Book Summary: How Big Things Get Done by Bent Flyvbjerg and Dan Gardner

Book Cover for How Big Things Get Done by Bent Flyvbjerg and Dan Gardner

Why did the Sydney Opera House ruin its architect’s career? What can we learn about megaprojects from Pixar’s movies? This summary of How Big Things Get Done by Bent Flyvbjerg and Dan Gardner explores why so many megaprojects are disasters and how to prevent these failures.

Buy How Big Things Get Done at: Amazon | Kobo (affiliate links)

Key Takeaways from How Big Things Get Done

  • Most megaprojects fail. Less than 1% of megaprojects come in on time, on budget, and deliver the claimed benefits
  • Planning is essential to a megaproject’s success:
    • Megaprojects are typically rushed, with only superficial or slapdash plans. There’s often an action bias — people want to see “shovels in the ground”.
    • Some argue that we underestimate our creativity under pressure, and that planning just leads to paralysis. But the data doesn’t support this.
    • Others argue that planning obstructs creativity. However, the authors argue planning enables creativity as you can experiment during the planning stage, which is relatively cheap.
  • How to plan well — good planning is:
    • Broad. Start by asking what your goal is. Projects are a means to a goal, not a goal in themselves. Are there other ways of achieving your goal?
    • Slow and detailed. Detailed planning may be expensive in absolute terms, but cheap relative to delivery. You want to flush out as many problems as possible in the planning stage. Bad planning leaves problems to be figured out at the delivery stage.
    • Active. Planning is not just sitting and abstractly thinking about what you’ll do. It’s about making a model or simulation, testing it and refining it.
    • Risk mitigation. You can study the black swans that tend to come up in your type of project and plan for them. One way is to simply reduce the project’s duration, leaving less time for black swans to crop up.
  • Experience is underrated, usually for two reasons:
    • Domestic bias. Politicians prefer to award contracts to local companies to create local jobs, instead of outsourcing to foreign companies who have the relevant experience.
    • Uniqueness bias. We like words like “biggest”, “unique” or “bespoke”, but we should really be wary of them because anything unique is inexperienced.
  • The megaprojects that go wrong can go really wrong (the distribution is “fat-tailed”).
    • However, modularity can reduce risks. If you can break down your project into smaller, repeatable chunks, you can learn as you go and things won’t get too far off track even if you screw up.
    • Modularity is a matter of degree. Some people are now looking at ways to make traditionally non-modular projects like nuclear or hydroelectric dams more modular.

The authors sort of summarise their own book by setting out 11 heuristics for better project leadership at the end. But, frankly, I think my summary is better 😊

Detailed Summary of How Big Things Get Done

Most megaprojects fail

Megaprojects are simply very big projects. “Big” is relative; for most people, a home renovation can be one of the most expensive and complex projects in their lives.

The vast majority of megaprojects fail to deliver the expected benefits on time and on budget. Based on a database of over 16,000 projects around the world that Flyvbjerg and his team compiled:

  • 91.5% of projects go over budget or are late;
  • 99.5% of projects go over budget, are late, or fail to deliver the stated benefits.
Example: Californian High-Speed Rail

In a 2008 referendum, Californians approved a high-speed rail project to connect Los Angeles and San Francisco, which is about 350 miles apart by air. The project was expected to cost $33 billion and be ready by 2020.

The project is still not completed. In 2019, the Californian governor announced that they would only complete half the route — 171 miles between Merced and Bakersfield, at an estimated cost of $23 billion. As of 2022, the time of writing, the current highest estimate is that it will cost $100 billion, but no one really knows for sure.

Map showing distance between Merced and Bakersfield

People often spin the numbers

One way in which megaprojects are special is the politics. In any big project, people and organisations will compete for resources and jockey for power. Various people have incentives to spin the numbers, making it hard to find the reliable ones. Even without spinning, there’ll be tons of numbers generated at different stages by different parties.

