You Can’t Measure What Really Matters

My resting heart rate is pretty good. I slept like shit last night, got enough protein this morning, and hit my move goal three times this week. I weigh about what I did in college but can’t bench what I could then, have great cholesterol for a guy my age, and spent too much time on Instagram Tuesday.

There’s no escaping the quantitative optimization of middle age. Maybe it’s a good thing. Maybe future me gets a few more years of mobility before the wheels come off. But is my life better for it today?

That’s not a shot at the technology. Instrumenting yourself works because it changes your behavior. Measure your sleep and you’ll sleep more. Measure your drinking and you’ll drink less. Measure your training and you’ll show up Sunday morning with the other MAMILs (Middle Aged Men In Lycra) and drop a couple of them on the climb. Some of it might buy you years, or make the last ones better.

But time and life are different things. Nothing on my wrist has data on the depth of my relationships, my progress toward being the man I want to be, the quality of what I learned this week, or whether the people closest to me are ok. Those are the things that make the extra years worth having. They’re also the ones nobody’s figured out how to track.

So we optimize what we can count, and assume the rest will follow.

It doesn’t.

The Same Trade, at Work

Your startup is making the same trade with AI.

AI will take some labor out of your workflows. It will draft the first pass, answer the tier-one ticket, reconcile the account, summarize the call, and eventually let you run the same business with fewer people in it. That’s real, and if you’re not doing it, you should.

Just be honest about what you’re buying. You’re buying parity. Your competitor has the same models, the same wrappers, the same playbook that shipped last quarter, and roughly the same prompts. Everything AI is great at today becomes table stakes tomorrow, because everyone will have it. An advantage anybody can put on a credit card isn’t an advantage for long.

What it won’t do is build you a brand, meaning what people say about you when you’re not in the room. That comes from strategic clarity, message discipline, and accumulated evidence of how your product and your people perform under pressure. No model can generate that for you. Only people can, working together, over time.

It won’t build you a winning culture either. Not the values on the slide — what people do when you’re not in the room. That gets built one uncomfortable decision at a time. Who you promote. Who you let go. What you refuse to tolerate from your best performer.

The Most Expensive Half-Truth in Management

In business school I learned you can’t manage what you can’t measure. Mostly true. Also the most expensive half-truth in American management, because of what it implies and never says: that if a thing can’t be measured, it probably doesn’t matter much.

Look at what’s on the other side of that line. Trust. Taste. Nerve. Whether your best engineer is still here in two years. Whether anyone tells you the thing you don’t want to hear while there’s still time to do something about it. Whether a customer would go out of their way for you.

Try building a dashboard for that.

Sustainable Advantage

In a world where almost anyone can build almost anything, managing what can’t be measured is the whole game.

Pipeline velocity still matters. So do gross margins, and NRR, and the rest of it, and AI will help you with every one. It’ll help your competitors just as much. What it can’t do is make the good decisions for the long run that don’t show up on the quarterly dashboard.

Those take a leader willing to say out loud that some of what matters most will never end up in the board deck — and then to defend the time, attention, and operational slack they need. The offsite with no agenda. The one-on-one that isn’t a status update. The hire who scores worse on paper and is obviously right. The quarter you spend on something that won’t pay back this year.

Anthony Bourdain hated planning meals before he got where he was going. Asked what travel had taught him, he said:

“Don’t be afraid to eat a bad meal. If you don’t risk the bad meal you never get the magical one.”

The equivalent in startups — and in middle age, I’d argue — is being willing to give up a few units of productivity in service to a few units of humanity. Leave a great person on the CS team for now, meet that friend for drinks. Optimize everything you can count and you may end up marginally better off than the guy next to you.

But you’ll never get the magical one.