"Fail fast, learn faster." "Celebrate failure." "If you're not failing, you're not innovating." The entrepreneurial world is full of calls to embrace failure as a learning tool. This idea is well-meaning and, in theory, correct. Real innovation carries uncertainty, meaning many attempts won’t succeed. Learning from these failures drives progress. In my post on Pixar Animations, I noted that embracing failure is an important element of its entrepreneurial culture.
However, there's a big gap between saying failure is valuable and actually learning from it. Failure without reflection is just failure. Reflection without structure isn’t true learning. What sets apart those who learn from failure is not a better attitude. It's the use of tools that encourages examining what went wrong and why.
The Failure Spectrum: Not All Failures Are Created Equal
Before discussing how to learn from failure, let's clarify different types of failure. Amy Edmondson, a professor at Harvard Business School, developed a useful way to classify failure. It shows the difference between failures that lead to learning and those that cost resources.
Preventable failures occur due to ignoring known processes, a lack of focus, or insufficient skills. Examples include medication errors in hospitals or production defects from poor quality control. These failures bring no value. They highlight execution failures, and the lesson is simply to do better.
Complex failures happen in systems where many small issues combine unpredictably. For instance, a hospital patient may face problems due to a mix of conditions, medications, and timing. A supply chain issue can affect a business in unforeseen ways. These failures are more instructive than preventable ones. They reveal interactions and dependencies that weren’t understood before.
Intelligent failures arise from thoughtful experiments with uncertain outcomes. A startup might test a new product feature that doesn’t perform as expected. A research team might pursue a hypothesis that turns out wrong. These failures are worth celebrating. They lead to learning that wouldn’t happen otherwise.
Learning from failure starts with categorising it. Treating intelligent failures as preventable can lead to blame and missed opportunities to learn. Conversely, romanticising preventable failures as intelligent can excuse carelessness.
Pre-Mortems: Imagining Failure Before It Happens
A useful tool for learning from failure is the pre-mortem, created by psychologist Gary Klein. It flips the traditional post-mortem approach. The team imagines that a project has failed and works backwards to identify possible causes.
The process is simple. At the start of a project, the team envisions it failing in six months. Each member writes down their reasons for the failure. These lists are then shared and discussed.
This method counters cognitive biases that can distort planning. Optimism bias may lead teams to underestimate risks. Groupthink can silence concerns, making people feel disloyal. The pre-mortem allows everyone to voice potential issues constructively.
A good pre-mortem produces a list of risks to address, assumptions to test, and dependencies to monitor. Some of these risks may be known but not fully articulated. Others may be new insights that emerge in a safe environment.
While the pre-mortem doesn’t prevent all failures, it shifts learning from after-the-fact to before they happen. This can lower the cost of making adjustments.
Post-Mortems: Examining Failure After the Fact
When a failure occurs, the quality of the post-mortem shapes whether it leads to learning or blame. A post-mortem reviews what happened, why it happened, and what can be learned. If done poorly, it becomes a way to evade responsibility. If done well, it becomes a valuable learning tool.
The key to an effective post-mortem is separating what happened from who is at fault. Blame destroys honesty. If people feel exposed during a post-mortem, they may be dishonest. They will protect themselves and withhold information.
A structured post-mortem follows a clear sequence. First, establish a timeline of events from various perspectives. This helps gather facts everyone can agree on. Next, identify the decisions that led to the outcome. Then, examine the assumptions that influenced those decisions. Finally, consider what could have been done differently.
The result should be specific, actionable changes. Without these actions, a post-mortem is merely a conversation. With them, it becomes a means to improve organisational capability.
The Five Whys: Getting to Root Causes
A simpler yet powerful technique is the Five Whys, from the Toyota Production System. When a problem arises, the instinct is to fix the immediate cause. The Five Whys encourages asking "why" five times to dig deeper into the root cause.
For example, a startup's product launch fails. Why? The customer acquisition cost was too high. Why? The advertising campaign's conversion rate was lower than expected. Why? The messaging didn’t resonate with the target audience. Why? The team hadn’t validated customer assumptions before launching. Why? There was no structured process for customer research.
The first "why" identifies a symptom, while the fifth reveals a deeper issue. Addressing this can prevent similar failures. Without repeatedly asking why, fixes tend to be superficial.
The Five Whys is straightforward and requires no special training or extensive time. However, it requires genuine curiosity and the willingness to explore uncomfortable truths. Without that, it may stop at a convenient answer.
Failure Taxonomies and Pattern Recognition
Individual failures are data points, but patterns across them provide insights. Organisations that learn effectively from failure don’t just review each incident separately. They track, categorise, and search for patterns that hint at systemic issues.
A failure taxonomy is a way of classifying failures. This makes patterns visible. What matters is consistently applying this method and reviewing the data.
Often, surprising insights emerge. A team might find that failures cluster at a specific process stage. They may discover that failures blamed on external factors actually stem from untested assumptions.
Recognising patterns in failures helps scale organisational learning. A single post-mortem improves responses to one type of issue. A taxonomy of failures, combined with the willingness to act on patterns, enhances the system as a whole.
Building a Learning Discipline
These techniques are tools, not solutions. Their effectiveness depends on the context. In a culture that punishes mistakes, people won’t share openly. In a learning-focused organisation, these methods can drive continuous improvement.
What matters is leadership's commitment to viewing failure as data, not inadequacy. A culture that encourages examining failures without defensiveness is key. This commitment is tough to maintain, especially under pressure to deliver results.
However, the cost of failing without learning is high. Accumulating failures without reflection, repeating mistakes, and ignoring capability gaps are all costly. Honest reflection is less costly than allowing issues to persist.
Summary
The entrepreneurial talk about failure often sounds inspiring, but it can be shallow. The real work of learning from failure is disciplined, structured, and sometimes uncomfortable. Organisations and individuals who engage in this process build a capability over time: to fail less expensively, recover faster, and gain insights to improve future attempts. That capability, more than merely tolerating failure, is what distinguishes those who learn from experience and those who repeat it.