In the world of civic projects, the first budget is really just a down payment. If people knew the real cost from the start, nothing would ever be approved.
— Willie Brown, former mayor of San Francisco, as quoted in How Big Things Get Done

Similarly, there are incentives to engage in “strategic misrepresentation” — deliberately distorting information to win a contract or get a project approved. This may manifest in a superficial plan that glosses over major challenges.

Keep optimism in check

Even when politics aren’t involved, most people still lowball their estimates. This is the well-documented planning fallacy.

When we try to estimate how much work we’ll do on the weekend, we generate a mental image of ourselves working and get a sense of how much work that will cover. We don’t generate a mental image of all the possible interruptions that we may face. Each interruption may be unlikely, but if you have many unlikely events, it’ll add up to a large probability that one of them will arise. [Dan Gilbert gives a similar explanation for why our predictions of our own future happiness are consistently wrong in Stumbling on Happiness.]

Humans are inherently optimistic and overconfident, which is useful in many situations. But in megaprojects, unchecked optimism means we’ll fail to foresee problems. Moreover, if we’re measuring the success of a project against our original targets (on budget, on time, delivering the stated benefits), we should make sure those targets are realistic. Otherwise, our demands will be impossible to meet.

Example: Robert Caro’s book, The Power Broker

When Robert Caro began writing a book about Robert Moses, he expected it to take between 9-12 months. This was a laughable underestimate. Caro had extrapolated from the time it took him to write a long article, approximately the length of a book chapter (3 weeks). He then multiplied it by the number of expected chapters (~12).

But a book, particularly a biography as detailed as The Power Broker, is not the same as writing a series of long articles. If Caro had asked other biographers how long their books their books had taken him, he would’ve gotten a far more realistic anchor for his estimate.

The book ended up taking Caro 7 years. In the meantime, Caro and his wife watched their savings run out and they had to sell their house to keep going. The book ended up being a success—it won the Pulitzer Prize and became a bestseller. But if Caro had known in advance how long the book would take him, he may have been able to mitigate the financial strain on his family.

We can get better forecasts by carrying out reference-class forecasting (RCF). This basically involves:

  • determine the reference class of projects;
  • find the average or “base rate” for that reference class;
  • adjust that figure up or down depending on the unique characteristics of your project. But be careful not to adjust too much, else you’ll let biases creep back in. Only adjust if there are clear, compelling reasons why your project will be well above or below the average.

Because it makes use of relevant real-world experience, RCF beats out other forecasting methods by a large margin. For over half of projects, RCF performed better by at least 30 percentage points on average. It’s now mandatory in both the UK and Denmark.

[You’ve probably heard of this before through Daniel Kahneman and Amos Tversky’s work. While those two developed the theories behind this forecasting method, Flyvbjerg was the one who developed it for practical use in policy and planning. I’ve covered some of this before in my post on Decision Hygiene.]

Planning is essential

Projects broadly have two phases:

  • Planning. All projects start with a vague vision. Planning pushes the vision to the point where it is researched, tested and detailed enough to act as a road map for the way forward. Planning is relatively cheap and safe, as it’s usually just done with computers, paper and physical models.
  • Delivery. Delivery is when the serious money gets spent—when shovels hit the ground. Changes at the delivery stage are much more difficult to carry out and costlier.

Most project leaders don’t get too stressed out by early delays because they’ll have ample opportunity to catch up later. This is dead wrong. Early delays cause chain reactions. Later delays are better because there’s less risk of a chain reaction. [I wonder if the reason most project leaders don’t get worried about early delays is because they don’t have much of a plan anyway. If you’ve planned everything out, it’s clear that an early setback will disrupt all your plans. But if you’re “winging it”, there’s no plan to disrupt anyway.]

Megaproject plans are typically rushed

Megaprojects usually do have a “plan”, but it is often superficial and slapdash. It may not be much more than a broad vision that leaves many things to be figured out later. Examples include the California High-Speed Rail, the Sydney Opera House, and the original plan for the Pentagon building (it was prepared in one weekend and never used).

