This is a reprint of a blog post I wrote for Product Creation Studio

Waterfall project management is the paradigm taught by the Project Management Institute, the grantee of the Project Management Professional(PMP) certification. It’s so ubiquitous that if someone says “project management,” waterfall is probably what they mean. It’s an incredibly powerful approach, allowing one to manage projects of immense complexity, efficiently and successfully. The intention of waterfall is that the tasks flow from one to another as effortlessly and inevitably as water falling down a mountain.

This approach is well-suited for highly complex, yet predictable projects, like constructing skyscrapers, building cars, and even organizing a national tour of the Rolling Stones. Just as you would never start building a skyscraper without a complete design or playing a concert without the songs written and rehearsed, you also need a complete plan before beginning a waterfall project. In waterfall, every task is listed and the amount of effort it will take is estimated. Project managers note the order in which tasks must be carried out and specific tasks are assigned to the responsible parties. Tools like Microsoft Project are used to turn these inputs into a schedule. As you move through the plan, materials like re-bar and tires are delivered to the right place at the right time. If they arrive too early, there is no place to store them, arrive too late and it delays the project. Some of these inputs, like booking a large venue for a concert, need to be scheduled far in advance. One of the advantages of waterfall is its ability to focus on the critical path. As a project manager, this helps you know where to focus your resources and attention.

The types of projects that are best suited to waterfall are those that exhibit little change from one project to the next. The vast majority of the workers, from people swinging hammers on a construction project, to people attaching car doors on an assembly line, will have worked on very similar projects using identical tools. This routineness vastly reduces the uncertainty and risk. It also makes estimating how long tasks will take fairly accurate, since they are tasks that the practitioners have done hundreds of times before.

This is a far different situation than in product development. Tools like miniature sensors, 3-D printers, and low-power microprocessors are constantly changing. This makes what was impossible a few years ago, straightforward today. While on the one hand these innovations are incredibly exciting, they also make the product development process far more uncertain and risky; it becomes impossible to assemble a precise project plan.

Before a project is launched, every effort must be made to reduce uncertainty (e.g. soil samples taken and computer models run). Waterfall does a bad job of dealing with uncertainty. When the unexpected happens (e.g. a big storm the night of a concert), these projects tend to break big. In product development, the unexpected is typical, making these breaks common.

The development of the F-35 fighter jet is a good example of why waterfall works poorly for product development. This program created variants of a fighter aircraft suitable for the US Air Force, Marines, Navy, and the militaries of at least ten other countries. The requirements for this program were incredibly challenging: the plane needed to have stealth capabilities, be able to launch ground attacks, maintain air superiority, and have the ability to take-off and land vertically. The plan was to design, test, and start production concurrently. The waterfall was very efficient, but it broke under the reality that the design was imperfect. Most of the people working on the project had little or no experience designing a fighter from scratch; the experience needed to build a solid plan didn’t exist. Even if it did, because they were using the latest technology, nothing in the project was routine. Testing revealed problems on planes that had already been manufactured and the manufacturing process needed to be changed, driving costs well beyond the projections. The F-35 is “ plagued with design flaws” and is at least $165 billion over budget.

In my own experience, I worked at one company that required a detailed waterfall plan for the entirety of a project — including commitments from managers on three different continents that engineers would be available on the exact days the plan said they’d be needed. Since the first part of the project was conducting feasibility studies and developing the architecture of the solution, it was impossible to accurately say how long it was going to take and when the engineers were going to be needed. One of the executives suggested we do statistical modeling to understand how long tasks were going to take, not understanding that these tasks were unique and had never been done before. We estimated durations based upon how similar projects had gone, but the uncertainty remained large. After failing to launch the project and restarting several times with new project managers, the company finally gave up.

Waterfall project management has many useful tools, like establishing dependencies and a critical path, which can aid in hardware product development. However, the paradigm of creating a complete plan at the beginning of a project is broken, based on the incorrect assumption that we know enough in the beginning to create an accurate plan all the way through to product launch.