Why decisions get reopened, and the threshold that stops it
You settled the pricing model in March. Everyone was in the room, nobody objected, and the meeting ended early. It came back in May, reframed, raised by someone who had been quiet the first time. You settled it again. It is now on the agenda for a third time.
The instinct is to treat this as a discipline problem. People are not committing. The team is not aligned. Decisions get reopened, the story goes, because somebody will not let go, and somebody needs to say the word final and mean it.
That reading is comfortable and usually wrong. A decision that returns three times was never closed, it was paused, and it was paused because nothing in your structure ever said what reopening would cost. When reversal is free, everything stays open by default, and the decisions that do stick survive because the losing party lost interest rather than because the system required them to stop.
- A decision that keeps returning was never closed. It was paused. Closure is a structural property rather than a feeling in the room, and it exists only when something in the system names what reopening would cost.
- The Reopening Threshold: before the decision leaves the room, write four things down. The door type (reversible or not), the trigger that would justify revisiting it, who may invoke that trigger and to whom, and the rule that the decision stays in force while any challenge is heard.
- Paul Nutt studied 356 decisions in medium and large organizations in the United States and Canada and found that half failed, meaning they were never adopted or fell out of use within two years. He traced those failures to the tactics managers used rather than to circumstances outside their control.
- A decision log records what you chose, but it does not govern what it costs to unchoose it. Minutes, logs and "disagree and commit" all fail in the same place, because none of them names a bar.
- Sull, Homkes and Sull surveyed 7,600 managers in 262 companies across 30 industries and found that conflicts between functions were handled badly two times out of three: resolved after a significant delay 38 percent of the time, resolved quickly but poorly 14 percent, or simply left to fester 12 percent.
- Reversible decisions should be cheap to reopen and irreversible ones should be expensive. One uniform bar produces either a team that relitigates everything or a team that cannot correct a mistake it has already spotted.
Why do decisions get reopened?
Decisions get reopened because nothing in the structure ever priced the reversal. Reopening costs the challenger nothing, so raising it again is the rational move, and the person who wants the decision to hold has no rule to point at.
Three mechanisms produce that, and they stack.
The first is that nobody held the pen. If the meeting produced a direction rather than a decider, then no one has standing to decline the second conversation. A challenge lands on a group, groups do not refuse things, and the item returns to the agenda. This is the same failure that keeps leadership meetings from producing decisions, read forward in time: an unowned decision does not just fail to close in the room, it fails to stay closed afterwards.
The second is that agreement in the room was not commitment. Silence at the end of a long meeting is frequently consent to leave, not consent to the decision. The person who never spoke has conceded nothing, so they act on their original view, and the gap between what was agreed and what is happening surfaces four weeks later as a fresh disagreement.
The third is the one nobody writes down. Reopening has no price. There is no statement anywhere of what would have to be true for this to come back, so any reason works: a new data point, a customer complaint, a change of mood, a bad week. Every one of those is arguable, and none of them can be refused.
Paul Nutt's work is the most uncomfortable evidence on this. Across 356 decisions studied in medium to large organizations in the United States and Canada, he found that half failed, where failure means the decision was never adopted or fell out of use within two years, and that tactics prone to fail were used in two of every three decisions he examined (Nutt, 1999). Note the scope: this is North American data from medium and large organizations, not a Romanian or small-company sample, and Nutt's attribution of failure to managerial tactics is his interpretation of his own case database rather than an experimental result. What survives the caveats is the shape. Decisions do not usually die in the meeting. They die quietly afterwards, in the space where nobody defined what closed means.
Is a decision log enough to keep a decision closed?
No. A decision log is evidence of what was chosen, and evidence does not close anything. It records the past and says nothing about the conditions under which the past may be revised, which is exactly the ground a challenger stands on.
The standard advice is to write things down. Keep a log, record the rationale, circulate the minutes. It is good advice and it is not sufficient, and the difference matters because most teams try the log, watch it fail, and conclude the problem is human.
A log tells you what was decided, by whom, and why. When someone reopens the question, it lets you establish that the conversation already happened.
But establishing that the conversation happened does not end it. The challenger's answer is always available: circumstances changed. A log has no reply to that. You end up with a well-documented argument instead of an undocumented one.
