How Big Things (Should) Get Done - Ep128: Prof. Bent Flyvbjerg

Transcript

Michael Liebreich

So Professor Flyvbjerg, thank you so much for joining us here on Cleaning Up.

Professor Bent Flyvbjerg

Thank you for having me, I'm really pleased to be here.

Could we start perhaps by you describing what you do, what you research, in your own words? Because I could make something up, but I'll probably get it wrong. So, what is your thumbnail sketch of yourself?

So, I'm a professor first and foremost, at Oxford University and at the IT University in Copenhagen. And what I research is big projects. So, really big, often multi-billion pound and multi-billion dollar ventures, capital investments, in all sorts of areas, from IT, to transport infrastructure, to mining, to defense projects, to energy projects, all sorts of energy projects. So, anything that's big and complex, that's my remit, and that's what I do. On top of the research and teaching that I do as a professor, I do a lot of consulting and executive education, training leaders; there's a real need for people to get better at this because we have difficulty with big projects. So, there's a real need and real demand out there for top leaders - who are responsible for these projects - to get the most recent knowledge and skills on how to do better. That's what I do.

And along those lines, my audience that watches on YouTube, which is about 20% - so, most people will be listening on the podcast - but the audience on YouTube might be able to see over my shoulder there, that I've got your book, the latest of I believe ten books, but it's called How Big Things Get Done. Ah, and you're holding up... For those on the podcast, you can't see, but Bent is holding up a copy of his book as well. So, that is part of your process of presumably educating executives and people like our audience on the challenges of these big projects, right?

Yes, it is. And like you said, I've written several books before, but this time, I wanted to do something I've never done before, something special, which would be two things: to take stock of where I am now, how much I know about these things and put it all in one book. So, instead of having some specialized book on just an aspect of the knowledge, I wanted to put my accumulated knowledge over my life up to this date into that book. And I wanted to do it - this is the second thing - I wanted to do it in a way where everybody could read it; anybody you know, anybody who would be in an airport bookstore, picking up a book, would pick up this book and look at it and say, whoa, that's actually relevant to what I'm doing. Because we all do projects, whether big or small, and we cover both big and small projects in the book. And by we, I mean, my co-author, Dan Gardner and myself. But anyways, that's what I wanted to do with the book: present my accumulated knowledge in a way where anybody could understand it, anybody who's doing projects.

And I've read the book, I read bits of it early, and then in preparation for this conversation, and I can vouch for the fact that it's extremely readable; there's lots of anecdotes, lots of stories. But also, a structure and a thesis. Can you start us off by saying, how big is the problem of projects, big things, not getting done properly? Can you put some metrics on... What would the world look like if we actually could do this stuff properly - if we all read the book and all followed its lessons? What would we save, or what would happen?

Well, to put numbers on the problem, we do that in the book, and we actually have a name for it, we call it the Iron Law of project management. And the Iron Law basically says, over budget, over time under-benefits over and over again. And that's a law documented by hard statistics, and we see that only around 9% of projects are on budget and on schedule. So, 9%, meaning that over 90% are not on budget and not on schedule. And now, if you add on benefits, then it's half a percent; that's half of 1% are budget, on time, and deliver the promised benefits. And that's a dismal statistic. This is based on the biggest database of its kind, with 16,000+ projects in it. At first, when he saw these numbers, we had a rather smaller database - just a few hundred projects, 258 to be precise - which was the world's largest database at the time. We had the same result, basically, but we thought it could be an artifact of the small sample. So, we started sampling more. So, we collected data for a few hundred more projects - still the same result. Then we collected data for a thousand more projects, still the same result. And now we have 16,000+, still the same result. So, it's a very robust number. And now, in the meantime, other colleagues have started to do research on the same thing as we do, so you don't have to trust us only - you can also look to our colleagues and they find similar numbers. So, we consider it a pretty firm result, a pretty firm finding, that projects are actually performing dismally; not only badly, but dismally. And on top of that, we have this thing about fat tails, you know, black swans; it's not only that they perform badly, but at large percentage of the majority of project types perform extremely badly, with huge blowouts; not just somewhat over budget, but a total blowout of hundreds and hundreds of percent over budget, and the same with their schedule and benefits. So, the situation is grim, that's the only way to describe it. Luckily, a lot of people are now interested in improving the situation.

So, let me give you two personal touch points. As I was reading the book, there were two things that really appealed to me. You talked about, actually, both of them in that comment. One is the reliance on statistics. Now, you may not know how New Energy Finance, the company that I started that I then sold to Mike Bloomberg, that is now Bloomberg New Energy Finance, the leading information provider globally on the net-zero transition... I'm just a senior author, I just write some pieces now. But that started because I wanted data: who's investing in wind, who's investing in solar, who's investing in fuel cells, all those sorts of things. And there was no good data, you couldn't do a trend analysis, you couldn't really grasp the problem without that data; how much money were we investing...? So I started by building databases, so that resonates with me... What you must have done is starting there, and opening up the spreadsheet and start putting structuring the data. The second touch point for me was I was on the board of Transport for London. And you may have... I'm sure you're familiar with what happened to Crossrail. So, I was on the board of Transport for London, I was on the Finance Committee, when the extent to which it was off-track was concealed from us. I was also on the board during our famous Garden Bridge adventure, when then-Mayor Boris wanted to build this garden bridge and the costs exploded and exploded, and almost looked like a chapter out of your book. So, it was very personal for me reading your book and thank you very much for both of those: building a dataset, and awakening these raw memories for me. Let's come back to that statistic though. It's one in 200 projects that is on time, on budget and delivers the projected benefits. One in 200, half a percent. Why is that?

