There is a particular kind of conversation that happens after a project goes badly wrong. Everyone sits in the room, the failure is acknowledged, and the discussion begins. Within minutes, something curious happens: people start constructing a story in which the outcome was obvious all along. "I knew that vendor was unreliable." "We all saw the warning signs." "The decision to cut testing time was clearly the turning point."

None of these people are lying. They genuinely believe what they are saying. But what they are doing is retrospective sensemaking — one of the most powerful and least understood processes in human cognition.

The phenomenon cuts across every domain of human activity. Doctors reviewing a diagnosis after learning the true illness remember their uncertainty differently than they reported it at the time. Investors recall their pre-crash portfolio confidence with distorted clarity. Military analysts reviewing strategic decisions after wars explain the "obvious" errors that led to defeats. What all of these have in common is the invisible, automatic reconstruction of the past to fit what is now known — a reconstruction so seamless that it feels like accurate memory rather than creative rewriting.

Understanding retrospective sensemaking matters not because it is a curiosity but because it is the mechanism through which humans learn — or fail to learn — from experience. Organizations that do not understand it conduct post-mortems that produce the wrong lessons. Individuals who do not understand it misread their own development and make worse decisions about their futures. The machinery of meaning-making runs continuously and largely below awareness, shaping every interpretation of past experience we ever form.

Karl Weick and the Theory of Sensemaking

The formal study of sensemaking in organizational contexts begins with Karl Weick, a social psychologist at the University of Michigan whose 1995 book Sensemaking in Organizations remains foundational. Weick argued that the traditional model of organizations — rational actors gathering information, analyzing it, and then deciding — gets the sequence backward.

In practice, Weick observed, people act first and understand afterward. We respond to ambiguous situations with available behaviors, and only by observing our own actions do we come to know what we were doing, what we believed, and what we intended.

"How can I know what I think until I see what I say?" — Karl Weick, paraphrasing E. M. Forster to describe the retrospective nature of sensemaking

This is the core claim: meaning is made backward. The act of making sense of experience is fundamentally retrospective — we look at what happened and construct a coherent account of why.

Weick's insight drew on earlier theoretical work in social psychology and organizational theory, but it was his systematic articulation of sensemaking as a process — rather than a content — that changed how researchers and practitioners thought about organizational cognition. Weick was not claiming that people are irrational or that their retrospective accounts are necessarily false. He was claiming something more subtle: that the construction of meaning is itself a process with identifiable properties, and that understanding those properties reveals things about organizational life that rationalist models of decision-making miss entirely.

The Seven Properties of Sensemaking

Weick identified seven properties that distinguish sensemaking from information processing or rational decision-making:

  1. Grounded in identity construction — we make sense of events in terms of who we understand ourselves to be
  2. Retrospective — meaning is constructed by looking backward at enacted experience
  3. Enactive of sensible environments — our actions partly create the environment we then interpret
  4. Social — sensemaking is inherently shared and collaborative
  5. Ongoing — it is a continuous process, not a discrete event
  6. Focused on and by extracted cues — we use small clues as anchors for larger interpretations
  7. Driven by plausibility rather than accuracy — we seek a good-enough story, not a perfect one

The seventh property is particularly important: sensemaking prioritizes coherence over truth. We are not running experiments to determine what actually happened. We are constructing a narrative that holds together, that fits our identity and our social environment, and that allows us to move forward.

Weick and the Mann Gulch Disaster

One of Weick's most influential applications of sensemaking theory was his analysis of the 1949 Mann Gulch forest fire disaster in Montana, in which thirteen smokejumpers died when a fire overran their position. Weick's 1993 analysis in Administrative Science Quarterly used the disaster as a case study in the collapse of sensemaking under extreme time pressure and threat. The central question was: why did the crew fail to respond effectively to options that, in retrospect, seem obvious?

Weick's answer was that the crew's collective sensemaking structure collapsed when the fire's behavior exceeded their mental models. The crew foreman, Wagner Dodge, invented an escape fire on the spot — a technique of burning away the dry grass immediately around you to create a fireproof refuge. But none of the crew understood what he was doing in time to follow him. They had no conceptual framework for the option being offered and could not incorporate it into their rapidly dissolving understanding of what was happening.

