Some meetings end with a decision. Most end with the feeling of one. This edition is about the gap between those two and what it costs.

If you follow the LinkedIn feed, you’ve seen the short version this month: a meeting’s real cost has three layers, and most organizations track only the first. This is the full argument, including the mechanism behind each layer, why the room can’t close, and why the P&L never names the cause.

It’s a diagnostic, not a checklist. By the end, you won’t have a new process. You’ll have a lens and a harder question to ask about the forums you sit in every week.

Picture the meeting. It starts on time. The right people are in the room, including a director, two senior managers, and the leads from planning, commercial, and finance. The discussion is sharp. Everyone understands the issue. Everyone contributes. The slides are clear. People nod. The hour ends, calendars reclaim everyone, and there is a shared sense that something moved.

Nothing was decided.

“Nothing useful was said.” Plenty was said. But no one left the room owning a call that wasn’t already owned when they walked in. The premise that brought eight people together is still open. And because the meeting felt productive, with the texture of progress, no one names the gap. You only notice that a meeting failed to decide when you go looking for the decision later and find it isn’t there.

This is the most expensive kind of meeting an organization runs. Not the long one. Not the badly chaired one. The one that mobilizes senior attention, performs the motions of resolution, and hands the problem back to the operation unchanged. A costly meeting is not one with senior people in the room. It is one where senior leaders come together, and the problem still goes back to operations, undecided.

The reason it stays invisible is that we price meetings wrong. We look at one layer of cost and ignore the two that matter.

Ask anyone what a meeting costs, and they will describe the visible layer. Who was in the room? How long did it run? How many people were pulled away from the work to attend? Multiply salaries by hours, and you get a number, the one that fuels “we have too many meetings” and the periodic campaigns to shorten them.

This layer is real. It is also the cheapest and the only one that most organizations ever measure.

It is the only one we measure because it is the only one that is legible on a calendar. The visible cost has a start time, an end time, and an attendee list. It can be counted without judgment. The other two layers have none of that structure. They do not appear in the same place, at the same time, or under the same name as the meeting that produced them. So they go uncounted, and “uncounted” quietly becomes “does not exist.”

That is the first diagnostic error: treating the measurable layer as the whole cost. Shorter meetings, tighter agendas, fewer attendees, all of it optimizes the cheapest layer while leaving the expensive ones untouched. You can run a forty-minute meeting with four people and have it cost the organization more than a two-hour meeting with twelve, if the short one fails to decide and the long one does not. The cost of a meeting is not the time it consumes. It is the decision that was supposed to be produced, and what happens when that decision is missing?

There is an asymmetry here worth naming. The part of a meeting that is easy to count is the part that costs the least. The expensive part, the decision that was or was not made, is the part that resists counting entirely. So measurement drifts toward the cheap, legible layer and away from the costly, invisible one, and the organization ends up managing the number it can see instead of the cost it is actually paying. Every productivity initiative aimed at meetings inherits this bias. Trim the calendar, cap the attendees, and install the no-meeting Friday. Each targets the visible layer, and none of them address whether the remaining meetings actually decide anything.

The second layer is recurrence, signaled by a phrase everyone has used: "let’s take it offline." Or its cousins: let’s align after, let’s reevaluate, let’s get the right people and revisit.

Sometimes that is a legitimate handoff. Usually it is not. Usually, it is the sound a meeting makes when it has run out of time, appetite, or authority, and the decision it was called to make is deferred without anyone saying so. The phrase is socially graceful precisely because it hides the failure. No one has to admit the room could not close the call. It simply rolls forward.

And it rolls forward onto a calendar. The same topic reappears the following week, wearing a slightly different agenda title. The same tension between planning and commercial gets re-litigated. The follow-up exists for one reason: the previous meeting did not reach a decision. But it does not look like a failure. It looks like diligence, a team staying close to a hard problem, meeting about it regularly, taking it seriously.

Consider a recurring supply review, the kind that exists in most operations under one name or another. A product family is carrying more inventory than it should. The review opens with the same chart it did a month ago. Everyone agrees the level is wrong. Planning has a view. Commercial has a different one. Finance wants the cash back. The conversation is intelligent, the hour runs out, and the meeting closes the way it did last time: someone will take it offline and bring options to the next session. The inventory does not move. The next session arrives, and the chart looks slightly worse.

That disguise is the danger. Recurrence reads as a healthy cadence. A standing forum that revisits the same issue month after month looks committed rather than stuck. The org chart says the topic is owned. The meeting series says it is being managed. Only a decision log, if one exists, would reveal that the same call has been opened and reopened a dozen times and never closed. The recurring cost is the meeting tax you pay over and over for a decision you never bought into in the first place. Because each instance is small and looks reasonable, the total is never added up.

The third layer is where the real money is, and it is the one furthest from the meeting that caused it.

When a forum does not produce a decision, the work does not stop and wait. It cannot. Operations run every day. So in the absence of a call, each function does the only thing it can: it defaults to its own interpretation and keeps moving. Planning protects the plan. Commercial protects the customer. Finance protects the margin. Each is rational within its own logic. Each is now pulling in a slightly different direction because the premise they should have shared was never closed.

This is where the cost stops being abstract. Plans stay open, so people execute against assumptions that do not match. Inventory sits longer than it should because the disposition decision that would move it was never made. Rework starts to look like normal coordination: the second and third passes at something that one decision would have settled on the first pass. Emails begin doing the job the decision should have done, threads multiplying as people try to manufacture the alignment a meeting failed to produce. And new meetings get scheduled to resolve the confusion created by the meeting that did not decide.

None of this shows up labeled as “cost of an undecided meeting.” It shows up as inventory carrying cost. As rework hours. As service noise, the expedites, the escalations, the exceptions that get handled heroically and never traced to their origin. As margin leakage that finance can see in aggregate but cannot attribute to a source.

