I program 5-axis CNC machines for fun.
That sentence either makes you curious or confused, and your reaction probably says a lot about how you think about the relationship between marketing and technology.
The False Divide
Somewhere along the way, business culture decided that "marketing people" and "technical people" are different species. Marketing does brand. Engineering does product. Never the twain shall meet.
This was always a lazy mental model, but it was functional when marketing meant buying media and engineering meant writing code. Those boundaries have dissolved.
Today's marketing runs on:
- Data pipelines that feed attribution models
- API integrations that connect ad platforms to CRMs
- ML models that optimize bidding and personalization
- Custom software that automates campaign management
If you can't read the architecture diagram, you can't lead the team building it.
What CNC Machining Taught Me About Marketing
Programming a CNC machine is an exercise in extreme precision. You're writing instructions (G-code) that tell a cutting tool exactly where to move, how fast to spin, and how deep to cut. A mistake of 0.001 inches can destroy a part worth thousands of dollars.
This discipline translates directly to data-driven marketing:
Toolpath thinking — In CNC, you plan the sequence of cuts to minimize tool changes and maximize efficiency. In marketing, you design customer journeys to minimize friction and maximize conversion. Same pattern, different material.
Tolerance management — Every CNC operation has tolerances: acceptable ranges of variation. Marketing metrics work the same way. A 2% conversion rate might be acceptable for cold traffic but unacceptable for retargeting. Knowing the tolerances lets you focus on what actually matters.
Feedback loops — Modern CNC machines have probes that measure the workpiece during cutting and adjust in real time. This is exactly what we're building with AI-powered marketing: systems that measure, learn, and adjust autonomously.
The Pattern
The best marketing leaders I know think in systems. They see inputs, processes, outputs, and feedback loops — not just campaigns and content. Technical fluency gives you this lens.
The Builder's Advantage
There's a massive difference between a CMO who commissions a dashboard and a CMO who builds one. Not because the dashboard is better (it might be worse, actually) — but because the builder understands the constraints.
When you've built a data pipeline, you know:
- Why certain metrics are delayed
- What data quality issues look like
- Where the integration points can fail
- How to ask engineering for what you actually need
This understanding transforms every conversation. You stop being the person who asks "why can't we just..." and start being the person who says "what if we..."
The Five-Year Horizon
Here's my prediction: within five years, the majority of new CMO hires at growth-stage companies will have shipping code on their resume. Not because they'll be writing production software — but because the ability to prototype, automate, and understand technical systems will be table stakes.
The CMOs who can:
- Build an AI agent that manages ad spend
- Design a data model for customer attribution
- Prototype a tool for their team using Claude Code or Lovable
- Read an API specification and understand its implications
...will outperform those who can't. Not marginally — dramatically.
What to Do About It
If you're a marketing leader who doesn't write code, here's my honest advice:
- Learn Python basics — not to become a developer, but to understand what's possible
- Build something small — a personal dashboard, an automation script, a simple tool
- Use AI coding tools — Claude Code and Cursor lower the barrier dramatically
- Study systems thinking — Donella Meadows' Thinking in Systems is the best starting point
- Get uncomfortable — the learning curve is real, but the payoff is exponential
The CMOs of 2030 are learning to code right now. Some of them are also learning to program CNC machines. I happen to think that's the more interesting path.
Yes, I really do program 5-axis CNC machines. I'm currently designing a custom CNC controller using LinuxCNC, ESP32 microcontrollers, and BLDC motors. The project is called Helios, and it's teaching me more about systems engineering than any management book ever could.