Building Bridges

How many times have I rolled my eyes at the assertion, “But, you can’t build a bridge with Scrum”? The claim annoys as much for its misunderstanding of Scrum as for its naïve view of projects. Curiously enough, I recently discovered support for Scrum in a recent book about big projects. Maybe you can build a bridge with Scrum after all.

Bent Flyvbjerg’s “How big things get done” is one of the most useful books on project management to appear in a long time./1/ Flyvbjerg is a leading expert on big projects, having studied thousands of so-called “megaprojects”. Thankfully, Flyvbjerg didn’t create project management method with this book. Rather, he explores the factors for success and failure in projects. With that, he addresses the central questions facing any project—even a kitchen renovation, which counts as “big” for the couple doing it. Why are we really doing it? How should we approach planning and delivery? How do we confront cognitive biases? What should the team look like?

Despite not being a methods book, it contains truly useful advice, which he further distills into heuristics at the end. Among these are the following:

• Ask why: successful projects have a clear understanding of their results and the true needs those results are meant to fulfill.
• Think slow, act fast: Successful projects use careful and sometimes long planning phases, but execute quickly. Planning is cheap, relatively speaking. The more questions, issues, risks etc. that are addressed in planning, the better the delivery. A fast delivery means less time where things can go wrong.
• Take the outside view: Avoid the fallacy that your project is unique. Rather, it is most likely part of a class of projects that have been done before. Therefore, you can use the experience and averages of that class for planning, estimating, risk mitigation etc.
• Build with Lego: take a modular approach to keep the focus on smaller, more manageable chunks that can be reviewed, adjusted and repeated.

Flyvbjerg also emphasizes how projects “don’t go wrong, they start wrong,”: either through a poor understanding of “why”, insufficient planning, bad estimates etc. He points to several examples where the rush to commit or to “start digging” dooms the project to a long, painful and expensive delivery. Reality always catches up with you. Whether it’s politics that corrupt an otherwise good idea or cognitive biases such as optimism, anchoring or uniqueness bias that mislead us, failing to address a problem during planning is guaranteed to cost more money and time during delivery.

Building Bridges

The reproach to Scrum (or any agile method) is that it would be inadequate for a big project or for something physical that can’t be corrected easily like software can: “How am I supposed to use one-week sprints for pouring concrete and welding steel?” Well…

One of the most important messages in “How big things get done” is that success comes through an iterative, experimental approach performed by an experienced team. Flyvbjerg highlights the planning methods of the architect Frank Gehry or the film-making process at Pixar. Both use extensive empirical testing and adjustment during design and planning. Gehry builds many physical and computer models of his buildings, reworking them to test every aspect—design, structural stability, usability etc. From Flyvbjerg’s telling, Gehry iterates sometimes multiple times a day. Pixar cycles through as many as eight iterations of a film using increasingly detailed mock-ups and gathering audience feedback before committing to production. Gehry’s projects such as the acclaimed Bilbao Museum typically come in on-time and within budget. Pixar’s films like Toy Story, Up and Inside Out have nearly all been highly profitable blockbusters. Both regard planning as the most important work in the project: “think slow, act fast”.

Scrum is nothing if not an empirical, iterative approach. It’s pillars are transparency, inspect and adapt. It sets maximal value on learning from past experience. So, while they might not be using sprints, dailies and a scrum master, Gehry and Pixar are doing exactly what Scrum or any lean/agile method advocates./2/

Moreover, both have ritualized how they move through their iterations. Pixar uses what it calls “Braintrust” meetings where open feedback is given to the film team. These meetings and the resultant feedback stick to rules that maximize candor, yet preserve a director’s creative freedom. Making them into a ritual reduces the overhead involved in coordinating the teamwork. That is, according to the Scrum Guide, why Scrum has its sprint structure: to reduce the friction around setting up meetings etc. while ensuring that the important things—planning, feedback and learning—happen./3/

