Moneyball
In 2002, the Oakland A's had the lowest payroll in baseball and had just lost their three best players to free agency. The conventional response was obvious: rebuild, tank, wait.
Billy Beane did something else. He stopped buying players and started buying runs.
The insight wasn't complicated. Baseball had spent a century evaluating players on batting average, stolen bases, RBIs — stats that felt like winning. Beane asked a different question: what actually produces runs?
Runs win games. And OBP was undervalued because nobody was looking at it.
Engineering organizations spend millions on tools that measure velocity, cycle time, deployment frequency — stats that feel like productivity. But none of them measure why teams are slow. The communication breakdowns. The specs that get reinterpreted. The context that decays between kickoff and deployment.
Everyone's buying players.
Almost nobody's buying runs.
System Over Stars
Phil Jackson inherited Michael Jordan — the most talented player in basketball history — and didn't build the offense around him. He installed the triangle: a system so complex it took years to master, designed not for stars but for spacing.
The triangle made role players dangerous. Suddenly Steve Kerr wasn't just a spot-up shooter; he was a decision-maker in a system that created advantages faster than defenses could react.
Most organizations rely on heroics. The senior engineer who "just knows" the codebase. The PM who can "translate" between teams. But heroics don't scale. Systems do. When communication is clear — when Product and Engineering actually share context — role players become dangerous.
The best teams don't have better players.
They have better systems.
Speed as Strategy
When Mike D'Antoni took over the Phoenix Suns in 2004, basketball was slow. Teams walked the ball up, ran set plays, ground out possessions. The Suns had Steve Nash — a point guard who could see passes before they existed — and D'Antoni had a theory: what if we just... went faster?
Seven seconds or less. Get a shot up before the defense sets. Turn every possession into chaos — but controlled chaos.
The league wasn't ready. The Suns didn't have the biggest stars or the deepest bench, but they had tempo. They made decisions before opponents could react.
Most engineering problems aren't solved by adding process. They're solved by catching misalignment before it compounds. By the time you're debugging a production incident or unwinding a misbuilt feature, you've already lost. The advantage goes to teams that detect drift early — in the handoff, in the spec review, in the first sprint — before defense sets.
Speed isn't about moving faster.
It's about seeing sooner.
Why Process Didn't Fix This
Here's what the basketball analogies don't capture: software already has its own version of systems thinking. We've been running plays for decades.
Each promised the same thing: a system that would make teams faster, more aligned, more predictable. And each solved something real — waterfall was broken, silos were deadly, coordination does matter at scale.
But none of them measured whether Product and Engineering actually understood each other.
They assumed alignment. They built rituals around the assumption. Sprint planning assumes shared context. Refinement assumes terms mean the same thing to everyone in the room. The entire edifice assumes that when Product says "simple" and Engineering hears "simple," they're picturing the same thing.
They're usually not.
The triangle offense didn't just add more plays. It changed what players paid attention to — spacing, timing, decision-making in motion. Scrum added more meetings. SAFe added more meetings about meetings. But nobody instrumented the thing that actually breaks: the translation layer between intent and execution.
Process can't fix a measurement problem.
Everyone Had the Data
Here's the strangest one. The Golden State Warriors didn't discover the three-pointer. It had been worth three points since 1979. Every team had the same data, the same shot charts, the same analytics departments.
But the Warriors acted. They built a roster around shooting. They spaced the floor differently. They took threes that other teams considered bad shots — and made them anyway.
The data about Product-Engineering misalignment exists. It's in your GitHub metadata. Your Jira comments. Your Slack threads. The patterns are visible — which specs generate the most back-and-forth, which terms consistently get misinterpreted, which teams drift furthest from original intent.
Nobody's looking at it.
Not because it's hidden. Because it's unscouted.
I'll be honest: I can't remember basketball between Jordan's second retirement and the Suns changing everything. The game existed, but it felt like teams were running out the clock on an era that had already ended. Then Phoenix showed up playing like the shot clock was a suggestion, and suddenly basketball was interesting again.
The Warriors hit differently. Everyone in the league was thinking about the three-pointer. Analytics departments had the same shot charts, the same expected value calculations, the same data. And Golden State still cracked the code first. They didn't just understand the math — they reorganized everything around it. Roster construction. Spacing. Shot selection. Culture.
By the time the rest of the league caught up, they had four rings.
Here's what stands out to me across all of these examples: analytics never replaced the point of the game. The point is winning. Moneyball wasn't about OBP for its own sake — it was about runs, which produce wins. The triangle wasn't about spacing as an abstraction — it was about championships.
The tension between analytics and intuition, between measurement and feel, is real. But it's healthy. The best organizations hold both.
I believe the same thing about DevEx. DORA metrics, SPACE frameworks, burn-down charts — these gave us visibility into things we'd never measured before. That's valuable. I'm not here to tear them down.
But they measure outputs. Cycle time. Deployment frequency. Velocity.
I want to help organizations win.
And winning means understanding why you're slow — not just that you are. It means catching the misalignment in sprint one, not sprint five. It means measuring the translation layer between what Product intended and what Engineering built.
The scoreboard matters. But the game is won in the possessions nobody's tracking.
The Unscouted Variable
DevEx tools tell you that you're slow. They can't tell you why.
They measure the scoreboard — cycle time, PR merge rate, deployment frequency. Useful numbers. Lagging indicators. Symptoms.
But the game is won in the possessions nobody tracks. The miscommunication in the kickoff meeting. The term that meant one thing to Product and another to Engineering. The context that decayed over three sprints until the feature shipped technically correct and fundamentally wrong.
Moneyball found runs hiding in on-base percentage. The triangle found championships hiding in spacing. The Suns found advantages hiding in tempo. The Warriors found a dynasty hiding in a shot everyone else ignored.
What's hiding in your organization?
The variable that changes everything is usually the one nobody's measuring.