My first delivery pipeline was shipped through the mail. I couldn’t avoid the mail at the time, but I did shrink the delivery process from 1 week to 4 hours in a few days. We’ve come a long way since then, and yet in many ways, many teams are often driving blind. In an age of near-limitless instrumentation, automation, and observability, we can still miss the forest for the trees. Many teams are drowning in data, and blindly automating their way under a pile of expensive errors. Implementing massive or disconnected systems is a huge, costly undertaking.

As we look to observe, inspect, correlate, forecast, detect, analyze and improve, what’s missing are cheap/easy/effective methods of seeing the big picture, choosing a destination and finding our way.

How can we decide what to observe and measure, and where? What can we do to improve flow before we have full observability? How can we build a roadmap based on where we are right now? This talk will introduce a few principles and practices I’m calling ‘Flow Engineering’ as a quick and effective way to get your bearings with Delivery improvement.

I’ve started using Flow Engineering after years of struggling to visualize and improve delivery pipelines of all kinds and communicate ideas, issues, and risks. It has helped me many times go from tearing my hair out (and I have the baldness to prove it!) to knowing exactly how to level up.