The lesson for organizations: maintaining sensemaking capacity under pressure — through experience, rehearsal, and flexible mental models — matters as much as possessing the right information. Sensemaking failures are not failures of intelligence but failures of the frameworks through which experience is interpreted.

Memory Is Reconstruction, Not Playback

Retrospective sensemaking does not operate on an accurate record of events. It operates on memory — and memory is reconstructive.

This distinction, established over decades of research by psychologists Elizabeth Loftus, Frederic Bartlett, and others, is critical. When you remember an event, you do not replay a stored recording. You reconstruct the event from fragments, filling gaps with inference, expectation, and subsequent knowledge.

Frederic Bartlett's foundational 1932 study, Remembering: A Study in Experimental and Social Psychology, established this with careful experimental precision. Bartlett had English participants read and recall a Native American folk tale, "The War of the Ghosts," at increasing intervals after the original reading. What he found was not degradation of a veridical record but systematic transformation: participants unconsciously filled in gaps, rationalized unfamiliar elements with familiar ones, and progressively assimilated the story to their own cultural schemas. The transformation was not random — it was coherent and consistent with each participant's framework for understanding events. Memory, Bartlett concluded, is fundamentally reconstructive, not reproductive.

The Misinformation Effect

Loftus's classic studies in the 1970s demonstrated that memories can be altered by information received after the original event. In one famous experiment, people who watched a film of a car accident were asked either "How fast were the cars going when they smashed into each other?" or "How fast were the cars going when they hit each other?" Those who heard "smashed" reported higher speeds and were significantly more likely to remember — falsely — seeing broken glass in the film (Loftus & Palmer, 1974).

The word "smashed" was not present when the event was observed. It was introduced afterward, in the question, and it retroactively altered the memory.

This is memory reconsolidation — each time a memory is retrieved, it becomes temporarily malleable and is re-stored in a slightly modified form. Our memories are literally rewritten every time we access them. The practical implications are significant: repeated recall of an event is not simply reinforcing a static record. It is re-creating the record, and each re-creation is subject to influence from the current context, current knowledge, and current emotional state.

What We Believe About Memory What Research Shows
Memories are stored like files and retrieved accurately Memories are reconstructed each time from fragments
Repeated recall strengthens memory accuracy Repeated recall can introduce consistent distortions
Confident recall indicates accurate recall Confidence and accuracy are poorly correlated
The passage of time degrades memories uniformly Later information can actively overwrite earlier memories
We remember the "gist" correctly even if details fade Gist-level memories are also subject to narrative shaping
Trauma creates especially vivid, accurate memories Trauma memories are reconstructive and subject to distortion
Our memories of our own emotional states are reliable Emotional state memories are strongly distorted by current mood

The Hindsight Bias as Memory Distortion

The hindsight bias — "I knew it all along" — is now understood as partly a memory distortion phenomenon rather than simply a judgment error. Baruch Fischhoff's foundational 1975 research showed that people who learned the outcome of a historical event consistently overestimated the probability they would have assigned to that outcome had they not known it. Crucially, they also could not accurately remember their original uncertainty — they believed they had been more certain than they were.

This finding has been replicated in clinical contexts with striking implications. Arkes and colleagues (1981) showed that physicians who were told the actual diagnosis of a patient case revised their recall of their own diagnostic confidence upward. Doctors who had actually been quite uncertain, once they knew the answer, remembered themselves as having been more confident — and were therefore less able to learn from cases in which their diagnostic process had been uncertain.

The physician-learning problem is general. If we cannot accurately recall how uncertain we were at the time of a decision, we cannot accurately assess the quality of our decision-making process. We can only assess whether the outcome was good, which conflates the quality of a decision with the quality of a result — a confusion that makes genuine learning from experience very difficult.

Post-Hoc Rationalization: The Brain's Story Machine

The most immediate form of retrospective sensemaking is post-hoc rationalization — generating rational explanations for decisions, beliefs, or behaviors that were actually produced by non-rational processes.