Well, that's one of the things that I've explored thoroughly in my research and that we describe in the book. And there's two levels that you can look at it, what we call causes and root causes. When people talk about causes, they typically will talk about, you know... it's just difficult to do projects, and they're complex, and in many projects, you have to do construction. And as soon as you start doing construction, you find things in the ground that you didn't expect, or the weather is bad, or whatever, you know. So, these are what we call causes. We find that underneath those causes, there are what we call root causes; it goes much deeper than these surface phenomena. And we say the root causes are two, and they're related to psychology. So, our human psychology is one root cause, the way our brains work. The second is power. And that's about how we jockey for position, and for funds and so on. Also, as human beings, it's more a social aspect than a psychological aspect. And if we take psychology first, it is very well-documented in behavioral science, behavioral economics - by giants like Daniel Kahneman and Amos Tversky and Richard Thaler at Chicago, and so on - that the human brain is biased, we have cognitive biases. The most prominent is probably optimism. So, we are all optimists, and you can see what happens with a budget if it's optimistic. Well, an optimistic budget is low, and then you will have a cost overrun, just as surely as the law of gravity works. I mean, if you're optimistic about a budget, it's going to be low, and you're going to have an overrun on that budget, on average. So, the same with a schedule: if you're optimistic about the schedule, you'll have a schedule overrun; if you're optimistic about the benefits, you won't have as many benefits as you thought you would have. So, that's just optimism, and cognitive sciences have documented over 200 biases to date, you know, and it's still increasing; new biases are being discovered all the time. So, other biases are the planning fallacy, overconfidence, bias and the ease of retrieval, and many others. And this means that the decisions that we make are actually not that good, the citizens are biased. And the solution to that, if you want to get beyond that, you need to de-bias, and that we also describe in the book. We've actually developed methods for de-biasing, including a method called reference class forecasting. But it actually doesn't matter much which method it is; the important thing is that you take de-biasing seriously, and you realize... So this is the first thing I teach decision-makers is: you're biased. Because you're biased, your biggest risk is you. The biggest risks are not out there in the world - even though they're also out there - but that's not the biggest risk, the biggest risk is the way we perceive the risks. Once we have processed the risk through our brain, they come out optimistic, they come out misunderstanding what the base rates are, the odds that we're playing, and so on. And in order to avoid that we need to de-bias things. So, that's the psychology side, and that's the root cause there. On the power side, it's similar, but it's deliberate. So, the psychology is not deliberate, this is something we do unconsciously; we actually don't think about it, mostly, when we optimistic, we are just optimistic. And we have all have had this experience: oh shit, I did it again, you know, I was optimistic again, I thought I would be at my destination at this hour, but look, now I'm 15 minutes late, again. So, this is something that's familiar to all of us in our behavior. The power thing is deliberate, it's very different. The result is the same, but it's deliberate. So, if you and I wanted to make our project look good on paper, that's easy, and in some situations, you want to do that because you're in competition with other projects for limited funding. So, how do you deal with that? You make your project attractive on paper. How do you make your project look good for the beauty contest, that is project selection? Well, you underestimate the cost so your project looks cheaper. Who doesn't like a cheap project? You underestimate the schedule so it looks like you can do the project faster. Who doesn't want their project fast? And you make the project look good by overestimating the benefits: this product is going to deliver a lot more nice and good things than will actually happen - and we know that - but we put it in the in the proposal. So, now we have a fantastic proposal for the beauty contest, and we have a higher likelihood of winning the beauty contest, that it's our project that gets funding, and we did it deliberately. And just as sure, again, as the law of gravity works, if we did all these things, we're going to have cost overruns, schedule overruns and benefit shortfalls, right? And this is exactly the same as on the psychology side. So, these two root causes reinforce each other; they create a double whammy of cost overruns, schedule overruns and benefits shortfalls with these two drivers.

And if anybody who is listening or watching this hasn't figured it out yet... the reason that this is all about the climate is because you found in your book, you looked at some of the worst projects - in terms of being on time on budget and delivering benefits - are nuclear, and some of the best ones, the best category, is actually solar. So, it's very relevant for the climate change challenge, is it not?

