An empirical study of failure: how to make mistakes your friend

Don’t be afraid to screw up — it’s all about the learning

I spent a year studying mistakes in creative processes and interviewing a bunch of guys about it — Here’s what I found out!

A few details about the study

Over a period of a year and a half, I gathered data from projects and workshops, and interviewed people from various fields: journalism, software development, game design, and education. I looked for common patterns in the data I gathered, to try and form a conclusion about two core questions: 1) what can mistakes teach us?, and 2) in what ways might we use mistakes to learn?

What can mistakes do for us?

It should go without saying that we can learn from the mistakes others have made, e.g. by reading post-mortems, case studies and failure stories. This is something we should definitely do, because it allows us to avoid repeating the mistakes others have made. In many cases, avoiding past mistakes allows us to avoid negative and pointless consequences. When possible, do NOT fail fast to succeed soon. Just go straight to succeeding :-)

It’s not always possible to learn from past experience, though. When we create something entirely new (i.e. when we design), we are dealing with complex or wicked problems. This study concerns mistakes in the context of these kinds of problems, and here’s what I found:

1. We can only learn through mistakes if we test. It is through trying things out that consequences (both good and bad) become apparent. This is probably not surprising to you.
2. Mistakes have an eye-opening effect. Failure calls our attention to something which isn’t as we would like. It doesn’t mean we necessarily know how to fix it, but the first step to solve a problem is to become aware of it, and failure plays a big part in that.
3. Mistakes help us answer questions. Diego Rodriguez said “a prototype is a question embodied”. A powerful statement which very elegantly makes a point. I would add that prototypes are also knowledge, assumptions and hypotheses embodied. Testing such an artefact provides answers about all of these things.
4. Mistakes generate new questions and hypotheses. When new insights are gained about questions, assumptions, hypotheses and knowledge, new questions and hypotheses emerge.
5. Mistakes show us opportunities where we might improve. When something works, we repeat it. It can become a convention, best practice or standard — routine stuff we just do without thinking about it. Conversely, things that fail show us an opportunity to improve, which means failure can be a path to iterative improvement
6. Mistakes show us the way. Conceptually speaking, mistakes are part of simultaneously creating the path and shining a light on the path through a design space. Both success and failure enable us to take one more step, and one more step, and so on. This means mistakes are not unwanted side effects, but an essential part of the way through a design process.
7. Mistakes can result in wildly original ideas. Finally, mistakes don’t necessarily result in unwanted consequences — they can also be a source of originality and result in radical innovation. Coca-Cola, Teflon and Gorilla Glass are three examples of products that were the result of somebody making a mistake, with unexpected but great results.

Why mistakes have a bad image

According to both observations and interviews in my study, people typically do not like making mistakes. You might even say it is stigmatised. The following factors contribute to mistakes having a really bad name:

1. Blame. Perhaps surprisingly, this is often a case of people blaming themselves, because they don’t want to let a team down.
2. Time pressure. Submit people to hard, unwavering deadlines and you leave no room for error.
3. Social, institutional and family pressure. Often there will be expectations and pressure about doing “the right thing”. We all know how pressure can affect us, and this is no different
4. Punishment. It’s a sure-fire way to make people recoil from mistakes if you punish them, rather than allowing people to fail, learn from their mistakes, and improve.
5. Bureaucracy. Complex rules and processes where deviation isn’t allowed, is not a good place for exploration and learning.
6. Teams that are too large. Large teams require lots of management overhead. Managers don’t like mistakes. They like Microsoft Project, work breakdowns, and no surprises.
7. An “expert” mindset. Basically meaning a lack of knowledge about wicked problems, and a belief that everything can be predicted by experts, as long as they are good enough.
8. Actual, serious negative consequences, such as injury, pollution, death obviously also factor into the stigmatisation of mistakes.

How to give mistakes a good name

First off, the study indicated it’s not actually about making mistakes per se. Rather, it is about learning — because in complex design problems, learning is necessary, and failing along the way is inevitable. This means we have to create a culture of learning.

You might (rightly) say this is not news. We know about it from the education system where teachers go out of their way to create an environment in which it is safe to screw up and learn. We also know about it from scientific work, where failures are often more productive than successes in terms of the data generated. Finally, we already knew about it from design research.

What is perhaps new in this study are indications about specific things you could do to create a culture of learning:

1. Examples of glorious success. Everybody likes to be on the winning team, so show people how experimenting, failing, learning, and improving is the path to creating better products.
2. Make it familiar. Like anything else, it takes time and experience to be comfortable with new ways of working. Here it may also be beneficial to have an ambassador with lots of experience.
3. Make it part of the rulebook. By “officially approving” learning via mistakes, it becomes part of the organisation you are in. It’s in the rulebook, which means you are safe and within your rights when you fail and learn.
4. Make failure a goal. More than just making it okay to fail, make it an explicit goal to fail and learn — although it would probably be better phrased as “make learning a goal”, since that is the outcome we seek.
5. Incentives and rewards. We get attached to our ideas and don’t want them to fail, but rationally speaking, killing an idea or a project should be rewarded and celebrated. Because once it’s dead, we can stop wasting time and money on it.
6. Social support. In order to share openly and without fear, there needs to be a supportive culture. This means trust, respect, honesty, openness, empathy, always seeking to be constructive, and a high degree of personal freedom.
7. Constructive feedback sessions. Showing flawed work-in-progress and getting feedback from your team lets you draw on the combined knowledge of the entire group — and the best part is you end up looking good!

Thanks for reading. At the beginning of this study, I had the notion that mistakes might somehow become a “tool” for making progress in creative problem-solving.

Along the way, it soon became apparent that this idea was somewhat naive. Mistakes aren’t actually anything in and of themselves — whether something is a mistake or not is entirely based on reflections and evaluations of the consequences that become apparent through experiments. So what we need to do if we want to learn from failure, is create a culture and a mindset that promotes experimentation, reflection and learning, as well as constructive criticism and collaboration.

Further reading

• Why We Fail: Learning from experience design failures by Victor Lombardi
• Success through Failure — the Paradox of Design by Henry Petroski
• Being Wrong: Adventures in the Margin of Error by Kathryn Schultz
• Creativity, Inc. by Ed Catmull
• A More beautiful Question by Warren Berger