When Improvising Is Better Than Planning: Learning from Mistakes at Work

Planning and improvisation solve different problems

Planning is valuable when work is repeatable, dependencies are known and errors carry serious consequences. Nobody wants a finance close, a security control or a production migration to depend on spontaneous inspiration.

But plans are built from what a team knows at a particular moment. In transformation work, important information often arrives only after delivery begins: an exception appears in real data, a user reacts differently from the expected journey or two systems interpret the same rule in different ways.

This is where improvisation can be better than more planning. Good improvisation is not doing whatever feels interesting. It is acting within a clear purpose and clear constraints, paying attention to what is actually happening and making the next useful move with the information available now.

A plan is a commitment to an outcome—not a promise to ignore better information about how to reach it.

Four signals that the team needs to improvise

The answer is not to improvise everything. The practical question is whether the current situation still rewards detailed prediction or now requires short cycles of action and learning.

• The assumptions are changing faster than the planning or approval cycle.
• No single function has the complete picture, so the next step must be built from several partial perspectives.
• A small, reversible test can create better evidence than another meeting about what might happen.
• The cost of waiting for certainty is higher than the cost of a contained experiment.

What this looks like in practice

Consider a system migration. A detailed plan is necessary, but a thin end-to-end test can reveal more than weeks of reviewing assumptions in separate functions. One realistic transaction moving through ordering, billing, payment and reporting may expose an exception that no individual team could see from its own component.

Or consider a customer escalation. The company needs clear roles, authority limits and compliance rules. Inside those guardrails, the team may still need to improvise: recognise the customer’s actual concern, combine information from several systems and create a response that no standard script anticipated.

The same applies in a workshop or leadership meeting. A facilitator needs an objective and a structure. If the room becomes stuck on a real disagreement, rigidly completing every planned activity can be less useful than changing the sequence, making the tension discussable and helping the group work with what has appeared.

In each example, planning provides the container. Improvisation makes the container responsive to reality.

Cultivate learning from mistakes—not more mistakes

The phrase ‘mistake culture’ can sound as if a company should lower its standards. That is not the goal. A useful culture distinguishes between errors that should be prevented, surprises that emerge from complexity and intelligent experiments designed to test an uncertain assumption.

Preventable or repeated errors need correction, stronger controls and accountability. Surprises in complex work need early visibility and shared sense-making. Intelligent experiments need boundaries, a clear question and a result that the team is willing to learn from—even when the original idea does not work.

The cultural advantage is speed of truth. When people can report a weak signal, a near miss or an incorrect assumption without first protecting their reputation, the organisation receives information while it is still cheap enough to use.

• Surface mistakes early, before people quietly compensate for them and hide the real process.
• Separate the person from the event: examine what made the decision reasonable at the time and what information was missing.
• Reward useful disclosure and learning, while still addressing negligence or repeated avoidable errors.
• Make the lesson visible by changing a checklist, assumption, test, role or decision rule.

Five routines that make the culture real

Telling people to ‘fail fast’ is not enough. If the same leaders punish the first visible failure, teams quickly learn that the slogan is decorative. The culture becomes credible through small operating routines.

• Name assumptions in the plan: write down what must be true, how it will be tested and when the team will look again.
• Define guardrails: be explicit about safety, legal, financial and customer boundaries that cannot be treated as experiments.
• Use reversible steps: test at a scale where the company can learn without creating an unacceptable consequence.
• Run a short learning review: What did we expect? What happened? What did we learn? What will we change now?
• Close the loop: an error that produces no change is only a story; an error that changes the system becomes organisational learning.

Psychological safety and high standards belong together

A culture that learns from mistakes is not a culture without challenge. Colleagues should still ask whether a risk was understood, a control was followed or the same failure has happened before. The difference is that the purpose of the challenge is to improve the work, not to find the fastest person to blame.

Leaders make this visible when they say which assumption they got wrong, invite inconvenient information and remain curious long enough to understand the event. They also protect the boundaries where experimentation is inappropriate.

The strongest teams do not choose between planning and improvisation. They plan enough to coordinate, improvise enough to respond and learn enough not to repeat the same mistake.