Absolutely, and I would say climate projects are probably the most interesting type of projects, of the 25 different types of projects that we look at. It's super interesting, and strange in the way that they completely polarize; at one end, you have the worst performing projects, which will be nuclear, like you said, but also hydroelectric power, so big dams, is the other energy project. And that's kind sad, you know, and a problem, because both nuclear and hydroelectric power, are carbon-free, right, more or less? So it would be really great if they performed well, and could be done fast, but instead, they have huge cost overruns, and they have huge delays, they take a long time to do, and they become very, very expensive. But all is not lost, because at the very opposite end of the scale, we have solar, as the best performing product type of not only energy projects - of any type of projects that we've ever looked at, solar is the best performing, meaning the smallest cost overruns; 1%, it's nothing, it's basically on budget on average, and being built very quickly on time, on budget. And cheaper and cheaper, you know, so it's getting cheaper and cheaper. The same for wind power; not quite as good but very, very good compared to other project types, wind power. Batteries, same thing. And the explanation is - which is something that we've looked into in depth, not only for energy projects, but for all projects - the explanation is that solar, wind, batteries are born modular, they're completely modular. So it's obvious for solar panels, right? The solar cell is the basic module, or the Lego as we call it. Basic module, you put enough solar cells on the panel, well, then you have a solar panel. You put enough panels together, you have an array. You put enough arrays out there, you have a solar farm. That's it, you know, it's just one module builds the next, which builds the next, which builds the next. And if you don't have enough in one farm, you just build more farms. So, this is what we call scale-free scalability. You can scale this at any scale, you can scale it up, you can scale it down; it's really easy, it's really cheap, and it's really fast. Wind, same thing. A wind turbine consists of four Legos: the first Lego is the foundation, the second Lego is the tower, the third Lego is the nacelle, with the turbine, and the third Lego is the wings, or the blades, that you put on the nacelle. Click, click, click, and you have a turbine. And literally, it used to take months to construct, on a construction site, wind turbines; now it's done in one day, because they are not constructed anymore, they are manufactured, in factories, and then you just bring them on to the assembly side and you assemble them. So, it's become incredibly efficient, and that's actually our luck in the situation that we're in now, same with batteries. We are so lucky that there are energy technologies that we can really scale up at enormous speeds and to enormous scale, which is exactly what we need now, in order to meet the climate goals for 2030 and 2050. We don't have time to mess around anymore; we don't have time to mess around with nuclear. I wish... I'm not anti-nuclear by any means. Some people think that sometimes, because I point out the numbers look bad for nuclear, then people say he must be against nuclear. Now if that isn't a dumb argument, I never saw one you know? It's about the numbers. I'm an economic geographer, I look at the economics, and the economics for nuclear are dismal. It's really bad, and I wish they were better, and I hope that small modular reactors might solve that eventually. But I wouldn't be holding my breath, because we don't even have a prototype yet. But the situation for energy in general is this interesting polarization, of really badly performing projects and really well performing projects.

Let's do this because... Rather than make it all about nuclear - and of course, there's going to be a subtext, because these nuclear power stations, and waste storage, they are big things, there's a clue in the book - How Big Things Get Done. Although, as you point out big solar, big wind, big batteries, giga-factories, and so on are also big things. But I don't want to make it just about nuclear. I actually want to try and extract.... because in your book, you've got a number of heuristics, and I think we would do well to spend a bit of time on some of the other heuristics, because they are relevant. And then maybe we can come and look at a range of different net-zero technologies, and see whether they are likely to be at the bottom end of the scale, or at the top end, the high performing or the low performing. You've got one heuristic, which you started the book with, which you call "Think Slow, Act Fast." What do you mean by that?

Yeah, so "Think Slow" means deliberating. So, you need to sit back and deliberate about what you're doing. So, one of our cognitive biases is availability bias. Our brains have a tendency to just grab the first idea that comes up and run with it. And there's probably good evolutionary reasons for that, that we needed to be able to act fast when somebody popped up in our brain. And so that's a pattern we have, and again, it's very well-documented by behavioral science. And this is the characteristic of thinking fast, you just quickly grab an idea. And this is actually why Daniel Kahneman's book is called Thinking, Fast and Slow, that's his famous bestseller where he summarizes his work, called Thinking, Fast and Slow. So thinking fast, when we talk about, is a nod to Daniel Kahneman; yes, this is the way to do it, he is right. This is his argument in his book, that we need to think, slow and not fast, because when we think fast, we make lots of errors. It's good... I mean, when you meet a lion on the savanna, and the lion is trying to eat you, it's very good to think fast and run fast, you don't want to sit down and deliberate: what do I do in this situation? So, that's why we have these things from an evolutionary point of view. But I'm sure that anyone can see that it's not all decisions that you can use that decision-making process on, and that's what Kahneman argues. So, we agree with him: thinking fast is not a good idea for billion-dollar projects, or any important project that you do in your life - it might be your wedding, or buying a new home, or the kitchen renovation that we talk about in the book. All of these things, you actually need to think slow - which doesn't mean taking a long time. It just means taking sufficient time, adequate time, to think through what it is that you're doing. And because, if you do that, then you're able to act fast. And that's the second thing that Kahneman doesn't talk about, but that we talk about, because we are about doing projects, so acting is absolutely crucial to what we do. And we are saying the preparation for acting fast is thinking slow, and only if you thought slowly, will you be able to act fast, and acting fast is extremely important to reduce your risks. The longer you take to do something, the more you expose yourself to risk, including black swan risks. So, we look at that exposure as a window, a window in time, and by acting fast, you make that window smaller, you shorten the time it takes to deliver your project. And by making that window small, fewer bad things can fly through the window, including big fat black swans; they can't get through the window if you make it small, and therefore you have a bigger chance of being successful with your project. So, that's what we called the rhythm of a successful project, is: think slow, act fast. And only if you do that, will you be successful.

And you give a great example, there's a few examples in the book, the Frank Gehry building in Bilbao. And I said I don't want to make this about nuclear, but nuclear is the opposite of that, right? Nuclear, where the window is so hard to close because it just takes long to build, and so all sorts of things can change: politicians, political regimes, earthquakes, all sorts of things can change. Is that why you think fundamentally that nuclear won't be able to address its disadvantage in this game?

