When Failure Leads to Innovation, and When It Doesn’t (Part One)
Successful innovation requires successful
experimentation, and successful experimentation requires eagerness to learn
from failure. This has become a cliché because it holds true, time and time
again.
The great experimenter Thomas Edison is famous
for comments such as, “I have not failed. I’ve just found 10,000 ways that
won’t work.” The design firm IDEO keeps up that spirit with their now-famous
phrase, “Fail often in order to succeed earlier.”
Successful Failures
But let’s distinguish between two different kinds
of failure: instructive failure and terminal failure. Apple’s
Newton (PDA) and Lisa computer were of the instructive variety: they were
failures to grow on, not failures to stop on. Circuit City
Early on, Hewlett Packard exploited the power of
instructive failure. According to Peter Sims, “Hewlett Packard cofounder Bill
Hewlett said HP needed to make 100 small bets on products to identify six that
could be breakthroughs. So, little bets are for learning about problems and
opportunities while big bets are for capitalizing upon them once they’ve been
identified.” Sims’ “small bets” are what we’d call experiments: exposure to
non-fatal failure that can teach you something.
The entrepreneur’s challenge can almost entirely
be summed up as ensuring that the learn rate exceeds the burn rate:
those who don’t learn fast enough go under.
Deliberate, inquisitive exposure to failure is an
experiment. And a clever experiment is like a clever investment: your downside
(risk) is manageable, and your upside (lesson) is spectacular. Of course, there
is a time to bet the farm, but that’s after you’ve learned which farm
to bet on.
Failing at Failure
Some people fail at failing: they fail without
gaining anything. What’s the difference between failure that’s experimentation
and failure that just failure? Maybe this: if you make a non-fatal mistake and
learn from it, then it was “experimentation.” But if you make a mistake and deflect
any lessons, then it was simply a failure. Lessons learned lead to
innovation; lessons flunked, as in school, tend to be repeated.
Here are some ways to flunk at failing:
- Finger pointing. When the question is, “Who screwed up?” instead of “What did we learn?” then the only thing that’s learned is how to keep your head down.
- Reasons, stories, and excuses. When an organization’s lousy results allegedly stem from “the poor economy,” or “difficulty finding talent,” or “tough competition,” then nothing is learned or even speculated about what the organization can do better. Part of Warren Buffet’s initial fame stemmed from his annual reports in which he gave blunt assessments of what he and Berkshire Hathaway could have done better. It showed shareholders that lessons were not wasted on him.
- Unclear success. Like a scientist
with an untestable hypothesis, a leader with an unclear goal can spend a
lot of time and money without learning much. For example, when any given
organization consolidates two departments to “capture synergies,” what
does “synergies” mean? Lower costs? Faster product development? Quicker
response? What? Without some sense of the measurable goal, it
will be impossible to get the
- Activity-based success. Of course, you can be clear about your success, but define it as an activity rather than as a result. In which case, failure and learning are equally unlikely. Again, no hypothesis is tested. For example, government officials often declare success after they’ve added programs or increased spending. That’s it! Nothing about goals set, goals met, or lessons learned. Costs go up, but learning stays flat.
All these problems function as organizational learning
disabilities: dysfunctions that block learning and therefore block
innovation. In the next post we’ll suggest some cures.
Meanwhile, what have you done that works? How
have you overcome your own organization’s learning disabilities? How can you
create a failure friendly environment, where team members feel comfortable with
experimenting and learning from their failures?