"Disagree and commit" has the same shape. Jeff Bezos described it in Amazon's 2016 shareholder letter as a genuine decision to back a direction he did not personally favour, framing it as "a candid expression of my view, a chance for the team to weigh my view, and a quick, sincere commitment to go their way" (Bezos, 2017). Read carefully, that is a norm about how an individual behaves after losing an argument. It works when people practise it and provides nothing when they do not, which is precisely the case you are trying to solve.
What both are missing is a bar. Not a record of the decision and not an expectation of good behaviour, but a stated condition that a challenge has to clear before the organization spends time on it again.
The Reopening Threshold: four fields that price a reversal.
Write four lines before the decision leaves the room. They take about ninety seconds and they are what turns a conclusion into a closed decision.
Bezos's distinction is the useful one here: some decisions are two-way doors you can walk back through at low cost, and some are one-way doors that commit capital, headcount or a public promise. Say which one this is out loud, because the answer sets everything below it. Teams that skip this field end up defending a font choice with the same energy as a market exit.
Name it specifically while you still have the context: a named customer segment behaving differently, a cost assumption missing by more than a stated margin, a competitor move of a particular kind. A trigger written in advance is a rule. A trigger invented during the challenge is a negotiation.
One name on each side. This is the field that keeps the threshold from becoming a wall: everybody can see the route back in, which is what makes the closure legitimate rather than merely imposed.
This is the field people forget, and it does more work than the other three combined. Without it, raising a challenge is itself a way of stopping the work, and a challenger who cannot win can still delay indefinitely.
Door, trigger, standing, force. The structure resembles the Seam Contract that closes a decision owned by two functions, and deliberately so. That one assigns authority before the decision. This one holds it afterwards. A team that has done the first and not the second will still be reopening cleanly assigned decisions, which feels like bad faith and is actually a missing field.
The pattern I run into most often, and I am describing a recurring shape across engagements rather than one company, is a leadership team convinced it has a commitment problem. We take the last quarter's agendas and mark every item that appeared more than once. The list is always longer than anyone expected. Then we ask one question per repeated item: what would have had to be true for this not to come back? Almost every time the answer exists, is uncontroversial, and could have been written in a sentence on the day. Nobody wrote it, because on the day the decision felt finished.
Which decisions should be easy to reopen?
Reversible ones, and you should say so explicitly rather than hoping people infer it.
A team that sets a high bar on everything stops correcting itself. Somebody notices in week two that a two-way-door decision is wrong, weighs the cost of raising it against a culture that treats reopening as weakness, and says nothing. That failure is quieter than relitigation and considerably more expensive, because the organization keeps executing a decision it already knows is wrong.
So the threshold is not uniformly high. It is proportional to what the decision commits. A reversible call gets a low bar and an explicit invitation to flag problems early. An irreversible one gets a high bar, a named trigger, and a single person who can hear the challenge. The point of writing the door type first is that it makes this calibration visible, instead of leaving each person to guess how final "final" is.
There is a coordination dimension too, and it is the reason reopening is often not really about the decision at all. Sull, Homkes and Sull surveyed 7,600 managers in 262 companies across 30 industries, mostly large firms averaging around 6,000 employees, with a third based in emerging markets. They found that only 9 percent of managers say they can rely on colleagues in other functions and units all the time, with 56 percent saying all or most of the time, and that conflicts between functions get handled badly two times out of three: resolved after a significant delay 38 percent of the time, resolved quickly but poorly 14 percent, or left to fester 12 percent (Sull, Homkes & Sull, 2015). Worth keeping the two figures distinct, since the article's summary chart reports the "all or most of the time" number and the running text reports the stricter "all the time" one. When a decision needs another function to act and that function's follow-through is unreliable, reopening the decision is often the only lever a manager has. The item comes back not because anyone rejects it, but because relitigating is easier than enforcing.
How do you install a Reopening Threshold this week?
You do not need a new operating rhythm. You need one pass over your own agendas and a habit at the end of each decision.
- Pull the last quarter's leadership agendas and mark every item that appears more than once. Use the documents, not your memory, because memory quietly edits out the third occurrence.
