Anjali Sastry & Kara Penn: Making a success of failure

Planning projects in repeatable chunks rather than in one grandiose shot and prioritizing tasks that test basic assumptions can help you learn from mistakes


Penn (left) and Sastry say useful failures reveal something new about the world or help rule out an idea.
Penn (left) and Sastry say useful failures reveal something new about the world or help rule out an idea.

If you’ve ever played Hangman or Bullseye-Magpie, you know the misses tell you nearly as much as the hits—especially early on in the game. In Fail Better: Design Smart Mistakes And Succeed Sooner, authors Anjali Sastry and Kara Penn talk about ways in which misses, or failures, can help improve the chances of success in business.

Sastry is a senior lecturer at the MIT Sloan School of Management and lecturer in the department of global health and social medicine at Harvard Medical School, US. Penn is co-founder and principal consultant at Mission Spark, a Colorado, US, headquartered consultancy for non-profits and social enterprises.

Sastry and Penn explain: “We’ve seen it all over the world, in companies, non-profit organizations and executive classrooms, not to mention the media: Everyone knows we should learn from failure. But there’s little guidance for how to do so. Stories of corporate heroes succeeding after they’ve endured setbacks may be inspiring, but they fall short of showing how to create the conditions for—and capture the benefits of—productive failures. We wrote Fail Better to fill that gap.” Edited excerpts from an email interview:

What is good failure?

Useful failures reveal something new about the world, rule out an idea, or suggest a new direction in ways that are small-scale, affordable, linked to broader goals, and reveal insights quickly.

This informational value of failure is not a novel discovery, as any expert in software design, product development, or process improvement would agree. But too often, potentially instructive experiences go to waste because people lack methods for extracting their lessons. Companies need to learn from all sorts of failures, large and small. There is much to glean from every bungled product launch and failed expansion strategy. But also think of the waste involved in creating reports that nobody reads or sitting in meetings that don’t contribute to the business. We’ve learnt that in all sorts of companies and organizations, employees can see where to make things better, but lack the framework, platform, even the language and permission to make needed changes. By equipping everyone to identify failures and seek better ways to do things, our method helps companies to learn from their own experience.

We aim to take this thinking one step further with an even bolder idea. What if you could proactively design the most informative failures—in a sense, orchestrate your own mistakes? To do all this safely and effectively, we learnt, companies need a systematic method. Our book provides a practical guide for teams to (1) design their work to maximize the chances that any failure is productive; (2) build in experimentation that reveals critical insights along the way; and (3) share what’s learnt in ways that can change future personal, team and company behaviour.

How do ideas like crafting a problem statement and defining the product help?

Fail Better—Design Smart Mistakes And Succeed Sooner By Anjali Sastry &Kara Penn, Harvard University Press, 256 pages, Rs603
Most teams begin a new project by defining deadlines, tasks and outputs. Do what’s necessary while minimizing formalities, but add some early steps. Mapping out the underlying thinking can allow future surprises to be more productive.

We tell teams to start by pinpointing the problem they aim to solve. This seems basic, but often is overlooked. Yet a common project failure mode is spending most of the effort solving the wrong problem, realizing the mistake only when it’s too late to change course. Defining the problem at the outset will prime the team to consider new perspectives.

Getting the problem right is not the only important part of a launch. Many teams focus on their official outputs: the report to be delivered, the slide deck for a client, the new product design. But in the end, their efforts are taken in pursuit of outcomes—behaviours or changes in the world that the project’s output should engender. We advise teams to ask, how do outputs from our project solve the problem we have just defined? Making explicit the links between desired outcomes and the actions that you think will get you there. Equip your team to pounce on any unexpected results whenever they arise—then to figure out how to adjust their plans as needed.

How can organizations encourage “learning by doing” among employees and teams?

Two concepts can be useful to build into every project to encourage learning by doing: First, look for opportunities to scale back on grandiose one-shot efforts. Iterate instead. Plan projects in repeatable chunks that allow ideas to be tried out early—and then revamp, iterate, and try again. Prioritize actions that will yield the most useful information first: new data that could challenge a core assumption, test a prominent concept, or suggest a novel solution to the problem at hand.

Second, find three things to take away from every project, and share them. As you wrap up your work, take time to extract and share actionable lessons. It need not be complicated. Try just three steps: Identify and document one thing you did that you should not do in future projects, one new thing you tried that is worth doing again, and something new you learnt about the world that could be useful to others in your company.

What are the key lessons here for managers?

Managers need to be both kind and discerning, because learning how to talk about failures in a professionally appropriate manner may take practice. It’s key to demonstrate compassion in the face of disappointing results, while recognizing that not every mistake is beneficial. Failures arising from lack of attention, insufficient effort or disregard for the data must be weeded out from the rest, or you risk tolerating poor performance. The current Ebola crisis and lack of early response by the global community serves as an example of a bad failure: Correcting outcomes is more costly, in lives, infrastructure and money, than it would have been had action been taken earlier.

More From Livemint