Example: The Sydney Opera House

An open contest was held to design a national opera house in Sydney. The winner was Jørn Utzon, a Danish architect. He was 38 years old at the time, and hadn’t yet done anything of note. Utzon’s plan was so sparse that one art critic called it a “magnificent doodle”. He had no idea what materials the shells would be made of or how they’d be built. Utzon hadn’t even consulted engineers.

The construction was also rushed. Joe Cahill, the New South Wales premier who had commissioned the project, had cancer and wanted to leave a legacy—and he wanted work to commence before a local election. So construction started with a half-baked plan. There were even points partway through where completed work had to be dynamited and cleared away.

Overall, the Sydney Opera House took 14 years to build, blowing past its original 5-year schedule. The final bill was 1,400% over the estimate—one of the largest building cost overruns in history. And when it was finally opened in 1973, it was acoustically unsuited for opera.

Politics incentivise shallow plans and lowball estimates so that “shovels are in the ground” quicker, so leaders can show tangible evidence of progress. And once shovels are in the ground (i.e. a lot of costs are sunk), people become very reluctant to depart from the original plan, as flawed as it may be. Scholars call this “lock-in” or “escalation of commitment”.

While the phrase “shovels in the ground” usually describes government infrastructure projects, bias for action exists in business, too. This bias is often helpful there — speed matters in business and many decisions are easily reversible. But megaprojects are different in that decisions there are usually incredibly difficult or costly to reverse.

Does planning lead to paralysis?

Some argue that too much planning stops us from doing worthwhile things. Perhaps Robert Caro wouldn’t have started his book if he’d known it would take 7 years.

Albert Hirschman, a renowned economist at Columbia University, argued that people also typically underestimate the projects’ benefits and their own creativity. He called this the “Hiding Hand” theory. Hirschman therefore argued that ignorance and optimism is good — if we know that big projects pose big challenges that can only be overcome by creativity, and we underestimate creativity, we’d never launch a big project.

Since we necessarily underestimate our creativity, it is desirable that we underestimate to a roughly similar extent the difficulties of the tasks we face.
— Albert Hirschman as quoted in How Big Things Get Done

To support his theory, Hirschman only produced 11 case studies (some of which actually turned out to be disasters). The problem with such stories is survivorship bias—we tend to remember and celebrate the projects that overcame obstacles with a burst of creativity to become smash successes.

The data does not support Hirschman’s theory. Flyvbjerg analysed over 2,000 projects comparable to those studied by Hirschman and published a paper showing that, on average, we don’t underestimate project benefits at all—we tend to overestimate them. Only 1 in 5 projects fit Hirschman’s theory.

On average, losses far exceeded gains. And we’re not even counting the full, human costs of failed projects—these failures affect people’s careers and lives. The shambles that was the Sydney Opera House ruined the career of its young architect. What about all the other architectural treasures he could have designed?

Does planning obstruct creativity?

There’s a widespread view that creativity is mysterious and spontaneous. The idea of just throwing yourself into a big project without planning and then hustling your way through challenges is deeply appealing and romantic. There are indeed examples of people just “winging it” and succeeding thanks to their creativity. One is Jimi Hendrix’s Electric Lady studio, which was designed by two inexperienced kids. The movie Jaws is another.

But planning doesn’t obstruct creativity—it can enable it. There is now substantial literature showing that stress has a largely (but not entirely) negative effect on creativity. So be creative and experiment during the planning stage, where costs are low, like Frank Gehry and Pixar. You don’t have to be in the delivery stage to be creative.

What about the “lean start-up” model?

The strong focus on planning might appear to conflict with the “lean start-up” approach common in Silicon Valley, which argues that start-ups should release a “minimum viable product” as quickly as possible. The start-up can then keep developing the product in response to consumer feedback.

However, the authors argue their advice is consistent with this approach. As explained below, the type of “planning” they advocate is quite active, involving testing and iteration just like the lean start-up approach. Moreover, megaprojects can’t be tested in the same way as technology products. You can’t build a minimum viable skyscraper, see how people respond, then knock it down and try again. Megaprojects can’t afford to have the glitches and failures that most start-ups’ products have.

