What Have You Never Measured?

Jun 29

The AI-powered build took an hour. The insight it required took me sixteen years.

This isn't really a post about what AI can build. It's about the strange power of naming a number you've never measured.

The Homework That Boomeranged

Two weeks ago I was preparing two teams of innovators (a public company and privately held government contractor) for a couple of AI ideation sessions. I sent them both a series of prep emails that built to a crescendo: "What's something that genuinely matters to your mission that nobody currently measures — because it's too hard, too fuzzy, or no one's ever asked?" (Yes, I was legitimately proud of that prompt.)

Then the homework boomeranged back on my own head.

Over the course of my teaching career, I've given hundreds of keynotes and led hundreds of workshops, but I’d never named the bridge between them. Honestly, I never even thought to.

Keynotes are wonderful. The room buzzes, imaginations sparkle, people line up afterward to share what they're going to try. But real talk? Inspiration isn't equipping. A keynote gets a team excited about what's possible. A workshop gets their hands dirty actually building it. And true transformation (the thing every leader who hires me ultimately wants) doesn't come from merely uncorking excitement. It comes from equipping.

Which means the metric that matters most to my mission is the bridge between the two: how often does an inspired room become an equipped one?

Hundreds of data points. And until this year, zero measurements.

The irony stings a little: I'd just sent those same teams my own piece profiling how the Portland Trail Blazers moved from efficiency to impossibility. In that piece, I argue that the real AI opportunity isn't doing old work faster, but attempting outcomes we weren't even pursuing before. I drew the conclusion for others, but I never pointed it at my own dashboard. (I'm an educator. We might be better at assigning homework than doing it.)

Naming Is the Ballgame

Why did this number hide from me for sixteen years? Not because I'd tried to measure it and failed. Because it had never been promoted to "metric" in my mind. It wasn't on my dashboard at all. And dashboards only show what somebody, at some point, decided was worth counting.

It's always been this way: the metric that matters next is invisible at first. When factories electrified, as Azeem Azhar astutely observed recently, they measured what they'd always measured — output per worker. The real transformation only arrived when Ford started measuring flow. (The number that matters most is rarely on the dashboard you inherited.)

I've spent years telling people to name their AI teammates. (Ask Russ Somers how differently he shows up for "Wendy Webinar" than he ever did for a content tool.) Naming changes the collaboration. Turns out metrics work exactly the same way. An unnamed metric isn't a low priority you've consciously deferred. It's invisible. You can't attempt what you can't see, and you can't see what you've never named.

Now, I know, I know, "not everything that counts can be counted." Granted. I'm not asking you to chase a number off a cliff. (Plenty of organizations have lost their way optimizing exactly what their dashboards told them to.) Naming a metric doesn’t mean you must enshrine it. But there's a real difference between a thing you've named and chosen not to measure, and a thing you've never named at all.

The first is a decision; the second is a blind spot.

And once you've named it, a loop opens up: you name the number, you build a modest attempt to move it, and then you measure to find out whether the move actually worked. Three beats. The naming is the one nobody does.

The Part That Didn't Flatter Me

Once I named it, the build was almost embarrassingly easy. My AI teammate already had what it needed: my prep-call transcripts, the audience, what leaders said they were wrestling with. (It had been sitting on all of it the whole time, politely waiting for me to ask.) Now, after every keynote, it drafts a customized proposal for a hands-on workshop designed to equip that exact team. Not a boilerplate proposal, but a customized offer that knows the room. One hour, start to finish.

That proposal engine wasn't built to watch the number. It was my brand-new attempt to move it. The whole point was to turn more inspired rooms into equipped ones. So once it was running, the real question wasn't "did I build the thing?" It was "is the thing I just built working?"

And then the measuring began. (Part of me wishes it hadn’t.)

The first audit didn't flatter me: customized proposals had followed nine of my last ten keynotes — and so far, exactly zero had converted to booked workshops. Painful? Sure. (The part I’m still sitting with now is that building the intervention was trivial, but building it didn't mean I'd necessarily solved the problem the measure revealed. An easy build is not a working one. Those are two different victories, and I'd only earned the first.)

But behind that unflattering number is a glimmer of hope: three of those proposals opened advanced equipping conversations with organizations I never would have heard from again. No new revenue yet. But brand-new conversations that simply didn't exist before I named the number, and then deployed an AI-powered solution to drive it. Measurement isn't supposed to flatter you. It's supposed to inform you.

As I said earlier, I really do believe that the build was the trivial part. The naming was the ballgame.

Your Turn

I suspect you have one of these, too. A number that would genuinely move the needle for your business or mission that has never appeared on any dashboard you own, either because it's awkward to count, or it falls between two departments, or nobody ever asked.

Don't ask, "How could AI improve our numbers?" (To be fair, that one is also interesting, but not nearly as interesting as…)

Ask, "What number have we never paid attention to?"

Then finish this sentence:

"A metric I've never measured is _______."

Don't worry about whether it's perfectly measurable yet. (Mine wasn't. Mine still isn't flattering.) Name it first. Then craft a modest AI-powered attempt at measuring that metric.

And because the build really is the trivial part, here's the whole thing in one prompt:

"Here's a metric I've never measured: [your named metric]. You already have access to [the raw material that's been sitting there — transcripts, CRM, calendar, call notes]. Build me the smallest possible intervention that actually tries to move this number after every [recurring moment — keynote, sales call, sprint, shipped feature] — and have it track whether it's working, so I find out the honest answer either way."

You're probably not even one tool away. You're just one named metric away. (The hour-long build takes care of itself. Whether it works — that's the part the measure will tell you.)

Serious question: what have you never measured?

Join over 50,147 creators & leaders who read Methods of the Masters each week.