Each of these has its own quiet mechanism. Rework compounds because the second attempt is rarely recognized as rework. It is filed as diligence, as another pass, as coordination, so it never gets counted as the cost of an earlier non-decision. Service noise compounds because every expedite and exception is absorbed by people who are good at absorbing them. The heroics that paper over a missing decision are exactly what keep it from being seen. The better the operation is at coping, the more invisible the original failure becomes. Competence downstream hides the gap upstream, not a gap in people, but in a decision system that never closed the call.

And this layer is the one most exposed to a comfortable misreading: that it is just a coordination problem, or a culture problem, or a sign that some people are not good at their jobs. It is none of those. People working competently from different assumptions will produce conflict, no matter how skilled or well-intentioned they are, because what they lack is not skill or goodwill. It is a closed decision. Cross-functional friction is rarely a people problem. It is a decision vacuum with an org chart drawn around it.

A meeting fails to decide for reasons set before anyone walks in. The failure seems to occur during the hour when the discussion drifts and the call gets deferred, but the conditions for it were established the moment the meeting was convened, without three things being named.

The first is the decision itself. Most meetings are scheduled around a topic, not a decision. “Inventory review.” “Supplier performance.” “S&OP alignment.” A topic invites discussion and has no natural end state. A decision is binary: it is made, or it is not. When a meeting is organized around a topic, it can run its full length, cover the ground thoroughly, and end without anyone being able to say what was decided, because deciding was never the defined output.

The second is the owner. Not the chair, who runs the room. The owner is the person accountable for the call and its consequences. When that role is unassigned, the meeting has no one responsible for forcing closure. Everyone can contribute, and no one has to conclude. Deferral becomes the path of least resistance because no single person carries the cost of deferring.

The third is the threshold, the condition under which action is required, agreed before the discussion, so the room is not negotiating the premise and the response at the same time. Without a pre-set threshold, every meeting reopens the question of whether this even warrants a decision yet. And “not yet” is always available, always defensible, and always cheaper in the moment than committing.

Name the decision, assign the owner, set the threshold, and a meeting has a forcing function. It has to close, because the structure leaves no graceful way to defer. Leave those three unnamed, and the meeting inherits full discretion not to decide. It will almost always use it, not out of weakness, but because nothing in the room is built to stop it. The recurring forum that never closes a call is not staffed by people who cannot decide. It is a room engineered, by omission, to make deferral the easiest available move.

Put the three layers side by side, and the diagnosis becomes clear.

The visible cost is incurred in the room during the hour and is attached to names. The recurrence cost occurs weeks later, on a calendar, spread across a series that no one reads as a series. The execution cost occurs elsewhere entirely, in a warehouse, in a margin variance, in a customer’s service experience, and it happens months after the meeting that set it in motion.

Three costs, three locations, three timelines. No system connects them. The meeting lives in one record. The inventory lives in another. The margin variance is a third. By the time the cost is large enough to notice, it is so far from the undecided meeting in time, in department, in vocabulary that no one traces it back. The P&L shows the symptom with great precision and says nothing about the cause.

This is why the decision that was never made is the most expensive item in the organization that no one can find. It has no line. It does not appear as “decisions deferred” or “premises left open.” It appears as if everything downstream of those things, scattered across functions that each see their fragment and none see the source.

An organization can be rigorous about every measurable cost and still bleed from this one, because rigor follows the data. No data indicates that this margin point traces back to a March supply review that ended with “let’s align offline.” The connection exists. The instrumentation to see it does not. That gap between a real cost and an untraceable cause is where the most expensive decisions in a company disappear.

And the gap is self-protecting. Because the cost surfaces far from its cause, the people who feel it are not the people who could have prevented it. The warehouse manager, living with the excess inventory, was not included in the supply review. The finance analyst flagging the margin variance has no line of sight into the meeting that produced it. Each owns a symptom, none owns the decision. So the cost is real to everyone and traceable to no one, which is the precise condition under which a problem persists indefinitely, funded quietly by the operation and never charged to the room where it began.

A diagnosis only matters if it changes what you inspect: three moves, one per layer.

Visible cost: pick your most senior recurring forum. Count the decisions it closed last quarter, not topics discussed or calls closed, with an owner. If that number can’t be reconstructed from any record, the absence is the finding.

Recurrence cost: scan the agenda titles of the last twelve sessions. Any topic appearing three or more times isn’t being managed. It’s being revisited, a decision deferred under a new name.

Execution cost: take one expedite, one write-off, or one overflow storage invoice from last month and trace it backward. What decision, made when, would have prevented it? Did a forum exist where that decision should have happened? It usually did.

None of this requires a new system. It requires reading the records you already have and being willing to name what they show.

This is a diagnosis, not a prescription. The point is not to install a framework by Monday. The point is a lens, and it is uncomfortable on purpose.

The next time a meeting on your calendar ends, do not ask whether it ran on time. Ask what it decided and who now owns that decision. Then ask the harder version: of the recurring forums you sit in, how many closed a call in the last quarter, and how many met again about the same open premise?

The meetings that feel productive are not the ones to trust. Productivity is a texture, and a meeting can have it while deciding nothing. The meetings worth their cost end with an owned decision, and those are often the shortest, least comfortable, and least ceremonial of the week.

The hour was never the cost. The decision that did not happen is, and you will pay for it later, in inventory, rework, service noise, and margin, with no line that says why.

If you want the operating-level diagnostics that don’t fit a long-form piece, that’s where I publish them: The Tuesday and Thursday posts on LinkedIn (linkedin.com/in/psegala) are where the smaller, sharper observations land first. This newsletter goes deeper into a single argument. The feed is where the next one starts.