Abstract
Professor Bent Flyvbjerg has spent much of his long and distinguished career pondering why infrastructure projects, from the routine to the exotic, typically take longer, cost more, and deliver less in benefits than promised. He has looked at pretty much everything in the built environment, highways and railroads, bridges and tunnels, dams and pipelines, schools and museums, and solar farms and nuclear power plants. Along the way, he has also developed what is probably the most extensive database on project performance in the world. Simply stated, Professor Flyvbjerg knows projects.
In his recent book with Dan Gardner, “How Big Things Get Done,” he lays out, clearly and concisely, not only why some projects fail but what he has learned over the years about what leads to success. To someone who has spent more than 50 years in and around the “infrastructure business,” the book reads as something of a memoir in a “been there and seen that” vein. As an example of how not to do a big project, he cites California High-Speed Rail. Approved by voters (including this one) in 2008, the project was to connect Los Angeles and San Francisco in a 2 ½ hour trip that would cost $86. At a total cost of $33 billion, it seemed too good to be true. Unfortunately, it was. As I write, the total estimated cost is now $128 billion and the only operable section of the system will connect Merced and Bakersfield, two cities in the Central Valley with a combined population of less than 500,000 at an estimated cost of $23 billion. To put this in context, this is more than enough to buy every driver in the two cities a new Tesla. So, what happened? Flyvbjerg and Gardner attribute this failure (and many like it) to a paradigm of “Think fast, act slow.” In other words, secure a funding commitment and then rush through the planning process so the “real” work can begin. However, successful projects adhere to the opposite; “Think slow, act fast.” Do the planning, analyze the risks, question assumptions, rinse, and repeat. Only once this is done is it time to proceed to construction and proceed smartly to the finish line. Both the Hoover Dam and Empire State Building, two iconic construction feats, followed this mantra and were completed ahead of schedule and under budget. It can be done!
The first question should always be “Why are we doing this and what do we want to accomplish? In the case of California High-Speed Rail, one of the drivers always seemed to be “The Chinese and Japanese have high-speed rail so we should too.” The fact that conditions in those countries that supported high-speed rail were not present in California was never really scrutinized to the degree it should have been. Ergo, California is working to deliver “a bullet train to nowhere.”
Although Flyvbjerg’s work has focused on big construction projects, his message applies equally well to any activity that requires planning and execution, whether renovating a kitchen in a 19th century brownstone or justifying a new library. From personal experience in the public sector, I can attest to the unpopularity of asking lots questions at project kick-offs when the primary goal often appeared to be getting started on justifying a pre-determined political conclusion rather than identifying a problem and crafting a solution. Of course, if the area where a new facility was wanted by constituents didn’t meet minimum demand thresholds, there were always public employees to blame for applying the wrong criteria. After a while, even competent and ethical professionals will say “whatever” but then have to explain why the user demand never met the plumped-up projections used to justify the project.
The meat of the message deals with successful projects and the people who brought them to fruition. For their examples, Flyvbjerg and Gardner chose Pixar (of movie animation fame) and the architect, Frank Gehry. Both take advantage of extensive testing and simulations in the pre-project stage to run through endless possible scenarios and use feedback to hone the final plan to a razor’s edge. Only when convinced that the movie script or final design is the best it can be do these organizations move to the implementation stage. They know that changes on the computer are cheap while changes in the field can bust through budgets and schedules. Although swashbucklers will occasionally astound the world by making it up as they go and pulling off an outstanding success, these “miracles” are more one-off anecdotes rather than a normative model of good project delivery.
There are two concepts called out for special attention. One is Reference Class Forecasting or RCF. Drawing on the exhaustive database he has compiled on many, many types of projects, Flyvbjerg can predict with considerable accuracy how much a particular type or class of projects should cost and how long the construction process should take. By using RCF, a manager should be able to tell whether budget and schedule estimates are realistic and whether milestones or “inchstones” as Flyvbjerg calls fine grained monitoring points, are being met. Perhaps not surprisingly, one of the obstacles to more widespread use of RCF is that many project advocates believe that their specific bridge or rail project is unique and not just one of many that have similar, if not identical, characteristics. However, at the end of the day, most projects in a class entail roughly the same steps and if a project’s estimated cost or schedule fall way outside the reference class, it should set off alarm bells.
The other concept is that of modularity. That is, building big things out of small pieces. Using the model of Lego blocks, he makes a powerful case for scaling up simple components into a larger whole. In one case, two or three classrooms may make a school. In another, it may require eight or ten. The key component is the relatively simple classroom module which can be combined in many ways offers and repeating what works. In the energy field, nuclear power plants are notorious for coming in over budget and behind schedule. This is mainly due to each one being unique or nearly so in design and built so infrequently that the opportunity of transferring knowledge from one project to the next is diminished. Wind and solar power projects, on the other hand, lend themselves to modular and scalable construction. Assembling factory-built components turns out to be both faster and cheaper than constructing a complex, bespoke facility on site. Should demand estimates change at some point, wind and solar can be dialed up or back with relatively little impact. Nuclear plants do not have this flexibility.
The final chapter lays out what could be considered the eleven commandments of good project leadership. Although there’s little novel in these heuristics, it does give one pause to consider why, for example, one would need to be reminded that assembling a good team or doing proper risk management is important. Unfortunately, the sad truth is that as one drills down into failed projects, you find that many of these commonsense activities were ignored or treated superficially. This is despite the enormous financial benefits that could be realized by doing projects better. With tens of trillions of dollars’ worth of infrastructure being constructed globally, even modest savings could amount to hundreds of billions. As Flyvbjerg notes, “These are world changing numbers.” He cites a German study from 2020 that estimates that the total cost to end global hunger by 2030 would be $330 billion over ten years. Regardless of the validity of this estimate, it does put the potential value of better project delivery in context. If the reader takes nothing else away from “How Big Things Get Done,” they will still have gotten far more than their money’s worth.