Manufacturing Mindset

Where this came from

This one starts on a catwalk above a Toyota Production System floor in Minneapolis, watching green, yellow, and red lights tell the truth about every station on a line that had been running for decades. Raw materials went in on one end. Finished product came out the other. Every step was visible, measurable, and improvable. Nobody argued about whether a machine was down — a red light said so.

Then I walked back to an analytics floor. Same instinct to build, none of the same discipline to see. Work arrived from every direction — Slack, email, hallway asks, an exec's afternoon text — and nobody could name what was actually in progress versus stuck. Finished products, the dashboards and models and analyses, disappeared into the void with no signal on whether they'd helped anyone decide anything. A factory with no lights.

What 'treat data like a factory' actually means

Most analytics teams believe they are in the data business. They are in the manufacturing business, and the sooner a team accepts that, the sooner it stops accepting waste as the cost of doing business. Data is raw material sitting on the left side of the warehouse. Pipelines and transformations are the machines in the middle. Dashboards, models, and insights are finished products that either sell or gather dust on the right.

Once you draw it that way, the questions change. Not "did we ship it," but "what is our yield" — what percentage of raw data actually becomes something someone uses. Not "is the pipeline built," but "what is its uptime." A factory that tolerated 80% waste would get shut down. Most analytics orgs run at exactly that ratio and call it normal, because nobody ever drew the line and asked where the losses were happening.

The habit that makes it real: walk the catwalk

Manufacturing executives walk the floor. Not a scheduled demo — an unannounced look at the actual work, in production, as a user experiences it. Most data leaders never do the equivalent. They get status reports that say green across the board, and they believe them, because the alternative is admitting they don't actually know what their platform feels like to use.

The fix costs two hours a week. Open a random dashboard as if you were the business user waiting on it. Search your own data catalog for something ordinary. Sit next to an analyst for thirty minutes and watch, don't ask. What surfaces is never what the status report said: slow loads, confusing names, silent workarounds, the export button doing more work than the dashboard itself. None of that shows up when you only read about the work. It shows up the moment you go see it.

Where this shows up in the book

This is the frame underneath nearly everything Sarah Chen does at ClassicMotors.io, made explicit when she finally gets her successor onto the factory floor herself — and it is the direct subject of the first leadership essay, where the Minneapolis catwalk moment gets told in full.