How to plan well

Start broad with the end goal

Instead of carefully considering the project’s purpose or end goal, people often jump straight to the solution. They don’t explore alternatives or try to identify all the potential risks. A plan may look incredibly detailed, but be based on a shaky foundation.

Good planning starts off broad with open-ended and exploratory questions. The most basic of which is: “why?”. Why are you doing this project? Are there alternative ways to get to the same goal?

Projects are not goals in themselves, but are merely ways of achieving goals. Project planning flowcharts typically lay out from left to right what needs to be done. The goal is in the final box on the right. Planning should start there—figuring out what’s in the box on the right.

Example: The Guggenheim Museum Bilbao

The local government in Bilbao, Spain asked Frank Gehry, a renowned architect, to renovate a Guggenheim Museum for it.

Instead of simply agreeing or refusing to do the work, Gehry asked the officials why they wanted to do the project in the first place. After he learned that it was to draw visitors to Bilbao and boost the economy, Gehry convinced them to build a new, dazzling museum on a riverfront site instead of a renovating an existing one.

Photo of the Guggenheim Bilbao in Spain
By Naotake Murayama from San Francisco, CA, USA – Museo Guggenheim, Bilbao, CC BY 2.0.

The resulting building was an overnight success. In its first three years, it drew almost 4 million people to visit the city. It even came in under budget.

Obviously, questions should come before answers. But projects commonly start with answers, because thinking from right to left doesn’t feel natural. It’s more natural to just focus on what’s in front of you.

Good planning is detailed and slow

People usually think that to get a project done ASAP, you should start immediately, set tight timelines, and make everyone work at a furious pace. But this is wrong. Getting a big project done quickly requires a good plan, and good plans take time.

Think slow, act fast: That’s the secret of success.
— Flyvbjerg and Gardner in How Big Things Get Done

Planning requires thinking, and careful thinking—cultivating ideas and coming up with solutions—is inherently slow. Good planning ensures that every detail is scrutinised and tested. Bad planning by contrast leaves problems, challenges, and unknowns to be figured out later.

Example: The Empire State building

Before construction even began, the architects knew exactly how many beams, rivets and bolts would be needed. The Empire State building was finished entirely on paper first. The construction period took merely 18 months and was completed exactly on schedule, on 1 May 1931. Moreover, it came in $9m under its $50m budget.

Detailed planning also helps correct a cognitive bias called the “illusion of explanatory depth”. Most people feel they understand complex phenomena far better than they actually do. The illusion quickly vanishes when you ask them to try and explain what they think they understand.

Planning is active and iterative

The word “planning” comes with a lot of baggage. Organisations can waste tons of time on meetings with meandering discussions that never go anywhere, reports, charts and schedules that are useless. For many, “planning” sounds like a passive, bureaucratic activity: sitting, thinking, abstracting what you’re going to do. And much planning does fit that bill.

However, the type of planning the authors advocate involves experimentation. You simulate the project, test and make changes to it, then refine it. Problems are inevitable—an iterative process increases the chance those problems will be flushed out in planning. After many iterations, the simulation becomes a rigorous and reliable plan.

Example: Pixar planning

In the movie industry, the planning phase is called “development” while the delivery phase is called “production”. At Pixar, movies spend years in development. After a decent script is written up, every scene is planned out with detailed storyboards. Each storyboard covers roughly 2 seconds of film time, so a 90-minute movie requires around 2,700 drawings. The movie is then performed multiple times to other Pixar employees, with the development team voicing the characters and simple sound effects added.

Each “screening” produces valuable feedback, which is then incorporated into the next version. In the early versions of the movie Inside Out, for example, there were more emotions, such as Schadenfreude and Ennui, and each emotion had a normal, human name. The test audience found this far too confusing, so they simplified it. This whole process is usually repeated around 8 times so that, by the time the movie enters production, things go relatively smoothly.

All this planning is an insane amount of work, and it’s not exactly cheap in absolute terms. But it’s cheap in relative terms. In the production phase, where highly skilled people produce animations and movie stars do voices, costs easily explode.

