes the cadence compound. Every meeting produces decisions with an owner and a date, in one place, and every subsequent meeting opens by checking the previous ones. Without it, the same issue is discussed monthly for a year and everybody experiences the meetings as pointless — correctly.
The evidence pack has an owner and a deadline, and the deadline is before the meeting. If the numbers get assembled during the session, the session becomes assembly, and the decision gets whatever time is left.
The escalation path handles what cannot wait for its rhythm. Named thresholds that trigger an immediate conversation rather than an agenda item — a client at risk, a delivery date that will slip, a capacity breach. A cadence with no escalation path teaches people to sit on urgent things until Thursday.
And the standing exception review: anything that was overridden, deviated from, or handled outside the process since last time. That list is where process improvements come from, because a recurring exception is the standard telling you it is wrong.
The failure edges
The first: meetings with no decision. They persist because cancelling feels like disengagement, and they consume the attention the real ones need.
The second: evidence presented rather than circulated. Twenty of the thirty minutes go on reading numbers aloud, and the decision gets made in the last five under time pressure.
The third: no decision record. The same conversation recurs, nobody notices it is recurring, and attendance quietly declines because everyone can feel that nothing follows from it.
The fourth: reviewing everything rather than the exceptions. A pipeline review that walks all forty deals has no time for the four that matter, and the format actively hides them.
The fifth: numbers with contested definitions. Twenty minutes go on reconciling two figures that are both correct, every time, because nobody wrote down what the word means. The definition belongs next to the number.
The sixth: the cadence survives only while the founder chairs it. If the meetings stop when you are away, they are your meetings rather than the business's rhythm, and the rhythm was the point.
The seventh: every meeting attended by everyone. Attendance should follow from who makes the decision, and a session with four observers is four people learning that their time is not respected.
The evidence packs, and how to start without breaking everything
The evidence packs are the only real engineering in this system, and they determine whether it survives. Each meeting's numbers should assemble themselves and arrive before the session — the delivery pack from the project tool, the pipeline pack from the CRM, the monthly pack from wherever the scoreboard lives. A cadence that depends on somebody spending three hours preparing slides lasts exactly as long as that person's patience.
Start by subtracting rather than adding. Take the existing recurring meetings, ask what each decides, and cancel the ones with no answer. That single pass usually frees more time than the new cadence consumes, and it makes the introduction of the new rhythm a net reduction in meetings rather than an imposition.
Then introduce the decision record before anything else, even into the meetings you have not restructured yet. It costs nothing, it takes two minutes at the end of a session, and within a month it makes the recurring conversations visible — which is usually what convinces everybody that the rest of the redesign is worth doing.
What done looks like, and what it takes to build
Done is a week where the risks surfaced before anyone was surprised, the decisions were made by the people closest to them, and a month later you can show what was decided and what happened.
The checklist: a named decision for every recurring meeting; an evidence pack per meeting with an owner and a pre-circulation deadline; exceptions-only formats for the weekly reviews; a decision record with owners and dates, reviewed at the start of each session; an escalation path with named thresholds; a standing exception review; and written definitions attached to every number used.
The build is mostly design rather than engineering — one to two weeks, and most of it is deciding what gets decided where. The engineering is in the evidence packs, which have to assemble themselves. A cadence that depends on somebody spending three hours preparing numbers will survive exactly as long as that person's enthusiasm.
The prerequisite is having numbers worth reviewing. A cadence built on metrics inherited from tool defaults will produce meetings that faithfully review the wrong things — the rhythm cannot fix the instruments, it can only consume them.
Every recurring meeting needs a decision it exists to make; the rest is status and belongs in writing. Circulate the evidence beforehand and open with the last set of decisions.