How Big Things Get Done

Well, I made a paper model after reading this one and talked to a bunch of people about it as I was reading it.

In the subtitle, "the surprising factors that determine the fate of every project", the "surprising" is clickbait. But that's a good thing I guess, maybe means I've learned something over the years. Nevertheless, the book is dramatic, interesting and quite readable.

Flyvbjerg, a professor and consultant for megaproject management, compiled a huge database of megaprojects. This book seems to be the fit-for-public compilation of his research an experience. Packed with fascinating stories of great success (Empire State building, Guggenheim Bilbao) and hard(ly) success (Sydney Opera House, Heaven's Gate), interviews with Edwin Catmull, Franck Gehry and Daniel Kahnemann.

I think the practical point that stood out most for me is that the most accurate forecasting method is to refer to other completed projects of the same class, the more the better. This helps account for unknown unknowns & black swans which we otherwise don't factor in due to latent human optimism.

So read about how the Sydney Opera house went 14 times over budget and the architect was forced out half way, leaving Australia never to return. Read about how Pixar films go through ~8 iterations of feature-length, hand-drawn slide shows before being developed. And how solar and wind are the most reliably on budget of all project classes.

(And this is definitely an engineering book. While conscious of climate change, it goes on to suggest carbon capture and storage as a serious technology. And one of the examples of success is the expansion of Heathrom terminal 5... it is steeped in Capitalism Realism. Still valid and useful and helpful for doing stuff, just don't come for your critical theory.)

If the title is a question, Brent has collected data across thousands of large projects and found an answer that he reveals early. Big things get done over budget, late, and deliver less value than people expected. Or they don't get done. For the most part. Not by a little bit, either - big things fail by a lot. In a database of "16,000 projecgts from 20-plus different fields in 136 countries" he finds that "99.5 precent of projects go over budget, over schedule, under benefits, or some combination of these."

And it shouldn't be this way for big things. These are HUGE EXPENDITURES. Stuff like dams, nuclear power plants, healthcare.gov, and similar massive projects that people depend on succeeding. There should be lots of incentives to get it right.

Brent explores why this happens over and over again. He doesn't duck the question, he has real answers, like:

• Many projects start without defining why they need to happen. Without a defined why, decisions aren't always steering to the north star and not everyone will have the same vision.
• People act too early and don't plan well before they get started. He goes into how to do that better - pointing out how Pixar did it really well with low stakes small versions.
• When people get started early, they get delayed on every new emergency, so it takes longer than expected - and every day late is a chance for a new unexpected incident or emergency. Fast is good once you have a tight plan.
• People estimate poorly because they think their project is special instead of finding a good match of things that have been done before.
• People lie about the costs so they can use sunk costs to demand more funding when they exceed the budget. (Robert Caro describes how Robert Moses did this repeatedly in The Power Broker). Some "estimates aren't intended to be accurate; they are intended to sell the project."
• The team doesn't have enough experience so they can't anticipate that similar problems will show up as happened on previous similar projects. If you've never built an underground train before, you won't know to keep replacement parts for the digging machine ordered in advance.

The stories are numerous and detailed in here. I was also pleased by how many echoes I found with my other reading. I'd read about the building of the Pentagon and the chaos of the the early building of it. So much was done so fast, but so much had to be redone. I'd listened to the 99pi/Cautionary Tales episode about the building of the Sydney Opera house and how it ruined the architect, who fled the country and never saw it completed. I'm reading Robert Caro's "The Power Broker" and it's a great peek behind the scenes to see how and why the author thought his 8 years of research/writing would only take a year.

This book was a big eye-opener for me, articulating things my experience had turned into instincts. I was able to use it before I even finished, referring engineers I work with to find "reference projects" to figure out timelines instead of doing the same things that didn't work for us and don't work for anyone else.

If you do things that take longer than a month, this is a pretty valuable book to read. If you do things that are new or haven't been done before, it's VERY interesting. I'm going to go on about this a lot to other folks about it.