How Big Things Get Done

Funny that a core principle here is 'make your projects small and modular, and then they can get big', and it's id=1000 in my personal book tracking site.

Noted on January 30, 2024

Every big project - like tunnels, rail lines, etc, is - super late and super over budget. This includes big IT projects like ‘converting the court to use digital records’ Why?

It wasn’t always the case. Empire State Building: 50M budget, done for 41M, on schedule. But actually that’s a legendary outlier.

1.
IRON LAW of PM: every large project is over budget, over time, under benefits.

Forcing timelines doesn’t work
Forcing budgets doesn’t work

Bent Flyvbjerg studied megaprojects all over the world and has built huge database of projects - time, budget, case studies, etc. He’s legitimately an expert.

2.
Commitment Fallacy screws up projects: not adjusting yr project after you get new info.

Bias towards action is great but will kill you if it’s not reversible. Act fast when you can pivot. THINK fast and act slow when you can’t.

Many big projects are intentionally underscoped so they can get started. The true cost and time would never be approved, so a fake smaller number is used. Then once there’s a big hole in the ground, time and money to finish will be found.

3.
Start w Why, think from Right to Left. Eg, start w goal and work backwards. Examples: house remodeling story, Bilbao Guggenheim started w a renovation but pivoted to a new site.

Amazon pitches new internal projects by writing a Press Release and a FAQ, to envision the final benefits concretely. Then they rework those in cycles until everybody’s on board with a raison d'être

4.
Sydney Opera House is classic case of bad plan, start digging a hole anyways.

Pixar plans out movies, big iterative cycles and feedback before they start animating anything.

Lean startup principles help. Plan, prototype, test, as quickly as you can.

5.
Experience matters. Young people are more likely to think things are easy. “Nobody’s done this before!” Is a warning sign, not a rallying cry.

6.
Your project is not unique. Hong Kong XLR/MTR project example. Forecasting and budgeting are hard but even harder if you're extrapolating out a small project into a large one.

Getting your reference class correct is the first step. Don’t estimate out from the ground up (a kitchen remodel has x feet of floor space, sf of countertop, etc), but instead start from a set of completed references. You’ll get better estimates. It can take out bias and compounding errors.

People don’t do this because it’s simple, and the hard way seems more correct. Don’t calculate; anchor-and-adjust your numbers. RCF = Reference Class Forecasting.

Pay attention to if your project is within a normal distribution or a long fat tail distriubition. If fat tailed, then look at the ‘black swan’ events that led to those cases, and see if you can be proactive about them.

Chicago Fire Festival - burn a fake house down to celebrate the chicago fire. Spent all their time mitigating how things could go wrong, but didn’t put enough planning into going right; the fire failed to start.

7.
Ignorance is not your friend. Is it better to not know and just start? Lots of compelling stories about this (Electric Lady, Jaws, etc) but the data doesn’t back it up. Sydney Opera House was a financial disaster but also killed the career of the architect!

8.
You need a single unified team. Your different stakeholders and contractors need to feel like they’re all on Team Project. Align incentives, etc

9.
Wind and Solar projects seem to be pretty good about not failing. That’s because they’re made of modular components that can be combined into a giant project. Find what your lego unit can be and get good at that brick. Then your improvements compound as you combine that module into a bigger project.

Tesla Gigafactory 1 in texas is actually 21 small factories.

Also in Bent’s category of thin-tailed projects (more likely to go right): roads, transmission lines, thermal energy. Small components multiplied out.

Many small things can be more effective than one big thing. (and much more likely to get done)

CODA
11 heuristics (ie mantras!)
- hire a master builder
- get the team right
- ask why
- build with lego
- think slow act fast
- take the outside view
- watch your downside
- say no and walk away
- make friends and keep them friendly
- build climate mitigation into your project

Noted on January 30, 2024