So, I wouldn't go so far as to say that it won't be able to address these issues forever. It might be in the future that it will be able to address these issues, but I would go so far as to say that so far, nuclear hasn't been able to address these issues in a way where they have been addressed across a very large number of projects, and that you can just count on, we have a formula for how to do nuclear, we can do one successful project after the other. The historical evidence shows the exact opposite, including right now, with the four nuclear reactors that have been built in the US, of which two have been given up, and the two that are being built in Europe right now that are going way over budget and way over schedule. So, it's not going well for nuclear, but that doesn't mean that it has to go bad forever. But it is a problem, and the key is that it just takes too long, it's too difficult, and you have to get it right the first time. And it's also a problem that the major part is happening on a construction site. So, nuclear is not built in factories, mainly, it's built on sites, and that's a problem. It's much more difficult to control something on a construction site than it is to control something in a factory. Factories are... that's what they're for, basically. Factories are for controlling things, making things predictable, making every step in a process predictable. So you can do it over and over again, and you get better every time you do it, and because you get better, it gets cheaper and it gets faster. For nuclear, the exact opposite is happening: you do it, and then next time you do it, it takes longer, because you found out that there are more things that you need to take into consideration. And because it takes longer, and because there are more things you need to take into consideration, it also gets more expensive. And that's what happens with nuclear, and that's what we call negative learning curves. If things get slower, and more expensive every time you do them, you have negative learning curves. You want to stay away from that like the [plague], you know - it's what you don't want in the projects that you do. You want positive learning curves, which are the opposite of course, which means that things get faster and cheaper every time you do them, because you learn more and more. That's exactly what you want to do, and that's what's going to make you productive and successful in the long run.

You mentioned just earlier in this conversation, small modular reactors: we had Tom Samson, Episode 116, who talked about Rolls Royce's modular reactor. It's not that small, it's a half a gigawatt, or 470 megawatts. But it's all made in a factory and brought to the site and only assembled on the site, and that's something you talk about in your book. I'm still on the fence as to whether that will... Because you've got a disadvantage of scale for making nuclear reactors small. I mean, the reason we made them big was you go to all that effort, you might as well make it huge, and you get the economies of scale. So the small ones start with a disadvantage, but might catch up. And I've got to be honest, I'm on the fence about how cheap it could it could become.

Yeah, and you're not alone. And I don't mean myself, I mean other people are skeptical about the physics of this; the physical laws actually dictate that the nuclear reactor should be bigger, one gigawatt and above. And, therefore, per definition, according to these people, small modular reactors will be inefficient, because they don't realize the efficiencies of scale, and that is an issue. I still think that it's worth experimenting with; this is a thing we talk a lot about in the book is experimentation. So, we don't think of planning like a bureaucratic exercise where you just produce charts and paper; we think of planning as experimentation: you actually do trial runs, you build mock-ups, you build experiments, and you try out things. And that's part of the learning that you will do during the slow thinking. And the successful project leaders that we study, they actually do it that way. And we encourage that for nuclear too, you know. There's still stuff to learn about nuclear. I mean, it would be a pretty amazing thing if we could make nuclear work safely and cheaply, that would really be something you know. So, I think the prize is so big that we should keep trying, and we certainly shouldn't close well-functioning nuclear power plants that are already up and running that some people are doing. That's just... that does not make much sense in the situation that we're in right now with the climate. So that's our position on this, and we would just say the verdict is still out, there's no clear answer, not even to the Rolls Royce modular reactor. You say that they are building it in a factory: wait a minute. They're developing a prototype now, hopefully, and we haven't seen it yet. So, they're actually not building anything in a factory yet, you know. But they will be, hopefully, in the future.

That's right. The present tense - they are building - was doing a bit too much work there. It is very interesting, because if you look at, for instance, what's happening in space - which also you talk about in the book - that you've gone from these huge NASA projects that are so complex that they kind of collapse under their own weight economically, and then you've got these entrepreneurs, people like Elon Musk, Jeff Bezos and others, who can just operate at a completely different speed in terms of experimentation and learning. And then... we sort of don't mind if they get one or two things wrong, if some of Elon Musk's rockets spontaneously disassemble, I believe it's called, we don't mind. We kind of would mind if a nuclear power station spontaneously disassembled...

See, that's the thing about nuclear: the demands are so much higher regarding safety and environment. So, the bar is so much higher than for any other type of project, that people actually have problems jumping over that bar. That's as simple as it is. It's just too difficult, with the knowledge that we have today.

Let's do a few more of your heuristics. You've got one called "Take the Outside View," that's the title of the heuristic. What does that refer to?

Yeah, it refers to... It's the opposite to the inside view, obviously, and the inside view is something, again, that it's been documented in behavioural science, that we tend to take an inside view on our actions, on our decisions, and on our projects. So, we understand things inside out, that's our spontaneous way of understanding something. So, if we're doing a project, we will think about: what are the constituent parts of this project, what do I need to produce or buy in order to do the project, what will each item cost? And you multiply the number of items with the unit cost, and you eventually get a budget, right? That's a typical inside-out way of budgeting. The outside view does the opposite: it says don't start on the inside, start on the outside. Simply look at similar projects to the one that you are now planning, that have already been completed, and look at what they cost. Then you will have a much more reliable estimate of your costs. And of course, that hasn't been biased by your thinking, by you trying to imagine what the project is going to look like, and what elements are going into it, and what these elements are going to cost. No, instead, you are empirical, and you simply ask: what did these projects cost in the past? So, that's a very simplified description of the outside view versus the inside view. And again, it's clearly documented that the outside view is producing much more accurate estimates of cost and schedule and benefits. So, that's why we recommend taking the outside view: it simply makes for a more accurate view of what it is that you're doing. You'll get a more accurate business case for your decision, or your project, if you do it that way.

