The foundational paradox

Organisations accumulate experience continuously. The average company with a decade of operation has encountered most of the problems it will ever face in some recognisable form. And yet the research literature is unambiguous: repeated mistakes are not exceptional failures of particularly dysfunctional organisations. They are the default output of the mechanisms by which organisational memory actually works.

Levitt and March’s 1988 Annual Review of Sociology synthesis remains the foundational account. Their central argument is that organisations learn by encoding inferences from history into routines that guide behaviour — but the routines that encode lessons are the same routines that perpetuate mistakes. What gets encoded is the lesson that was drawn at the time, shaped by the cognitive, political, and psychological conditions of the moment. The routine persists until sufficient disconfirming experience overcomes its inertia, which may take years — or may never happen if the conditions that would provide that disconfirming experience are themselves avoided as a consequence of the routine.

Three mechanisms of organisational forgetting

MIT Sloan Management Review’s treatment of organisational forgetting identifies the primary pathways through which hard-won lessons disappear. Personnel turnover is the most straightforward: in fast-growing organisations where headcount doubles every eighteen months, institutional knowledge evaporates faster than it can be codified. A startup that experienced a near-catastrophic production incident two years ago and has since tripled its engineering team has effectively lost the institutional memory of that event. The people who learned from it are now a minority; the new majority has only whatever formal documentation exists, which rarely captures what the experience actually felt like or what the early warning signals looked like.

Political suppression is more insidious. When psychological safety is absent, postmortems become performances in which people share what is politically safe rather than what is operationally true. The lessons captured are sanitised versions of reality. The organisation’s memory becomes corrupted — accurate in its formal record, false in its actual content. Over time this produces a specific and recognisable inefficiency: problems that feel new but mirror what has already been done, mistakes that appear situational but are systemic, and a growing library of postmortems that bears decreasing resemblance to operational reality.

Superstitious learning is the third mechanism, identified by March and Olsen as among the most destructive. The organisation takes action, observes an outcome, and encodes a lesson — but the causal link is misattributed. A product launch that failed due to poor market timing is attributed to poor product-market fit; the encoded lesson is that more customer research is needed. The next launch has better customer research and worse market timing. The mistake repeats because the encoded lesson addressed the wrong cause with genuine conviction.

NASA’s Challenger and Columbia: the same mistake seventeen years apart

The most consequential documented case of repeated organisational mistake is NASA’s shuttle programme. The Columbia Accident Investigation Board identified in 2003 that the same organisational culture that normalised the deviance producing Challenger in 1986 remained substantially unchanged seventeen years later. The organisational memory of what went wrong had not been retained in the routines, norms, and decision-making processes that governed the second disaster.

This is the oscillation pattern that Organisation Science research identifies: genuine learning in the immediate aftermath of a crisis, followed by gradual reversion as budgetary pressure, schedule pressure, and personnel turnover erode the changes that learning had produced. The lesson was learned, encoded, and then forgotten — not through any single act of deliberate forgetting but through accumulated competing demands, the replacement of personnel who had lived through the original event, and the basic difficulty of maintaining institutional memory without active, repeated investment.

The myopia of success

Research consistently shows that organisations learn less from failure than from success, despite failure being the more informative signal. Successful routines crowd out the exploration that would learn from failure. Organisations that are succeeding are least likely to make the changes that would prevent the next failure, because the success validates existing routines rather than questioning them.

If the patterns described here — the sense that your organisation is making the same mistakes it has made before, that lessons are not sticking, that the same conversations keep happening — are significantly affecting your wellbeing, speaking with a psychologist is the right next step. UK: Samaritans (116 123, free, 24/7). Mind (0300 123 3393). BACP: bacp.co.uk/search/Therapists. International: internationaltherapistdirectory.com.

The competency trap

Madsen and Desai’s study of the global orbital launch vehicle industry found that success led organisations to exploit existing routines more intensely while failure led them to explore alternatives — but the learning from failure was slower and more vulnerable to being lost. The industry kept experiencing similar failure modes across organisations because each organisation’s learning from its failures was insufficient to overcome the exploitation pressure that success elsewhere maintained.

The competency trap names the specific mechanism: the more an organisation has refined a particular approach, the more its routines, skills, and cognitive frameworks are calibrated to it — and the less capacity remains to recognise that the approach is becoming obsolete. Familiarity is mistaken for reliability. The organisation keeps making the same mistake because becoming competent at the old approach reduced its ability to notice that the approach was now wrong.

What actually produces durable learning

Three conditions the research identifies as necessary. First, psychological safety in postmortems — without it, the lessons captured are sanitised and the memory becomes corrupted before it is encoded. Second, explicit codification that survives personnel change — lessons held only in individuals’ heads are not organisational memory; they are personal memory that leaves when the person does. Third, regular deliberate reactivation — the evidence that institutional memory erodes under competitive and budgetary pressure means that the learning investment must be recurring rather than one-time.

The organisation that runs an excellent postmortem once has not produced durable learning. It has produced a document. Durable learning requires the postmortem’s lessons to be incorporated into hiring, training, process design, and the frameworks through which future situations are recognised and interpreted.

A book worth reading alongside this

The Fifth Discipline by Peter Senge is the most widely read applied treatment of how organisations can build the learning infrastructure needed to avoid the repeated mistake problem. Senge’s five disciplines — systems thinking, personal mastery, mental models, shared vision, and team learning — are the practitioner answer to the Levitt and March diagnosis. His argument that organisations systematically fail to learn because their members cannot see the systemic structures producing the patterns they experience is the most accessible available account of why individual intelligence and organisational intelligence are not the same thing, and what the gap between them costs.

Have questions about this article?

If any part of this article raised questions you want to explore further, courbot.co is built for exactly that. It is courben.co’s AI assistant, designed around the psychology of entrepreneurship. Ask it anything from this article.

If the organisational dynamics described here are significantly affecting your wellbeing, speaking with a psychologist is the right next step. UK: Samaritans (116 123, free, 24/7). Mind (0300 123 3393). BACP: bacp.co.uk/search/Therapists. Crisis Text Line — text HOME to 741741 (US, UK, Canada, Ireland). International: internationaltherapistdirectory.com.

This article is for educational and informational purposes only. Sources: Levitt, B. & March, J.G. (1988), Annual Review of Sociology, 14, 319–340. March, J.G. & Olsen, J.P. (1975), European Journal of Political Research. Madsen, P.M. & Desai, V. (2010), Academy of Management Journal, 53, 451–476.