In research
published in Industrial and Corporate Change, I examined this issue with PeterW. Glynn and Hayagreeva Rao. We did not embark on a study of failures in famous
firms – it has been done so many times before. Instead, we modeled different ways
of designing organizational structures to see how well they performed. The idea
behind the model is that, in addition to whatever else it does, any firm faces
a sequence of incoming problems that it needs to spend time solving, and it
matters how quickly these problems are solved. This is an easy conceptual model
that takes some advanced math to solve; and looking at it we got some surprises.
First, problem
solution times can vary tremendously, especially when the firm is busy. This is
important because efficiency comes from not having too many resources at hand,
which means that efficient firms are busy. It is one reason that success and failure
are so closely linked.
Second,
managers can make the problem worse. Adding any level of approval of a solution
multiplies the time it takes to finalize, which is very problematic when the
solution time is already long. Large span of control (many subordinates), as we
see in many lean organizations, makes this problem worse. It is another reason
that success and failure are so closely linked.
Third, teams
can solve the problem. If they communicate well, a team will much more rarely
have very long solution times than an individual, even if the individual has
exactly the same expertise as all team members combined. This sounds
surprising, so I should be quick to add that communication slows down teams
significantly.
Fourth, smart
rule-breaking can help solve the problem. Specifically, a worker who selectively
solves some problems without forwarding the to the manager for approval can reduce
the number of problems that take a long time to solve. Similarly, a worker who
simply tosses out some problems instead of solving them can radically reduce
the number of problems that take a long time to solve. Naturally, rule-breaking
is only good for the firm if the workers can identify problems that don’t need
approval or don’t need a solution.
These are simple facts useful for putting together an organizational structure
and its processes. A designer who knows them can make the right choices
between efficiency today and resilience tomorrow, because they dictate how much
slower reactions to surprises are when a structure is set up for efficiency. Perhaps
the main lesson is exactly in understanding that this is a trade-off. Efficiency
is great until the day it isn’t, and it instead becomes the source of failure.