So, an example you use in the book is a couple renovating their kitchen, and then it becomes their house, and so on. This was very raw, because I'm just about to do a renovation of the kitchen in the house. And so when the builder turns up with that plan, that Gantt chart - you know the one that unrolls, and it's got all the little things on it - I should say put it away: how long did the last one take you, how long did the one before take you? And just go with that?

Yeah, but the builder probably won't tell you that, so you have to be careful who you ask. You should ask, actually, your family and friends. Just find out who has done... If you were doing a house renovation, who has done a house renovation over the past 5, 10 years amongst your friends and family and colleagues? And then you just ask them, how much did that go over budget? And some of them will say zero, it'll be very rare. Some will say 30%, and some will say 100% - that's actually the average, 100%. Some will say 200% or 400%, like the poor couple in our book with their house renovation that started as a kitchen renovation. And then you will simply take the estimate that the builder gives you, and you would multiply it by this cost overrun that you obtained from your friends and family, and now you have a reliable estimate. The builders estimate will not be reliable, and it's simple why: the builder cannot imagine all the unknown unknowns that are in your building; there might be mould, hidden mould; there might be electrical installations that are not up to the current standards, but as soon as you start ripping them out, you have to put in new ones; as soon as you start ripping walls and floors out, you actually by law have to put in electricity installation that will live up to the current standards, and so on and so forth. And the builder doesn't know any of this because they haven't seen what's behind your walls yet, so they can imagine it. But they won't tell you that, and therefore you need this other way of going about it: you combine the builders estimate with your empirical cost overruns from your friends and family.

So, I'm smiling because this is also relevant to something that I've done a lot of work on, which is hydrogen heating in homes, where we have people say, well, we did it in the 1970s; we went from town gas, which was 50% hydrogen, to natural gas, which is, of course, mainly methane, and it cost this; so, we adjust for inflation, and that's what it should cost to go backwards. So, it's like, wait a minute, we went from a dangerous gas to a safe gas; we went from leaky homes, 40, 50 years later, even in the UK, we now have more air-tight homes, and now you want to go back...?

I would question the assumption that you have air-tight homes, from personal experience. But that's a different question, let's not get into that...

Ah, empirical data: compared to 1970, we absolutely do. But the point is, where do you get your estimate from? Do you listen to somebody who is, for instance, very interested in selling hydrogen? Probably not, I suspect.

Exactly. That's something you always need to take into account. I would say, I mean, I know builders that I've worked with for so many years that I trust them to a degree where I say, come on, give me the real estimate; don't give me any bullshit, give me the real estimate. And they will do it, and I would trust them. But it is more reliable to get a larger sample, where you ask numerous people. And I would say you should have a minimum of 10, 15. But even if you only have five, it's a lot better than having nothing, I can guarantee you that.

That's right. The lesson is, if in doubt, build a database!

Exactly, exactly.

Let's do another of your heuristics. You've got one called "What's your Downside?" And I suspect this is - well, I know, because I've read the book - this is all about the fat tail. This is about the fact that.... There's a certain class of project that doesn't go 10% over, or 10% under, it does something completely different, right?

Yeah, they blow up basically. And the vast majority of project types that we have studied have fat tails. Actually, solar and wind power, and energy transmissions, and roads, and pipelines, don't; they have thin tails, which means the risks are small and easy to manage. All other project types have fat tails, and that's why you really need to watch your downside when you're doing projects, because that's where the fat tails lie, of course. And people often talk about the opportunities and risks as if they were symmetrical, you know; that you need to risk something in order to get the opportunity. And that is true, of course, but there's a very fundamental asymmetry between opportunity and risk, and that is: risk can kill you. Or, put you out of business, but it can kill you in some instances, like if you're a pilot, or doing extreme sports, or mountain climbing or whatever. And in business, it can kill you, your business, right? That's what risk can do to you. There's no opportunity that can compensate for that; if you're dead, you're dead, and you are out of business. So, that's why the main focus needs to be not getting out of business, right? This is what people like Warren Buffett understand, that they always look at the downside first; and then, of course, you go for the upside - once you've got your downside covered, you can go for the opportunities. But you never, ever forget the downside. Never. You can forget the upside for a period of time because you can survive, still, if you don't get knocked over by your downside. So that's the explanation, why we say "What's your Downside?", and keep an eye on that. And it might be, existentially, the most important piece of advice you can give anybody - a person or an organization.

That's funny. So, I do some angel investing, and one of the things I do with any company I invest in is I say, I want to see that you've got a risk register; startups don't have risk registers, generally. But I want to see it, because the thing that I slightly worry about: I know you're going to pivot, I know normal difficulties are going to happen, but what are the things that can really kill your business, that can really take it out completely? Unexpected or expected or whatever. So, even startups I try to get them to watch their big downside.

Yeah. I think that a risk register is a good start, buy my experience with people is that either people have this or they don't. It's much more of a mental attitude than it's about having a register. So somebody who doesn't understand downside, who develops a risk register, is not very well-helped by that risk register, compared to somebody who actually, it's in their guts, this feeling that downside is dangerous, and that's what I need to protect myself against. And if you have that, you can actually do without a risk register; I wouldn't recommend going without a risk register. But it's the more important thing for person to have - and I would say, especially at the startup phase, where things are small and nimble, and not so bureaucratic - that you already have risk registers and so on.