We’re good at learning by tinkering—which is fortunate, because we’re terrible at getting things right the first time.
— Flyvbjerg and Gardner in How Big Things Get Done

Mitigate risks

All sorts of black swans may disrupt a project, such as a pandemic, death or change of government. It doesn’t even need to be a dramatic event. In complex systems, minor changes can frequently combine in ways to disrupt the project.

… when delivery takes decades, the unpredictable becomes inevitable.
— Flyvbjerg and Gardner in How Big Things Get Done

To mitigate risk, you don’t need to predict the exact circumstances that lead to disaster. Risk mitigation is a matter of probability, not certainty. We can study the types of black swans that tend to come up in particular projects and look for ways to reduce the risk.

An effective way to reduce risk is to look for ways to shorten a project, particularly the delivery stage. A project’s duration is like an open window. The longer the window is open, the more opportunity there is for a black swan to crash through.

For example:

  • In 2003, Madrid greatly expanded its subway network by 131 km of rail and 76 stations. To make the project go faster, they got up to 6 boring machines working at once, instead of using just one. Each machine costs between $20m to $40m, which is cheap given the time and other costs it would save. They managed to deliver in around half the time and half the cost compared to the industry average.
  • Archaeology often disrupts high-speed rail projects because the law requires that if the digging uncovers a historical relic, the work must stop a qualified archaeologist surveys the site and ensures that nothing significant is lost. So experienced managers make sure they have an archaeologist ready.
  • Essential machinery may break down. Instead of then ordering parts from a manufacturer and pausing the project until they arrive, you can keep a wide range of spare parts on hand.

We’ll never achieve certainty, but we can move the probabilities in our favour.

Experience is underrated

Experience is invaluable in big projects. Experienced project leaders have a lot of tacit knowledge—knowledge that they cannot put into words. Not only can they draw on their experience to develop better plans, they may also have more experience navigating the messy politics of most big projects.

Though the value of experience may seem obvious, it’s often overlooked or misunderstood. Politics tends to prioritise other things. The Olympic Games are a prime example. The Games aggressively marginalise experience, because they’re run in a different city, by different people, every time.

The Olympics are forever planned and delivered by beginners—a crippling deficiency I call “Eternal Beginner Syndrome.”
— Flyvbjerg and Gardner in How Big Things Get Done

Every Olympics since 1960 (both summer and winter) has gone over budget, with an average overrun of 157%. The highest overrun was the Montreal 1976 Olympics, which soared to 720% over budget.

Domestic bias

In public infrastructure projects, governments like to award contracts to domestic companies so as to make influential friends and “create jobs” for locals, even though the domestic companies will rarely be as experienced as foreign competitors. This seems to be partly why the Californian High-Speed Rail became such a mess.

Example: Canadian icebreakers

The Canadian government decided it wanted to buy two icebreakers (large ships designed to move through ice and clear the way for other vessels). Instead of buying them from experienced manufacturers in other countries, it decided to award the contracts to Canadian companies.

What’s worse was it decided to give the contract for one icebreaker to one company in Quebec, and the other contract to a second company in British Columbia. So the Canadian companies couldn’t even learn from their experience building the first ship. The result was that the cost blew out from the original CA$2.6 billion estimate to CA$7.25 billion.

Uniqueness bias

When leaders try to build the most unique project—the first, or biggest, or tallest, etc—experience again gets sidelined. Words like “bespoke”, “unique” or “original” carry positive connotations, but we should really be wary of them because anything new or unique is inexperienced.

Example: Seattle’s biggest tunnel

Seattle decided that it wanted to make its Route 99 tunnel the biggest one in the world. Since that would require the world’s biggest boring machine, Seattle placed a custom order.

The machine cost $80 million—more than double the cost of a standard borer. After boring around 1/9th of the tunnel, the borer broke down. Getting it out of the tunnel, fixing it, and putting it back in took 2 years and another $143 million.

If the city had instead just drilled two standard-sized tunnels, it could have used off-the-shelf equipment that was widely used and more reliable.

Experience isn’t something only people have—technology can be experienced, too, if it’s tried and tested. But we often assume newer technology is better.

