(2025-02-10) Procopio The Dirty Secret About Productivity And Ai

Joe Procopio: The Dirty Secret About Productivity and AI. There’s a nasty surprise waiting for business leaders at “peak productivity”*

One of the best affirmations of real business advice I ever got was from my friend Ricky Spero, a local CEO

“I don’t worry about whether our people are working enough. I worry obsessively about whether we’re working on the wrong things. If you worked on the wrong things, your productivity was 0%”

I could have dropped a way more controversial headline on this column, like “Peak productivity is the enemy of innovation” or “How the tech industry has weaponized the chase for peak productivity.”

I want to start with this controversial shot:
Productivity can’t be optimized, because it can’t be measured. (metric)

Another way to say that is, “Hey. MBA guy. Get your bullshit formulas out of my real world.”

Fine, let’s take it literally then. Productivity, in its simplest form, is output over input.

the problem isn’t in the premise, it’s in the execution.

This is a great equation when talking about assembly lines, commodities, and repeatable tasks with known outcomes.

Next counterargument. Productivity aims to put a stake in the ground for improvement to happen.

Well, define input.

When the output in the productivity equation is a goal, or anything non-quantifiable, the input, time, becomes meaningless. When the output is quantifiable, like dollars, the equation then assumes that the input mechanism, the labor, is completely interchangeable.

Did I just disprove my own argument? That’s what leadership is trying to fix! By measuring and improving the outcomes of each labor component!
But that’s not leadership. If dollars are your only goal, you have an assembly line. You don’t need leadership

True productivity is the result of real leadership, and like all leadership-driven results, it is not quantifiable. Just like satisfaction, or loyalty, or happiness, all of which impact productivity.

next counterargument. If you’re not measuring the making of money, you’re doing business wrong.

the true numbers-based-business is the holy grail of the marriage of MBA schools and technology.

Back when I was at Automated Insights and we were building one of the first NLG/Generative AI platforms, starting in sports

Couldn’t we develop a better betting spread than the current human-driven model, which is often some Vegas guy shooting a dart and then immediately following the money as it came in?

Algorithms will tell you yes. The real world laughed at me. There are too many factors, too many variables, too many unknowns.

If sports really worked that way, where outcomes could be forecast by models, no one would care about it enough to make the bets to fuel the money to make the quest for the better betting spread worth pursuing.

Next counterargument. We can get close enough to make the right decisions.

We’re not getting close to a decision. We’re getting exactly right on two or more static, numeric axes.

we can’t escape the fact that cost is never the sole driver for the customer. If that were the case, then the cheapest product would win every time.

There are other factors, and a lot of them are non-quantifiable.

So if we want to optimize productivity, we need to stop trying to measure it, and optimize leadership instead.
And I know the counter-question: What’s the formula for that?