But what I'm actually doing, I think, as an investor... what I think I'm doing with the management is not asking them to do a bureaucratic exercise; actually, I'm signaling that I, as an investor, care about massive downsides, and I'm giving them a license to worry about things like, you know, cybersecurity. Because a lot of investors they're like yeah, yeah, let's go, and we want you to build the thing, and "move fast, break things." But I'm actually saying, hang on a second, yeah, but don't make stupid mistakes that cost you the whole company. So, I'm sending a certain message, rather than doing a bureaucratic exercise. I think.

Yeah, no, no, I think that you would be doing that, and I'm in total agreement with that approach. I'm just saying that some people think if they have a risk register, this thing on paper, then they're covered. That's not enough, that's all I'm trying to say.

Oh, the other thing I do is when bad things happen, I force them to pull out the register and say, well, was this bad thing - you said that now we're delaying the launch of a product - was that on your... was the reason on your risk register or not? You have to close these loops, don't you?

I like that.

I was very interested that you've got.... one of the projects that you've mentioned, that outperforms that, that is up there with the sort of more Lego projects, like solar and wind and so on, is transmission. And I was interested in that, because transmission is at the heart, by the way, of the challenges around the net-zero transition, right? Because we're going to electrify so much, and because renewable energy comes from new places, and it's variable, transmission is really, really important. And it's taking way too long: 10, 15 years to build transmission. And yet, here it pops up, and you're saying no. Presumably, you're saying, it may take long, but at least it's hitting budget, and it's doing what it says on the tin?

Yeah, I think it can be done. I saw a really worrying article the other day, I think, in the Financial Times about how the UK's grid is not even able to... it's not even able to connect all the wind farms and so on that are being built now, so there are billions and billions of pounds’ worth of investments that cannot be used. This is a tragedy.

Bent, let me just say: it's global. There's a connection pipeline, a connection queue, in the US that is bigger than all of the renewables that have been built to date. There's the same.... even in Scandinavia. I actually chaired a conference... And the problem between northern Norway and southern Norway, there's a break, or almost a break, in the grid, and it's gonna take a decade to do something about it. This is a universal problem.

We need to figure out a way to do that faster. And the way to do that is to build on the parts of the grid that are modular, and there are lots of parts that are modular, and just make that more efficient, like we've made everything more efficient for solar and wind and batteries. We need to do the same for the grid, and it looks like the grid has inherent elements that will actually allow that. So, that's something that needs to be given higher priority. I don't know if somebody has been sleeping, and not thinking that the grid needed to be brought up to standard in order to be able to accommodate all those new energy sources that are going to be plugged into it. And that's just one thing, then we also need the big connectors between different countries and so on; Norway and Denmark has a connector where you can run electricity back and forth, which is useful because Denmark is wind and Norway is hydro, so Denmark can use Norway as a big hydro battery, which actually happens. Certain days, the country produces 150% of their electricity needs, so 50% would have to be lost if it wasn't sent to Norway to pump water into hydroelectric dams. And this is happening elsewhere. This is what we need more of: interconnectors. I mean, the concept of the grid that we have today, or have had up until today, is too small; we can't have grids that are so local, they need to be much, much bigger geographically in order to even out loads and even to send loads between different grids.

Absolutely, I agree. You asked the question, why were people asleep? I think what you've got is because of energy efficiency improvements - for instance, LED lighting or more efficient motors - what you actually saw over the last 15 years is electricity demand in the developed world declining. So, the UK in 2005 used 37% more electricity in the domestic sector than it is using today. Against that backdrop, somebody has to go and 'OK' these huge, monumental investments for the stuff you're talking about; so, lots more electricity, more renewables, interconnections, and so on. And I'm trying to think how you would modularize what's, essentially, a backbone. The transmission grids - all grids - they're not quite monopolies, but they are place-to-place monopolies. How do you modularize... I can see how you can modularize installation of a cable or building pylons, but the actual grid itself kind of feels like one of a kind?

Well, no, I don't think so. Maybe when it's built out, but the individual elements of the grid are not one of a kind. So, substations are not one of a kind, for instance, and cables are not one of a kind, they can be built in modular fashion. And the way to drive efficiencies on that is to create competition between the people who are delivering this. So, I will see that the owners of the grid - that's the demand, you know, they're demanding these things - but there's a market that they're demanding it from. That market needs to be incentivized to become more efficient, and I think that that can be done, just as it's been done in other areas.

So, I know that one of the guests on Cleaning Up from a couple of years ago, Greg Jackson, who's the CEO of Octopus, I know that he would agree that if National Grid can't build a link that he needs, he should be allowed to go in and bid and get regulated, and win the right to do that. But the planning permission, the thing that takes a lot of the time, is actually getting social license; actually doing the consultations; only one company can do the engineering business case creation for a new link connecting to the grid. So, there are some things that feel hard to break up into your... to Lego-ize.

Well, that's the process itself. But actually, we give examples in the book about how you can modularize processes; it's actually important to think about both processes and products here, and not just think modularity applies to products - it also applies to process. But I don't think that's the only problem here. I do think that that the process of approval needs to be made more efficient, but another thing that is even more important, or just as important, is that people need to think ahead. So, the example that you just gave, that electricity demand has been falling because of increased efficiency: that's all very well, but it's been clear for a long, long time that the future is electrification. So, it's been clear that we would need these types of grids for quite a while, and there's actually no point in building all the wind farms and solar farms and so on if we don't have a grid that they can plug into. So I think that this indicates to me that the people responsible for the grid haven't been thinking slow, actually; thinking what is actually happening here? We might have increased efficiencies right now, but what will it look like in 5, 10 years? And 5, 10 years from now, it's going to be very different.

