I told you in yesterday’s update that I had finally, finally finished the book I have been working on, Don’t Screw Up Your Program. This was supposed to be an “easy” book — one that I thought I could whip out in a few weeks because the experience it was built around was so recent.
And also traumatizing. As in, it turned out I needed months away from it before I could write with any sense of detachment. Or without throwing things at the computer.
But it’s done now except for edits, and several reviewers have it in hand to give me their feedback. Since it’s a workbook, with pages to fill out to help you plan your program, it will need a bit more graphic design work, but at this point I think it will be fully out by September. You can sign up for advance notice and pre-order opportunities here.
One of the points that underpins the book is that if you make bad assumptions in the beginning, that will trigger a cascade of problems for you - and the people you’re trying to help. One seemingly small thing can result in the whole system crashing, if you’re not paying attention.
One of the examples that business schools still use to teach about how cascading errors happen was the Great Northeast Blackout of 2003. And since I was pretty much at the place where that started when it started, it made sense to start there.
Here’s the story as it appears in Don’t Screw Up Your Program. I hope you’ll notice how utterly normal, mundane, unexceptional the experience was.
Until it wasn’t.
An anatomy of Cascading Errors
It's the summer of 2003, and I am visiting my mother outside of Cleveland, Ohio with my 2- and 5-year old sons. One particularly warm afternoon, bright and sunny, the electricity shuts off.
It's an old town, in an area with an old power grid. An outage on a sunny day is a little weird, but not unheard of.
As the sun set, as we sat on the porch and fanned ourselves, we noticed something strange.
The streetlights weren't coming on, either.
We'd never seen that happen before.
I don't remember how we found out about the bigger story that we were playing a tiny role in, along with millions across the eastern U.S. We weren't just having a power outage - we were experiencing the most massive breakdown in the national electrical grid, ever.
It took weeks after the blackout for the whole timeline to get untangled. When that happened, we learned that it started with a routine, miniscule glitch: a tree branch (ironically, only a couple of miles away from my mother's house) sagged onto a power line. That caused the local system to fail, which caused the regional system to overload and fail, which pushed most of the eastern U.S. and Canada into complete blackout.
All of that turmoil caused by one branch sagging under its own weight.
Business writers call this kind of situation cascading errors. One seemingly minor thing goes wrong, but because of the way systems have gotten interconnected (often in bits and pieces over time, without anyone thinking ahead about how those connections might play out), that glitch turns into a disaster.
And if you look at the post mortem of most disasters, you will find cascading errors.
More times than not, cascading errors happened because no one thought that chain of events was possible. People assumed that everything would work as it was intended, without asking the key question:
What if it doesn’t?
—--
For us, the power outage was uncomfortable, but not a crisis. My kids remember eating dinner and playing on the house's big brick porch without having any idea why that happened.
Three years later, however, my mother was using an oxygen system that required electricity.
If the timing had been different, that cascading error could have threatened her life, as it did thousands of poor, unhoused, sick and disabled across the eastern United States. Nearly 100 people died as a result of that power outage, and losses in business revenue, wages, spoiled food and more were estimated to reach nine digits.
Your program probably won’t run risks like that. But the world in which your program operates is even more VUCA (volatile, uncertain, complex and ambiguous) than it was in 2003, when some of the solutions that your predecessors have tried were still in someone’s imagination – and politics, businesses, communication and more only partially looked like they do today.
So the likelihood of cascading errors is even higher today than it was during the Great Blackout.
And way higher than you might think.