05 Jul 2006  Research & Ideas

The Accidental Innovator

Many important innovations are the byproduct of accidents—the key is to be prepared for the unexpected. Professor Robert D. Austin discusses his research and practical implications on the concept of accidental innovation. Key concepts include:

  • Innovation can't always be planned—accidents happen.
  • Be prepared to recognize serendipitous opportunity.
  • Understand the nature of breakthrough inventions in your industry and plan accordingly.


Companies spend many hundreds of billions of dollars on R&D each year, but the microwave oven was conceived from a melted candy bar, saccharin from an accidental chemical spill, and the Daguerre photo process via a shattered thermometer. Accidents happen—and we're all better off because they do.

In their recent working paper "Accident, Innovation, and Expectation in Innovation Process," authors Robert D. Austin and Lee Devin explore the concept of accidental innovation, how it works or doesn't, and how good accidents can be encouraged. Austin is an associate professor at Harvard Business School, while Devin is professor emeritus from Swarthmore College.

Sarah Jane Gilbert: Can you explain what accidental innovation is? What led to your interest in researching this concept?

Robert Austin: Historical accounts of many important discoveries and inventions tell of fortuitous accidents. A surprising number of important discoveries and inventions are associated with stories about spillage, breakage, and other manner of unintended action that led to valuable, though unexpected, outcomes.

Probably the most famous is Alexander Fleming's discovery of the antibiotic properties of penicillin. Fleming accidentally left a dish of Staphylococcus bacteria uncovered for a few days and returned to find the dish dotted with bacterial growth, except in one area where a patch of mold (Penicillium notatum) was growing. Fleming himself said of this event, "I did not ask for a spore of Penicillium notatum to drop on my culture plate... When I saw certain changes, I had not the slightest suspicion that I was at the beginning of something extraordinary.... That same mould might have dropped on [any one] of my culture plates, and there would have been no visible change to direct special attention to it." Similarly, Daguerre, who invented photography, made his breakthrough when he put an exposed plate into a cabinet in which a thermometer had earlier shattered; mercury vapors from the broken thermometer developed the photographic image unexpectedly.

We have to be careful of these stories, in part because they make such good stories. Some scholars are skeptical of them, but the sheer number of them is interesting. And many scientists, like Fleming, talk very explicitly about the role of accident in their work. Some even argue that the orderly way people sometimes describe processes of discovery and invention, of the progress of science, is nothing less than fraud.

Actually, though, I would not really label this "accidental innovation." The innovation itself can't really be said to be "accidental," even though it involves accident. It takes a considerable capability to see the value in an accident, and to build upon it to create even more value.

I became interested in this subject when I was interviewing artists about their creative processes and discovered that many of them seemed to rely on accidents to generate interesting and creative outcomes.

Q: How important is the role of accident in the creative process? Does this happen often?

A: These are very difficult questions to answer in any definitive way. We don't really attempt to be conclusive in our research, although we propose some ways of looking at the questions.

Part of the problem is that the word "accident" is rather imprecise; not all accidents are equally accidental. If you bump into your neighbor at a local grocery store, without planning to, you might call that an accident. But that would be less accidental and remarkable than meeting the same neighbor in a tattoo parlor in a country on the other side of the world (assuming you hadn't planned to meet there).

We propose a way of defining the "intensity" of an accident, by which we mean, roughly, how far outside intentions (and expectations) an outcome really is. If you are looking for a new ulcer drug and discover one in a surprising way, that's not quite as amazing as if you are looking for a new ulcer drug and discover instead a new artificial sweetener (this is how Aspartame, a.k.a. "NutraSweet," was invented). The former is a lower intensity accident.

Many artists seem to rely on accidents to generate interesting and creative outcomes.

How often accidents are involved in important innovations is also difficult to say. How do you define the time boundaries of an accident? What does "often" mean? Ernst Mach, a philosopher and scientist of note, argued that the most important inventions tend to be associated with accidents. Whether or not accidents happen "often," if the most important inventions or discoveries are associated with them, that's pretty interesting. These are all hard questions to answer, but we suggest in the paper that it is important to ask them and see if we can arrive at rough answers, even if we can't definitively answer them.

Q: What are some examples of innovations that were discovered by accident?

A: There are many, many examples. Here are just a few: anesthesia, cellophane, cholesterol lowering drugs, cornflakes, dynamite, the ice cream soda, Ivory soap, NutraSweet (and several other artificial sweeteners), nylon, penicillin, photography, rayon, PVC, smallpox vaccine, stainless steel, Teflon. The list is a lot longer, but this gives you an idea.

Q: Your research involved interviewing artists. Can you tell us why you chose this group? What did you find by talking to them?

A: I was interviewing artists as part of a larger study of the principles, processes, and practices that lead to reliable innovation. (I'm working on a book on reliable innovation.) I found that many artists are very upfront about the importance of accident in their processes. Some artists see accident as a good way to produce something that they would not have been able to think of in advance and that is, therefore, quite new—something they wouldn't have even thought to try to create. One artist showed me how one of his important pieces came out of some experiments with unusual tools; he wasn't trying to do a piece of work, he was just trying out the unfamiliar tools, and something interesting happened. Another, a potter, showed me how he would create beautiful pots and then, while they were drying, whack them with a stick. Sometimes they just broke, but other times he'd get an interesting shape that he'd never seen before. Then he'd make a whole series based on that new shape. He was trying to get outside of what we would call his "cone of expectations and intentions" to create something truly new.

Q: Is there a way innovators can encourage good accidents? In other words, is there anything we can control to foster this process?

A: Great question. Artists think they develop a talent for causing good accidents. Equally or perhaps even more important, they believe they cultivate an ability to notice the value in interesting accidents. This is a non-trivial capability. Pasteur called it the "prepared mind." There's an interesting analogy to evolutionary models of creativity here. In 1960, Donald Campbell proposed that we think of creativity as "Random variation + Selective Retention." That is, we need two processes, one to generate things we can't think of in advance, and another to figure out which of the things we generate are valuable and are worth keeping and building upon. In science, the arts, and other creative activities, the ability to know what to throw away and what to keep seems to arise from experience, from study, from command of fundamentals, and—interestingly—from being a bit skeptical of preset intentions and plans that commit you too firmly to the endpoints you can envision in advance. Knowing too clearly where you are going, focusing too hard on a predefined objective, can cause you to miss value that might lie in a different direction.

In business, there's a saying that goes "if you don't know where you're going, any map will do." You can almost always get managers to nod in agreement with this suggestion that you might as well not start something if you don't have its end objective well defined. Working without a clear definition of your objective is considered wasteful, inefficient. But if you are trying to get outside what you can anticipate and see in advance, if you are going after the truly new and valuable, this way of thinking can be a problem. This is one truth about innovation that artists seem to understand a lot better than managers.

Q: What can managers learn from your findings and apply to their businesses?

A: Probably the most important practical implication has to do with the design of innovation processes. If you think that most breakthrough inventions—blockbuster drugs, say—are arrived at through careful, scientific, deductive process, without accident being much involved, then you'd design one kind of innovation process. On the other hand, if you think there is a very significant overlap between accident and important innovations, you'd design the innovation process differently. You might want to design in some "accidents," and you'd want to nurture your capability for "selective retention"—your ability to know what to throw away and what to keep. There's a whole arsenal of ways we might do this, many of them that we find in the practices of artists, but to see what they all are, you'll have to stay tuned for the completion of this research.

About the author

Sarah Jane Gilbert is a content developer at Harvard Business School's Baker Library.