But I didn’t understand what it meant in those relatively early days of Agile. The expression conjured up missed deadlines, broken promises, and an undisciplined approach to software development. I found the term threatening—it went against everything I believed in as a project manager. This quote by a CIO in a recent article summarizes how I felt back then: “I really don't like the term 'fail fast.' I don't like the term 'failure.' It's got so much laden on it. I understand the folks in Silicon Valley; they mean something different.”
As I learned more about Agile and became certified in 2010, I came to understand that in many respects failing fast was a reaction against the “do it right the first time” slogan of the quality movement of the 80s and 90s. I learned that failing fast does not refer to a lack of success—simply a different meaning of success. Failing fast is a way to experiment quickly and be OK with throwing away unsatisfactory results. It‘s also a way to ensure that “sunk costs,” money already spent, is not a factor in future decisions.
So what exactly does it mean to fast?
Most definitions of ‘fail fast fail often’ include some of these elements:
- Speed of execution. A project where the “speed of execution is a lot more important than perfect execution.[i]” In other words, it’s more important to get results quickly than to get perfect results. A feature of the old quality models, which required doing things right the first time, was having zero defects. Having implemented several large projects right after this kind of quality training, I now realize that the teams spent an inordinate amount of time trying to ensure that every aspect of the project would work perfectly. Back then even small defects were discouraged. In other words, the cost of preventing defects was far greater than the benefit of an earlier implementation.
- Taking a large project, breaking it into small pieces, and time boxes (iterations) mean that more features can be completed in shorter amounts of time.
- Experimentation and extensive testing. This includes trying new things, learning from these trials, and refining the inputs and/or tests in order to achieve different results. The number of tests is less important than learning from the results and making changes.
Failing fast in the age of pandemics
But let’s get back to the topic at hand—the need to fail fast, particularly as it relates to Covid-19. Many in the health care industry are realizing the advantage and even the necessity of “failing fast.” This is particularly true in the area of vaccine development. Vaccines usually take anywhere between 2 and 5 years or longer to develop, test, manufacture, and distribute. There are many reasons for this. Before being approved for manufacturing, vaccines typically go through many phases from animal trials to extensive testing on a variety of different human demographics.[ii] What, then, makes epidemiologists think that a Covid-19 vaccine can be made to fail fast and come to a successful conclusion? Here are some examples of how this is being done.
- One of the best examples of failing fast is by concurrently developing vaccines at the same time manufacturing and distribution channels for that vaccine are created. For this to work, several things are necessary. Industry and scientific leaders, regulators, and others, groups who do not always work well together have to do just that. And they have to share rather than withhold information--internationally.
- Traditional thinking in vaccine development was that the creation, manufacturing, and distribution of that vaccine happened sequentially rather than concurrently. There was too much financial risk in thinking about manufacturing and distribution before the vaccine went through all the trials and was approved. Since most vaccines never make it to human trials, let along through all the trials to approval, why incur the huge cost to manufacture and distribute something that was going to get thrown away? That’s different with Covid-19.
Failing fast requires organizations to assume risk that was not thought possible in the past when it was too great a cost to develop these channels before vaccine approval. According to Dr. Fauci, this concurrent development can shave many months off the time that it typically takes to develop vaccines.[iii] Because of the need for speed and financial benefit, this kind of unprecedented international collaboration has begun.[iv]
- Partnering with other organizations to develop the same vaccine. This is what’s happening in the development of the Covid-19 vaccine. Unlike vaccine development in the past, many organizations are working together on the same vaccine in order to speed up development and approval.[v]
- Using AI to study the disease’s mutation patterns. This new coronavirus probably mutated from animals to humans and continues to mutate. Most of these mutations appear to be minor, but new evidence suggests that newer mutations help the virus better penetrate the body[vi]. Trials need to take these mutations into account and AI speeds up this process.
- Using AI to determine what an infected cell looks like. AI can look at the many and complex cell attributes by looking at the problem holistically and predicting which potential vaccines are most likely to succeed in clinical trials. Humans are not good at understanding what a sick cell looks like.
All these fail-fast measures require strategic and innovative thinking, strong executive leadership, and a commitment to work collaboratively rather than competitively. But in a world-wide pandemic, failing fast and failing often is exactly what’s needed.
[i] Forbes, Sunnie Giles, 4/30/18, https://www.forbes.com/sites/sunniegiles/2018/04/30/how-to-fail-faster-and-why-you-should/#758c5b92c177
[ii] Rob Grenfell & Trevor Drew, The Conversation, February 17, 2020, https://www.sciencealert.com/who-says-a-coronavirus-vaccine-is-18-months-away
[iv]World Economic Forum, Charlotte Edmond, May 14, 2020., https://www.weforum.org/agenda/2020/05/coronavirus-covid-19-vaccine-industry/World
[vi] https://www.washingtonpost.com/science/2020/06/29/coronavirus-mutation-science/?arc404=true, Sara Kaplan and Achenback, June 29, 2020.