Finally, Flyvbjerg cites the example of the new Terminal 5 at Heathrow, which, despite initial hiccups with the baggage delivery, is regarded as massively successful. Not only was it on-time and within budget, it received accolades for two aspects: team responsibility and the working conditions. Despite needing hundreds of contractors, the British Airports Authority (BAA) created a “one team” identity within the project, in part by structuring contracts so that the contractors would be incentivized to solve problems together rather than play the blame game. The BAA also invested heavily in team dynamics and ensuring that workers’ needs (like safety glasses etc.) were met. They also tasked workers with creating their own standards and approaches, a hallmark of all lean and agile methods. The sum of these measures amounted to highly committed teams and enormous professional pride in being part of “T5”. The list of Scrum Values matches nearly one-to-one the T5 approach: commitment to a common goal, focus, openness, respect and courage.

The fallacy in the bridge question comes partly from naïvely thinking that projects are mostly about pouring concrete (i.e. building) in a sheer linear fashion. We learn from Flyvbjerg, however, that the thinking slow phase—where we understand the why and where we plan, model and design—lends itself well to an iterative, experimental approach. And that seems to be key to their success. With good planning, modern project delivery has low uncertainty and proceeds in a factory-like fashion. Even his “build with lego” suggestion to use a modular approach succeeds due to iteration and empirical review.  According to Flyvbjerg, IT projects (some of which surely were done with Scrum) have an average cost overrun of 73%. Applying his lessons, IT projects would do well to spend more time on active planning and benefits analysis./4/

The bridge question also fallaciously sees Scrum as mainly about sprints, stand-ups and retros, even though these elements are only there to support iterative, experimental learning.  

To emphasize: “How big things get done” is not a Scrum or agile book. Neither “Scrum” nor “agile” appears in the main text. Flyvbjerg does refer positively to the “lean startup” model, where an iterative planning phase is key to understanding customers’ needs./5/ But, that’s it. Flyvbjerg, I suspect, would roll his eyes, too, at the suggestion of using Scrum. Nevertheless, the book shows that the very factors that lead to successful big projects match agile and lean thinking.

The book is truly a “must read” for all project managers and senior executives running projects. Whatever “bridge” they want to build, they will profit from the book, whether they decide to use Scrum or not.

Notes

/1/ Bent Flyvbjerg and Dan Gardner, How Big Things Get Done (New York: Currency, 2023). German translation: Bent Flyvbjerg and Dan Gardner, How Big Things Get Done: wie Projekte gelingen: von der Küchenrenovierung bis zur Marsmission, trans. Anke Wagner-Wolff and Gisela Fichtl (München: Droemer, 2024).

/2/ For more on Pixar’s ways of working see Edwin E. Catmull and Amy Wallace, Creativity, Inc: Overcoming the Unseen Forces That Stand in the Way of True Inspiration, expanded edition (New York: Penguin Random House, 2023).

/3/ Jeff Sutherland and Ken Schwaber, The Scrum Guide: The Definitive Guide to Scrum: The Rules of the Game (ScrumGuides.org, 2020), https://scrumguides.org/docs/scrumguide/v2020/2020-Scrum-Guide-US.pdf; Jeff Sutherland and Ken Schwaber, Der Scrum Guide: Der gültige Leitfaden für Scrum: Die Spielregeln (ScrumGuides.org, 2020), https://scrumguides.org/docs/scrumguide/v2020/2020-Scrum-Guide-German-männlich-male-version.pdf.

/4/ John L. Ward and Elizabeth Daniel, Benefits Management: Delivering Value from IS & IT Investments: Delivering Value from IS and IT Investments, 1. Auflage (John Wiley & Sons, 2005). Ward and Daniels hypothesize that with better benefits analysis, about half of IT projects would never be approved because their imagined benefits are so marginal.

/5/ Eric Ries, The Lean Startup: How Today’s Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses (Crown Business, 2011).