Learning from failure or success

In a round up following KM Australia, back in August, Shawn Callahan has challenged the notion that we learn best from failure. I think he has a point — the important thing is learning, not failure.

Harris Hawk missing the quarry

Here’s Shawn’s critique.

During the conference I heard a some speakers recount the meme, “we learn best from failure.” I’m not sure this is entirely true. Anecdotally I remember distantly when I read about the Ritz Carlton approach to conveying values using stories and I’m now delivering a similar approach to a client on the topic of innovation. Here I’ve learned from a good practice. As Bob Dickman once told me, “you remember what you feel.” I can imagine memory being a key first step to learning. And some research shows it’s more complex than just learning from failure. Take this example. The researchers take two groups who have never done ten pin bowling and get them bowling for a couple of hours. Then one group is taken aside and coached on what they were doing wrong and how they could improve. The other group merely watches an edited video of what they were doing right. The second group did better than the first. However there was no difference with experienced groups.

I wish I could access the linked study — Shawn’s summary and the abstract sound very interesting. Here’s the abstract.

On the basis of laboratory research on self-regulation, it was hypothesized that positive self-monitoring, more than negative self-monitoring or comparison and control procedures, would improve the bowling averages of unskilled league bowlers (N =60). Conversely, negative self-monitoring was expected to produce the best outcome for relatively skillful league bowlers (N =67). In partial support of these hypotheses, positive self-monitors significantly improved their bowling averages from the 90-game baseline to the 9- to 15-game postintervention assessment (X improvement = 11 pins) more than all other groups of low-skilled bowlers; higher skilled bowlers’ groups did not change differentially. In conjunction with other findings in cognitive behavior therapy and sports psychology, the implications of these results for delineating the circumstances under which positive self-monitoring facilitates self-regulation are discussed.

Based on these summaries, I would draw a slightly different conclusion from Shawn’s. I think there is a difference between learning as a novice and learning when experienced. Similarly, the things that we learn range from the simple to the complex. (Has anyone applied the Cynefin framework to learning processes? My instinct suggests that learning must run out when we get to the chaotic or disordered domains. I think we can only learn when there is a possibility of repeatability, which is clearly the case in the simple and complicated domains, and may be a factor in moving situations from the complex to one of the other domains.)

The example Dave Snowden gives of learning from failure is actually a distinction between learning from being told and learning by experience.

Tolerated failure imprints learning better than success. When my young son burnt his finger on a match he learnt more about the dangers of fire than any amount of parental instruction cold provide. All human cultures have developed forms that allow stories of failure to spread without attribution of blame. Avoidance of failure has greater evolutionary advantage than imitation of success. It follows that attempting to impose best practice systems is flying in the face of over a hundred thousand years of evolution that says it is a bad thing.

In the burned finder scenario, success (not touching a burning match) is equivalent to lack of experience. Clearly learning from a lack of experience will be less effective than learning from (even a painful) experience. By contrast, the bowling example provides people with a new experience (bowling) and then gives them an opportunity to contemplate their performance (which was almost certainly poor). However, whatever the state of their performance, it is clear what the object of the activity is and therefore ‘success’ can be easily defined — ensure that this heavy ball leaves your hand in such a way that it knocks down as many pins as possible by the time it reaches the far end of the lane. As the natural tendency of learners at early stages in the learning process is to concentrate on the negative aspects of their performance (I can’t throw the ball hard enough to get to the end of the lane, or it keeps going in the gutter), it is understandable that a learning strategy which focuses on success could have better results than one that merely explains why the bad things happen.

In the bowling experiment, no difference was found between the negative and positive approaches when experienced bowlers were studied. All this suggests to me is that we need more work in this area, especially considering learning in the complicated or complex domains. Even for experienced bowlers, the set of variables that affect the passage of a bowling ball from one end of the lane to the other is a predictable one. There is not just one cause and effect, but the laws of physics dictate that the relationships between all the causes should have predictable outcomes. By contrast, much of what interests us with regard to knowledge and learning in organisational environments does not depend on simple causal relationships.

In those complicated or complex organisational situations, I think we can learn more from our own failures than other people’s successes (which I think is the point that Dave Snowden is making). I think Shawn is also right to suggest that we can learn from our own successes too. However, that can only be the case if we take the time to analyse exactly what was the cause of the success. So we need a commitment to learning (which brings us back to deliberate practice, amongst other things) and we need the insight into our actions and activities that allows us to analyse them effectively. I think the will to learn is often present, but insight is often missing when we consider successful initiatives, possibly because the greater distance between cause and effect means that we cannot be confident that success is a product of any given cause. On the other hand, it is usually easier to identify causes of failure, and the process of failure also provides an incentive to work out what went wrong.

As for the quality of the lessons learned from failure or success, I am doubtful that any firm conclusion could be drawn that as a general rule we learn better from failure or from success. However, as we become more experienced and when we deal with fewer simple situations, we will inevitably learn more from failure than success — we will have more experience of failure than success, and other people’s successes are of limited or no value. So, although we can learn from our successes, my guess is that more of our learning flows from failure.

It feels like there is more research to do into these questions.

2 thoughts on “Learning from failure or success”

Comments are closed.