Failure is information. Whether institutions extract that information depends on the conditions they create for honest post-failure analysis.
What Learning From Failure Requires
Learning from failure — genuinely extracting the diagnostic information that failures contain and converting it into improved institutional design — requires a set of conditions that most institutions do not naturally create. Safety for reporting: the people with the most direct knowledge of how failures occur must be able to report that knowledge without facing career consequences for having been present or associated with the failure. Honesty about causes: the analysis of failure causes must reach the structural and systemic conditions that determined the failure's possibility, not stop at the individual decisions that most directly preceded it. And commitment to change: the lessons extracted must be connected to institutional changes that actually address the identified causes, rather than to symbolic responses that satisfy the demand for visible action without changing the conditions that produced the failure.
Each of these conditions is in tension with the natural institutional dynamics that failure produces. Failure triggers blame attribution, which makes safety for reporting difficult to maintain. Failure attribution follows institutional power dynamics, which directs it away from structural causes and toward individual ones. And the political pressure for visible response favours symbolic changes over structural ones, because symbolic changes can be implemented quickly while structural ones require the time-consuming and politically difficult work of redesigning systems and processes.
The Learning Organisation's Actual Practices
Institutions that genuinely learn from failure — that have produced the sustained improvement in performance that learning is supposed to enable — do so through specific practices that overcome each of these natural barriers. They separate the accountability process from the learning process, recognising that combining them produces accounts that are defensive rather than diagnostic. They require post-failure analysis to reach structural causes, using root cause analysis frameworks that by design go beyond individual decisions to the systemic conditions that shaped those decisions. And they measure the completion of institutional changes against identified failure causes, rather than treating the learning process as complete when the after-action review is written.
Failure is the most expensive education available. The institution that pays for it but does not extract the learning is paying full price for a course it did not attend. Learning from failure is not a natural institutional behaviour — it is a deliberately designed process that must be built and maintained against the institutional dynamics that work against it.
Discussion