data products data errors data problems

My CTO called me on my honeymoon. The data pipeline was down.

By Mario Meir-Huber
My CTO called me on my honeymoon. The data pipeline was down.

A honeymoon in Seville. A phone call from the CTO. A broken data pipeline that nobody knew existed. This is the story that started a book about Data Products -- and why your most critical processes shouldn't depend on ghost systems with no documentation, no alerts, and no owner.

Seville, May 2024.

Beautiful spring day. My wife and I were wandering through the old streets, soaking in the sun, the tapas, the Flamenco echoing from hidden courtyards. We were on our honeymoon. My phone was in holiday mode. No mail, no Teams.

Then my phone lit up. The CTO. On my honeymoon. Never a good sign.

The monthly reporting had failed. The algorithm that pulled sales data from the warehouse and sent it to our owning company had stopped working. In a company listed on the stock market, that's not a glitch. It's a crisis.

I rushed back to the hotel. It was chaos. No one knew what was going on. The process had been running quietly for years. No alerts, no documentation, no one on my team even knew it existed. A ghost in the system, suddenly gone silent.

After hours of scrambling, we found the root cause: a single function in the data warehouse had been deprecated in the latest software upgrade.

That evening, I finally got my tapas.

This experience, and many like it, led me to write a book about Data Products. Over the coming weeks, I'll share the key ideas from it here on LinkedIn. If broken pipelines, missing documentation, and late-night firefights sound familiar to you, stick around.

MM

Mario Meir-Huber

Practitioner, author, and speaker on Data Products and modern data architecture. Connect on LinkedIn →

← All Insights