I have sat through enough retrospectives to recognize the moment a team stops clearing its impediments and starts curating them. Someone reads the board aloud. The blocker at the top has been there for three sprints, worded a little more precisely each time, because the team has gotten good at describing it. People nod at how accurate the description has become. The card does not move. The list grows more articulate and never any shorter.

The retrospective was built as the place where a team looked hard at its own work and then changed something about it. In a lot of organizations it became the place where a team grew fluent in its own dysfunction. And the ritual scales upward. What the impediment board is to a delivery team, the engagement survey is to the whole enterprise. The instinct is the same and only the instrument gets bigger. A company now owns more ways to name what ails it than at any point in the history of management, and somewhere along the way the naming started to feel like enough.

The diagnosis got better while the fixing got harder

A theory-building paper in this year's Academy of Management Proceedings traced how a century of management scholarship handled dysfunction and arrived at an uncomfortable place: as the explanations for dysfunction grew more sophisticated, the capacity to act on them narrowed, and the corrective insights that did survive tended to be the ones that left the existing decision-making structures intact.1 Read that slowly. The better an organization got at explaining why it was stuck, the more its explanations were shaped so that nobody in charge had to move.

The complaint is old, only newly instrumented. A quarter century ago Jeffrey Pfeffer and Robert Sutton named it the knowing-doing gap and argued that most companies already knew what they needed to do and simply did not do it, falling instead into what they called the smart talk trap, where plans and presentations become a substitute for the deed rather than a route to it.2 What has changed since 2000 is not the gap. It is how much money and software now sit on the knowing side of it. We can measure sentiment weekly and have it benchmarked against millions of other employees on a dashboard before lunch. None of that machinery touches the doing.

The easy half and the hard half

You can watch the split in the numbers on change itself. Perceptyx, a firm that sells the listening tools in question, drew on more than 23 million employee responses across 490 organizations for its 2026 benchmark and found only 55 percent of employees rating their organization's change management favorably. The lowest-scoring practice it measured was whether leaders lead change well, at 61 percent, with whether an employee's opinion gets considered sitting just above it at 65.3 Their own analysts put it more plainly than most vendors would: "Communicating updates without inviting input covers the easy half of change and skips the half that decides whether employees adapt."3

The easy half is the naming. Circulating the survey, presenting the findings, sitting in a room and agreeing that yes, this is the problem, is comfortable work that produces artifacts and photographs well in a board deck. The hard half is redistributing something real once the room empties, a budget line, or a decision right that currently belongs to someone who would rather keep it. That half yields no artifact until much later, and it costs a specific, named person something now.

An organization that can describe its problem in fine detail and still not move on it already holds every piece of information it needs. What it is missing sits a level deeper, in the willingness to give something up, and another round of listening will never reach that.

Why precision starts to feel like progress

Here is the trap inside the trap. A vague worry nags at you. A precisely worded finding with a percentage beside it feels handled, even when nothing has been done, because the mind treats a well-formed sentence about a problem as evidence the problem is being managed. The more articulate the diagnosis, the stronger that false calm. I have watched leadership teams leave a session lighter for having named a thing they had no intention of changing. The relief of description had done the work that repair was supposed to do.

The tell is easy to check, if you want to. Pull last year's engagement results, or the transformation health-check from the year before, and read the top three items. Then ask what specifically moved because of them: what got funded, and who was handed authority they did not have before. If the honest answer is that the items were noted, socialized, and carried into this year's version of the same report, the instrument has stopped diagnosing your organization and started documenting a standstill in increasingly good prose. It is the mirror image of something I wrote about with a strategy that everyone can recite long after it has gone stale, where saying the direction out loud was allowed to stand in for pursuing it. None of this asks anyone to measure less. The ask is narrower: treat a diagnosis as a debt you have taken on rather than a task you have completed. A retrospective earns its keep the day a single impediment leaves the board because someone with the authority to remove it finally did. So here is the question to carry into the next planning cycle: what would come off your board this quarter if naming the problem no longer counted as handling it?