Somebody serious about lifting keeps a notebook, and what goes into it is short. Sets, reps, and the weight that actually came off the floor. That is close to the whole record, and it is enough, because those numbers let the lifter make the only decisions available to them. Push harder here. Ease off there. If the weight on a compound movement has not gone anywhere in five weeks, that is a plateau, and it means either something in the approach has to change or they have found the ceiling of what their body will give them for now. The notebook works because the person holding it can read it and do something about what it says.

Now picture a coach who wants more than that. He begins recording foot placement to the centimeter, the precise path of the bar through the lift, how many breaths were taken before the pull, joint angles at the start and the finish, and the color of the shoes. A few of those have some real relationship to lifting. Most cannot be reproduced from one session to the next, because the subject is a person having a Tuesday rather than a rig in a laboratory. Collecting all of it takes real time. The chart that comes out the other end is dense, technical, and would look tremendous on a wall.

None of it moves the weight. You either move the weight or you don't, and what decides which one happens is a kind of learning that goes on inside the lifter, about how their own body behaves under load. The average person in the gym does not understand the bar path data and has no use for it. What they might well do, if the coach is confident enough when he explains it, is agree to have it collected. Agreeing to have something collected is a different matter from getting any use out of it.

The business version of that coach

I have spent a fair portion of my working life on the corporate edition of this, and the thing I try to get clients to see first is that the elaborate collection is almost never an attempt to deceive anyone. It is usually earnest. Somebody wants to be rigorous, and reaches for the most data-shaped object within arm's reach. The cost arrives later, when the attention that belonged on the two or three measures that actually decide something has drifted onto a wall of numbers that decide nothing at all.

A few years back I sat in on a retrospective for an engineering group that had gone fully remote and, about a month earlier, had switched on activity tracking across the team. The dashboard was excellent. Active hours up, idle time down, everything pointing the right way. Then one of the senior engineers said the thing that made the room go quiet. He had stopped reading code before he changed it, because reading does not move the mouse, and the tool scored a still mouse as nothing happening. He had been there nine years. That habit was most of the reason his changes rarely broke anything.

The research on this is not ambiguous

Personnel Psychology published a meta-analysis in 2023 that pooled 94 independent samples covering 23,461 workers, and it found no overall link between electronic performance monitoring and better performance, whether the researchers measured how fast the work went or how much of it there was.1 Stress went up reliably, and largely regardless of how the monitoring had been set up. Job satisfaction came down alongside it, as did people's sense of being treated fairly, while the feeling that something private had been taken went the other way. The authors are careful to say the effects are not always large.1 The study is three years old and resurfaced in the science press this month, because the question became a good deal more expensive once monitoring started arriving bundled inside AI productivity tooling.2

Buried in that analysis is a result worth more to an executive than the headline. The damage was not fixed in advance. Organizations that monitored more transparently and less invasively saw better attitudes from the people being monitored.1 The software was doing much the same job in both cases. What differed was whether anybody had bothered to explain why it was there, and how far into someone's working life the thing was allowed to reach.

The saying everyone quotes was written as a warning

There is an aphorism deployed to justify all of this, usually as what gets measured gets managed, sometimes improved, sometimes done, depending on who is repeating it. It is commonly credited to Peter Drucker and occasionally to Jack Welch, and the evidence for either is thin enough that the Drucker Institute has gone on the record saying he never wrote it. The trail runs back instead to a 1956 paper in Administrative Science Quarterly by V. F. Ridgway, titled Dysfunctional Consequences of Performance Measurements.3 Ridgway's argument was that quantitative measures are useful tools whose indiscriminate use, and the confidence placed in them, run well ahead of what anyone actually knows about their effects.3 He was cataloguing the damage, not recommending the practice.

Simon Caulkin put the missing half back in the Guardian in 2008, noting that what gets measured gets managed “even when it's pointless to measure and manage it, and even if it harms the purpose of the organisation to do so.”4 The line on the conference slide is the opening clause of a sentence whose closing clause was the entire point.

Velocity, and who it is for

The clearest case I run into is team velocity. It is a genuinely useful number, and it is useful to precisely one audience, which is the team that generated it. They know what a point meant to them last month, what got in the way, and which of those things is likely to recur. An outside observer, including the most attentive manager two floors up, has none of that context and therefore cannot read the number, let alone suggest how to move it. Mandating an improvement in it is stranger again, because the demand assumes there is something wrong to begin with. Improvement toward what, exactly?

Velocity can also be gamed in an afternoon by anybody who understands how it is calculated, without one thing changing about what reaches a customer. Once a measure can be moved without the underlying reality moving with it, collecting that measure is theater. It yields a number that goes up, a meeting in which the number is discussed approvingly, and no effect whatsoever on the world outside the room.

So the working rule I give clients is this. A measurement has to make sense to the people who are affected by the decisions it enables, or there is no utility in collecting it. The lifter's notebook passes because the lifter reads it and acts on it. Velocity passes inside the team and fails the moment it leaves the room. Activity tracking fails at the first step, since the person generating the data cannot use it for anything and can only be judged by it. The same failure shows up in the engagement survey that cannot find a toxic manager, an instrument that reports back whatever it was built to notice and stays silent about all the rest.

Being able to collect something has never been a reason to collect it. Before the next renewal lands, put a plain question to your own leadership team. What have you actually decided differently because of something a dashboard showed you? Not what you looked at. What you changed. If the honest answer is that the numbers have mostly reassured people who were already anxious, then you are carrying a license fee and a stress cost in order to purchase a feeling, and the people underneath it adjusted their behavior to feed the measurement a long time ago. Somebody on your team has stopped reading before they write. You will find out which one the next time something breaks that should not have broken.