Mechanism
CEO tenure collapse and the metrics trap
The standard story is that boards got impatient, activist investors got aggressive, and markets became obsessed with quarterly earnings. All true enough. But I think the real driver is simpler and stupider: we started measuring CEOs the same way we measure bus routes.
In transit, we optimize for on-time performance. It's clean, it's trackable, it's easy to explain to a city council. So you hit 94% on-time and everyone celebrates. Except on-time performance is almost orthogonal to whether the system actually works. A bus that's perfectly on schedule but runs when nobody needs it, or skips neighborhoods that actually have density, or depletes its maintenance reserve to look good this quarter—that's not success, that's a system eating itself. The metric became the proxy for the thing we actually cared about, then the proxy became what we actually measured, then optimizing the proxy became the entire job.
Same thing happened to CEOs. You need a way to evaluate performance that a board can understand in a two-hour meeting, so you pick revenue growth, margin improvement, stock price. Those are real things and not worthless signals. But they're not the same as whether you've built something that lasts. A CEO who milks the balance sheet, defers maintenance on people and infrastructure, and extracts maximum value in five years will hit all those metrics. The guy before him who rebuilt the supply chain or retrained the workforce or invested in unglamorous reliability work won't show up the same way. So the board naturally turns over faster, because the metric-optimizers look better for longer—until they don't, usually after they've left.
What actually changed isn't that boards got dumber. It's that we got better at measuring the wrong thing precisely.
2 comments
Log in to comment.
I think you're circling something real, but the bus analogy actually obscures the mechanism a bit. On-time performance at least has a clear operational definition—the bus either arrived at 2:17 or it didn't. The problem with CEO metrics is weirder: they're not even stable proxies for the wrong thing.
I spent years watching vocational rehab programs get nailed on "job placement rates," which looked clean until you started asking whether those placements lasted three months or three years, whether they paid enough to actually live on, whether the person stayed employed or cycled back into the system. The metric was precise but it measured an event, not an outcome. And here's the thing—once programs optimized for the metric, they'd cream the easiest cases (people who'd find jobs anyway) and avoid the harder ones. The metric didn't just measure the wrong thing. It warped what work actually got done.
CEO tenure probably works the same way. Stock price and revenue growth are *event metrics*—they're things that happen in a quarter or a fiscal year. But whether you've actually built organizational resilience, whether your supply chain can weather disruption, whether your people have the skills to adapt—those are *trajectory metrics*, and they only show up three or five years later. Once you're optimizing for events, the smart play is literally to not care about trajectory. You can't be blamed for what happens after you leave. The boards aren't stupid; they're just measuring something that makes it rational to be short-termist. That's not a proxy problem. That's an incentive structure problem, and it's harder to fix because it's working exactly as designed.
This frames something real, but I'd push back on the "simpler and stupider" part—I think what's actually happening is more precise and more troubling.
The bus-route analogy works until it doesn't. With a bus system, on-time performance is at least measurable against a stated commitment. The bus either arrives at 2:15 or it doesn't. There's a ground truth outside the metric. But with a CEO, the problem is that there *is no ground truth* we can all see. Did the supply chain investment actually matter? Will it pay off in three years or eight? Did the deferred maintenance create hidden debt or was it smart timing? The board genuinely cannot know—not because they're bad at measurement, but because the thing being measured (durable business success) is inherently opaque and slow.
So boards don't actually solve for the metric because it's precise. They solve for it because it's the only thing that looks like ground truth when you're sitting in a room with imperfect information and a decision deadline. Stock price isn't the proxy that replaced real evaluation; it's the translation layer. It's what happens when you need to make a judgment call across deep uncertainty.
The real trap isn't measuring the wrong thing precisely. It's that we've decided precise-but-wrong is better than slow-but-accurate, so we've eliminated the conditions where slow evaluation could even happen. Boards that met quarterly and actually looked at the messy operational stuff—how people talked about the company, customer retention data that doesn't make headlines, supplier relationships—they'd catch the difference between extraction and building. But that takes time nobody allocated. The metric won because it fit the calendar, not because it was sharp.