I think there's another piece to it, which is defining corridors - clean energy corridors, clean power corridors - and just pre-permitting, and getting all of the kind of social license piece done, ahead of when you actually need the power to be built. But there's a lot of things that would need to be done. In fact, there's a nice example in your book; Terminal Five, you talked about... what did you call it? You turn the milestones... you talked about using "inch-stones": if you really have got one big project, really make sure you're hitting your inch-stones, not just your milestones.

Yeah, so this is a concept we actually developed on a project we were called on to in Hong Kong, a high-speed rail project. The only 100% underground high-speed rail line in the world, with the biggest underground high-speed rail station in the world, smack in the center of Hong Kong. And it wasn't going well, so midway through, the project was broken with cost overruns, schedule overruns, a flooded tunnel, a tunnel boring machine underwater, and so on. So, really bad, but actually not so much worse than any other project - it happens much more than you think, and that's what explains the numbers that we started out talking about. And that's where we invented inch-stones. So, we basically reprogrammed the whole thing, and said, okay, we need to know from now until the opening date, everything needs to be broken down, and it needs to be completely clear what happens and who's responsible for what. And also how do we escalate this, if it starts going wrong, again; what happens, who's responsible, in order to make sure that we wouldn't have further delays while people found out who was responsible, and what to do and so on. And that's what inch-stones are really good for, and I highly recommend them for projects where you really want to make sure that you meet the deadlines and the budget.

And you have something to celebrate along the way. Every inch, you get a bit of a....

That's true, and we also develop early warning systems, where immediately, within a week, if something goes wrong, it will be flagged to the leadership. So, often what happens is things start to go wrong, and people are afraid to escalate it to the top, so bad news travels slow. And that's a slowness you don't want; you want bad news to travel fast, so you need to develop a culture where it's actually not only legitimate, but encouraged, that if something is beginning to go wrong, everybody on the project has a duty to report it to the leadership. So, we develop systems like that too. Again, something that is hugely recommended. And Terminal Five, were very good, Terminal Five at Heathrow that you mentioned. We made a point in this book of actually going out and identifying a lot of successful projects, because it's just too depressing to keep talking about the failures, I find. And I wanted this book to be a joy to write, so we made sure to find lots of examples of successes, where we could pick the brains of the people who were responsible for these successes - see how they think, see how their cognitive biases are different from ordinary people's cognitive biases and so on.

Or did they maybe get lucky? And the reason I say that is, on the Terminal Five build, that's a chap called Andrew Wolstenholme, who was later responsible for Crossrail; and I was on the board, I was on the Finance Committee, and basically, it went horribly. It did all the things that you talk about, all the bad things in the book, and information was withheld from us on the Finance Committee of Transport for London. I don't know whether he individually was responsible, but certainly by the management of Transport for London, there was key information about this thing spiraling out of control. And the Mayor, interestingly, had taken... There was one board member of TfL, who was also on the board of Crossrail, but he was a conservative. And so, as soon as Sadiq Khan took over, he was fired from that role, and then we had no governance link; we had no direct board to board Information link, it all was filtered through management. And when we were probing, the information was not provided. But that was under the same executive that you have as presiding over one of the one of the success stories in your book.

Yep, that is true. And we don't we don't ask for anybody to be perfect, you know. But we do ask, when we study success, that it's highly unlikely that it was luck. And we do that by looking at people who have been able to deliver success several times, and Andrew Wolstenholme actually has been able to deliver success several times. It's very unfortunate what happened with Crossrail, but it's a classic. There's even an article in this week's Economist about it, and even they get it wrong - and they're usually pretty good journalists at The Economist. They say that the reason that we had the delays and the cost overruns is because it's very complex to build under a city like London; they actually got that pretty well, you know, they were very good at that. It was the IT, and it's often the IT that is the problem. So, the signaling systems, the safety systems and so on, had been underestimated. So, this is optimism, you see how we get right back to the root causes here. It actually turned out - and I've discussed this in depth with Andrew Wolstenholme directly, face to face, we actually do executive training together at Oxford University, and elsewhere - and he acknowledges that they were optimistic about how long it would take to make the signaling systems and the safety systems reliable.

Maybe. There may be a little bit of bias in his diagnosis. Because certainly, after it became absolutely clear... And I mean, just for the listeners who don't know, we were supposed to believe that in July or August 2018 - and the Queen, bless her was supposed to come and open it in December 2018 - we were told it was still on track, we were being told it's still on track. And of course, it ended up being delayed by four years. Afterwards, all the stories emerged about how everybody working on it knew it was off track. The only people who didn't know - or didn't want to know, or didn't want to believe it - were the higher echelons within Crossrail, and particularly within Transport for London, including the commissioner, who - evidence came out afterwards - had actually interrupted that information flow and changed information that was supposed to be given to us on the board. And he's now going off and doing other projects, was given his honours, and is now doing... I think it's the renovation of the Houses of Parliament project. This was not a well-managed project with an IT problem.