Experimentation and Experience are linked

Test everything. The less proven something is, the more it must be tested.

Experimentation is how we gain experience. So try to maximise experimentation with an iterative planning process. At that stage, failure is relatively safe, meaning you can take more risks and try new ideas.

Megaprojects can go really wrong

Most megaproject types have “fat tails” — i.e. cost overruns do not follow a normal distribution but a “fat tailed” one with higher chances of extreme outcomes. For example:

  • NASA’s James Webb Space Telescope was 450% over budget;
  • Canada’s firearms registry (an IT project) was 590% over budget; and
  • Scotland’s Parliament building was an incredible 978% over budget.

Fat-tailed distributions are typical within complex systems. Even smaller projects like home renovations can have fat tails. And because ordinary people have limited funds, a fat-tailed overrun can easily wipe them out.

Reference-class forecasting with fat tails

Ideally, you’d want to know if your project is part of a fat-tailed distribution or not. If you’re unsure, err on the side of caution and assume it is (it usually will be).

While it’s fine to use the average (mean) cost in RCF for a normal distribution, the average is not as reliable for fat-tailed distributions where things can go really wrong. So instead of trying to forecast a single outcome (“The project will cost X”), you should try to forecast risk instead (“The project is X percent likely to cost more than Y”).

In a typical fat-tailed distribution in project management, you get about 20% chance of ending up in the tail. That is too much risk for most organisations. But you can mitigate that risk by following the advice in the rest of this book.

Modular projects are less susceptible to fat tails

Modularity is valuable for pretty much all projects, but it’s essential for megaprojects. Modular projects deliver faster, cheaper, and better than non-modular ones.

What is modularity?

Modularity just means making a big thing from small things repeated many times. Wedding cakes are mostly made up of several smaller, identical cakes. Skyscrapers can have many repeated floors. A wind farm is just a bunch of repeated windmills.

Example: 20,000 schools and classrooms in Nepal

In the early 1990s, Denmark and some other governments agreed to build around 20,000 schools and classrooms in Nepal, in the poorest and most remote regions. Flyvbjerg was the planner for the project.

Not only did the project finished on budget, it finished 8 years ahead of schedule. Independent evaluators thought the schools worked well. One reason for this project’s success was its modularity. Classrooms are small and relatively easy to build. Once they’d built a few and gotten some schools going, they could look at what worked (or didn’t) and make changes.

The opposite of modularity is to build one huge thing. One huge hydroelectric dam, or nuclear power plant, or IT project. Building one huge thing is more likely to be slow and complex, particularly if it’s highly bespoke and doesn’t use standard parts or machines.

How modularity reduces risk

In an Appendix, the authors set out a table showing the average cost overruns for 25 project types, based on their database of over 16,000 projects:

  • There are 5 project types—solar power, wind power, fossil thermal power, electricity transmission and roading projects—which are not fat-tailed. They rarely go disastrously wrong because they’re all modular to some degree.
  • Nuclear power and IT projects, on the other hand, are very fat-tailed. Around 18% of IT projects have cost overruns exceeding 50%, with the average cost overrun being 447%!

Modularity reduces risk by allowing for:

  • Experimentation. You can test out an idea on something small and see if it works. You can’t do this with the “one huge thing” approach. If you don’t get it right the first time, you’ll waste a lot of money.
  • Learning. Even if you don’t change what you’re doing, you’ll naturally learn and get better by practising the same thing over and over. Thanks to modularity, Denmark managed to reduce the cost of wind power by 60% within just 4 years. When you’re building “one huge thing”, however, your experience may have limited transferability. For example, the few nuclear power plants that have been built tend to be designed for a specific site. So it’s hard to get the experience to become an expert in building them.
  • Partial success. A wind farm that is 90% complete can start generating returns, whereas “one huge thing” projects are all-or-nothing by nature. A nuclear power plant that is 90% complete is still useless. This leaves a larger window for a black swan to crash through.
Example: China’s modular hospital

When the Covid pandemic first hit China, a company that makes modular housing modified an existing room design and then cranked built them at speed in a factory. Manufacturing in a factory and assembling on-site is far faster than traditional construction because a factory is a controlled environment designed for efficiency. So it won’t be disrupted by things like bad weather.