- For each repeated item, write what brought it back. New information, an absent stakeholder, or nothing at all. The third category is usually the largest, and it is the one this fixes.
- Take the three items with the highest cost per repetition and write a Reopening Threshold for each retroactively. Door, trigger, standing, force. Circulate it as a rule rather than as a verdict on the previous rounds.
- Add the four fields to the close of every decision from now on. Ninety seconds at the end of the item, while the context is still in the room. Doing it later means doing it without the context, which is how you get a vague trigger.
- Review after one month, and count. Not whether people liked it, but how many items came back, and whether the ones that came back cleared their own trigger. Items that returned without clearing a trigger tell you the standing field is not being honoured.
This connects directly to what Leadership Architecture measures. Organizational Execution Capacity asks whether decisions get implemented quickly, in a coordinated way, and without reopening. Decision Clarity asks whether authority is defined and applied consistently, which is the upstream condition for anyone having standing to hold a decision closed at all. Repeated agenda items are the cheapest visible signal of a gap between the two. If the pattern shows up across many decision types rather than a few, the underlying issue is likely to be decision rights rather than closure discipline, and the one-hour decision rights audit is the better next step.
Here is what I would ask a leadership team to notice. When a decision comes back for the third time, the conversation is almost never about the substance. It is about whether the first two conversations counted. That question is not a matter of character, and no amount of insisting on finality will settle it. Somebody has to write down what closed means, once, before anyone needs it.
Frequently asked questions
Why do decisions keep getting reopened in leadership meetings?
Because reopening costs the challenger nothing. Three conditions usually stack: no single person held the decision, so nobody has standing to decline a second conversation; agreement in the room was consent to end the meeting rather than commitment to the outcome; and no one ever stated what would have to be true for the item to return. The third is the structural one, and it is the only one you can fix in a sentence.
What is the Reopening Threshold?
The Reopening Threshold is a four-field rule written at the moment a decision closes: the door (whether the decision is reversible), the trigger (what new information would justify revisiting it), the standing (who may invoke that trigger, and to whom), and the force (that the decision stays in effect while any challenge is heard). It prices the reversal in advance, which is what turns a conclusion into a closed decision.
Is a decision log enough to stop relitigating decisions?
No. A log is evidence of what was decided and gives you a record to point at, but it says nothing about the conditions under which the decision may be revised. The standard challenge, that circumstances have changed, is unanswerable by a log. You need a stated bar the challenge must clear, not just a record that the earlier conversation happened.
Should some decisions be easy to reopen?
Yes, and you should say which ones out loud. Reversible decisions, what Bezos called two-way doors, should carry a low bar and an explicit invitation to flag problems early, because a team that treats all reopening as weakness will keep executing decisions it already knows are wrong. Irreversible decisions get the high bar. Naming the door type first is what makes that calibration visible instead of guessed.
How is this different from "disagree and commit"?
"Disagree and commit" is a norm about individual behaviour after losing an argument, and it works when people practise it. The Reopening Threshold is a structural rule that operates whether or not they do. One asks people to hold the line voluntarily; the other states, before anyone is invested, what a challenge has to clear. They combine well, but only the second one still works on a bad week.
Sources
Paul C. Nutt, Surprising but true: Half the decisions in organizations fail, Academy of Management Executive, 1999, 13(4), 75–90. Studies of 356 decisions in medium to large organizations in the U.S. and Canada; source of the half-fail finding, the two-year follow-up definition of failure, and the finding that tactics prone to fail were used in two of every three decisions studied.
Donald Sull, Rebecca Homkes & Charles Sull, Why Strategy Execution Unravels and What to Do About It, Harvard Business Review, March 2015. Survey of 7,600 managers in 262 companies across 30 industries, averaging 6,000 employees, one-third based in emerging markets; source of the reliance figures and the 38, 14 and 12 percent split on how cross-functional conflicts are handled.
Jeff Bezos, 2016 Letter to Shareholders, Amazon, 2017. Source of the "disagree and commit" passage quoted above and of the two-way door framing.
Client examples are anonymized composites; no figures are invented.
How many of your decisions are actually closed?
Five minutes. No account. A structural read on whether your decision system, not your team's commitment, is the real constraint.