Okay. I mean, you may have inside information that I don't have, but all I can say is that that's a classic, that the information doesn't percolate to the top where it should go, and that's exactly what you need to avoid. That's how we designed - like the high-speed rail project that I explained about in Hong Kong - we designed it in a way that would do exactly the opposite of what happened on Crossrail. And it actually worked, you know. That was delivered to the schedule that we developed like that, that we programmed with the inch-stones, and it was delivered to the budget that we estimated at that time, halfway through the project. So, we eliminated the optimism that has been in the project from the outset. Midway through, we re-baselined everything, reconfigured everything, re-programmed everything, and de-biased everything, including the optimism bias, and got a realistic estimate. Like we talked about for the kitchen renovation or home renovation before, that's what we made for the high-speed rail line. And they delivered to that.

To be fair... To complete on the Transport for London experience, we also had the Garden Bridge, where the only thing I ever signed off on, on the Finance Committee, was I think it was £4 million to do a feasibility study, where I tried to set the condition that we would not put money into it, because it was supposed to be all private sector. And it ended up... I believe we ended up spending £40 or £50 million; I still don't know where it went, I have no idea. I don't think anybody knows where the actual cash ended up. The budget exploded to £200 million plus, and it eventually got killed by Sadiq Khan. Eventually, after a lot of to-ing and fro-ing. I'm conscious of time, Bent. I want to finish if I could, rapid-fire, if I might: some of the other technologies around the net-zero transition, your thoughts about whether they're likely to be... if you could kind of keep it short, just say well, sounds like it's going to be easy, that one makes me worried, if that's okay with you. So, I've got a long-distance cable... I'm an investor, it's a cable from Morocco, all the way around the coast, laying it under the sea, and coming into the UK. Is that going to be a top of the table, or a bottom...? Is it going to be a bad performer, or down with solar and wind, to be a good performer?

It's going to be hugely expensive but if done right it's a good performer.

Okay, what about heat pumps? We need to switch to heat pumps... I'm prejudging, I've already revealed my bias against hydrogen for heating. So, rolling out heat pumps - gonna go according to plan as a program, or not?

Depends if you're talking about air-to-air heat pumps or ground-to-air heat pumps. In Denmark, where I'm from, there's some resistance to the air heat pumps because they make noise, so you get them in a neighborhood, you get noise, and that's a problem, and it would be nice to have that solved. But overall, I think heat pumps, they can be delivered hugely successfully, fast. They are very modular, and the good thing about them is that you don't need a whole system, you can do it individually on each home. And it's being done now, there's lots of heat pumps going in around Northern Europe.

So, Episode 122, we had Professor Sir Chris Llewellyn Smith, who's looked at what do you do for long-duration storage? I completely agree, interconnections and so on, but there are times - the Germans call it the dunkelflaute - where there are weeks with low wind, and what Llewellyn Smith has discovered is, they might happen year after year for a few years; you could have some bad wind years, and then have a dunkelflaute. So, his solution is hydrogen storage, deep underground in salt caverns.

Right. There are lots of different solutions. So, I'm not a specialist on the on all the different alternatives of energy storage.

But would that be an easy type of project to get [done]? If you were consulting to somebody trying to build one would that fill you with... Would that be an easy or a difficult in your book?

So, I wouldn't be able to call that one. I don't know enough about the details. But I do know that there's enough alternatives that we're going to find something, and there's no question that we need something for that kind of situation; there has to be some kind of baseload, that we can always deliver, even when the wind is not blowing and the sun is not shining.

Another one, where again, I will confess that it's a company that I advise in this case: Episode 58, John Redfern, closed-loop geothermal. You go very far down - this is high temperature geothermal, not ground-source heat pumps - you go three, four kilometers down, you go a few kilometers out, you're drilling, drilling, drilling, and then you're linking it up to make an underground radiator to bring out the heat. Is that the sort of thing that tends to go well? Or is that going to be a fat tail type one?

Again, I'm not a specialist here. But I've heard people that I trust saying that that's extremely risky. Drilling very long holes into the ground like that is very risky. So, that's all I know about that. I also know that geothermal has been used for a long time in Iceland, for instance. So, there are good geothermal plants that actually work.

To this one, I think John would say, ah, you know, Professor Flyvbjerg is talking about fracking, or about having to... So, he's doing only the drilling.

I was not associating this with fracking at all. I was thinking of exactly that kind of drilling. And what I'm saying is, I don't know enough about it, I'm not an engineer on this issue. But I've heard people say it's risky.

Okay, so that's a few of the new business models, a few of the things that we're likely to see evolving during the transition. But overall, then, final question, with all your knowledge about projects, how they can go wrong, most of them 99.5% currently, but hopefully, less and less if people read your book, which they should. But the transition: overall, are you optimistic, or are you not optimistic?

If you're talking about the energy transition, I am optimistic. And this is actually one of the things I discovered while doing the research for the book is that, wow, we're actually lucky that the things that we need are being developed at speed and at scale right now, following the basic heuristics, if you wish, that we that we lay out in the book. So, I'm optimistic about the situation. I think that it's all about electrification now, if we're talking about the energy transition. And then we need to produce that electricity on the basis of renewables. And from what I've seen, that can be done, we can do that.

Bent, Professor Flyvbjerg, enormous pleasure speaking with you today, thank you so much for sharing time with us here on Cleaning Up.

The pleasure is mine, thank you so much for having me.

How Big Things (Should) Get Done - Ep128: Prof. Bent Flyvbjerg

Listen On

Featured Episodes

Recent Episodes