Within a mere 9 days, China opened a 1000-bed hospital with 1400 staff in Wuhan.

Modularity is a matter of degree

A skyscraper’s floors may not be made out of exactly identical parts but its floors can be designed to be as similar as possible. When William Lamb designed the Empire State Building, he minimised variety so that construction crews could learn during the process.

Even with traditional “one huge thing” projects, there’s been a shift to modularity:

  • Nuclear. People are now thinking of building small modular reactors (SMRs). However, at the time of writing, this is still an unproven technology.
  • Hydroelectric. Instead of building one giant dam, you could divert some of the river’s flow, run it through small turbines to generate electricity, and return it to the river. This is “small-scale hydroelectric”, and involves less environmental damage, cost, or risk.

Why Flyvbjerg wrote this book

At the end of the book, Flyvbjerg explains that we need to get better at planning and delivering big projects if we want to have any hope of tackling climate change.

To get to the “net zero by 2050” target that most countries have committed to, we’d need to roll out massive projects at scale—without the high failure rate.

My Review of How Big Things Get Done

I really enjoyed How Big Things Get Done. It’s clear Flyvbjerg is an expert in this area — he’s spent decades building up the biggest megaproject database in the world, and governments around the world have approached him for help with their megaprojects. He also did a great job addressing the obvious counterarguments to his points (e.g. the Silicon Valley mentality and Hirschman’s “Hiding Hand” theory).

Books by academics tend to give you excellent ‘bang for your buck’ — I’ve never read a book by an academic in their field that “could’ve been a blog post”. That said, one thing I don’t like about academics’ books is that they can be difficult to read. Not necessarily because of the material itself but because of bad writing — unwieldy sentences, unnecessary jargon, or excessive detail. How Big Things Get Done does not suffer from this. It’s very easy to read and the examples are accessible and engaging. Perhaps more academics need to partner up with people like Gardner, who are experienced in writing for lay audiences.

It would be easy to read this book and think that a key takeaway is just “plan more”. But that would be wrong. The truth is that most of us don’t work with high-stakes megaprojects, and the value of planning does not apply to the same extent when we deal in lower-stakes. The authors acknowledge this themselves when they talk about the Silicon Valley “fail often, fail fast” mentality. That said, I do think action bias exists and is underrated, even in our everyday lives. So I think the better takeaway is to consider more carefully the trade-offs between planning vs doing.

Let me know what you think of my summary of How Big Things Get Done in the comments below!

Buy How Big Things Get Done at: Amazon | Kobo <– These are affiliate links, which means I may earn a small commission if you make a purchase through these links. I’d be grateful if you considered supporting the site in this way! 🙂

If you liked this summary of How Big Things Get Done, you may also enjoy:

4 thoughts on “Book Summary: How Big Things Get Done by Bent Flyvbjerg and Dan Gardner

  1. I like examples like the Bilbao Guggenheim museum, where Gehry was able to better describe what they were after. And I hadn’t realised how planned the Empire State Building had been.

    I think my issue is I find gantt charts and detailed planning really boring…

    Minor typos: in a couple of points I think you’re confusing Caro and Moses.

    1. I think a lot of people find planning boring, which is part of why there’s an action bias. Planning comes with little or no immediate feedback, so we don’t feel the same “reward” we feel when we start doing stuff. For this reason, it can also be hard for third parties like managers, politicians, or customers to distinguish between bad, superficial planning (which may involve lots of useless charts and reports) from truly good planning. The kind of active, experimental planning that Flyvbjerg advocates does come with some feedback (e.g. Pixar’s early screenings) but not all projects can be tested in this way.

      And thanks – I’ve fixed the Caro/Moses mix-ups now.

    1. Hey, thanks for the comment 🙂 I checked out your summary too, and it looks really good! I particularly liked your use of pictures. Funny how two people doing the same book can end up with very different summaries.

      Good luck with your Substack!

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.