Gazzaniga's Interpreter

Neuroscientist Michael Gazzaniga conducted a series of experiments in the 1970s and 1980s on "split-brain" patients — individuals whose corpus callosum (the bundle of fibers connecting the two brain hemispheres) had been severed to treat epilepsy. This unusual condition allowed Gazzaniga to communicate separately with each hemisphere.

In one famous experimental setup, Gazzaniga instructed the right hemisphere (via an image shown only to the left visual field) to perform an action — say, to stand up or to laugh. The patient would perform the action, but their left hemisphere (responsible for language) would have no idea why. When asked why they stood up or laughed, patients consistently generated explanations: "I wanted to stretch," "I thought of something funny."

Gazzaniga called the left hemisphere the interpreter — a module that monitors behavior produced elsewhere in the brain and generates plausible explanations for it, creating a seamless narrative of rational self-directed action (Gazzaniga, 1985). This research suggests that the experience of making conscious, rational decisions may often be a post-hoc interpretation of processes that had already concluded. The sense of "deciding" may be the story we tell about a process that had already happened.

This is not a minor point. It challenges the intuitive model of the self as a unified, rational agent whose conscious intentions precede and cause actions. The Gazzaniga interpreter model suggests that conscious intentional reasoning is often a narrator that runs after the fact, constructing stories that maintain the fiction of conscious agency. Whether this is universally true is debated, but the evidence from split-brain patients, social priming research, and unconscious decision studies all converge on a picture in which conscious rationalization is more central to the experience of decision-making than to its actual causation.

Haidt's Moral Dumbfounding

Social psychologist Jonathan Haidt extended this analysis to moral judgments. In studies of "moral dumbfounding," Haidt presented people with scenarios that triggered immediate moral disgust — for instance, a story about consensual incest between adult siblings who take careful precautions and never tell anyone.

Most people immediately judged this to be wrong. When asked why, they would generate reasons: "It could cause psychological harm." When told those reasons were stipulated to be false in the scenario, they would generate new reasons. When those were undermined, many participants reached a dead end — they could not articulate a reason, but maintained the judgment. They were "morally dumbfounded."

Haidt's interpretation: moral intuitions come first, and moral reasoning is constructed afterward to justify them (Haidt, 2001). The reasoning is not the cause of the judgment; it is the post-hoc narrative we construct to explain a judgment already made. This social intuitionist model of moral psychology has been extensively debated and refined, but the basic finding — that articulated moral reasons often track rather than produce moral intuitions — has been replicated across many paradigms.

The implication for everyday decision-making is uncomfortable. Many of the reasons we give for choices we have made — career decisions, relationship choices, political positions — may be constructed after the fact to explain and justify intuitive responses, rather than being the actual causes of those responses. We are not necessarily lying when we give these reasons. We genuinely believe them. But they are stories, not causes.

How Retrospective Sensemaking Undermines Learning from Failure

Organizations spend significant resources on post-mortems, lessons-learned exercises, and failure analyses. Much of this investment is wasted — not because the intent is wrong, but because retrospective sensemaking systematically distorts how we interpret failures.

The Hindsight Bias in Organizations

The hindsight bias, documented extensively by Baruch Fischhoff since the 1970s, is the tendency to see past events as having been more predictable than they actually were. Once we know an outcome, we cannot accurately reconstruct what we knew and believed before the outcome occurred.

This has a devastating effect on learning. If we believe a disaster was "obvious in retrospect," we attribute it to negligence or stupidity rather than to the genuine complexity and uncertainty that existed at the time. We then implement safeguards against negligence when the actual vulnerability was a process failure that could affect any competent actor.

Aviation safety researcher Sidney Dekker has documented this pattern extensively in the analysis of aviation accidents. In case after case, post-accident investigations that searched for the "human error" responsible for a crash found the error — and recommended training and discipline as remedies — while leaving intact the systemic conditions that made the error easy to make. Accidents recurred because the actual cause was not addressed. The retrospective narrative had been too neat.

Narrative Gravity

When teams conduct post-mortems in traditional formats — "what went wrong, who was responsible, how do we prevent it" — the social dynamics of the room exert what might be called narrative gravity: pressure to arrive at a clean, causal story with identifiable actors who made identifiable mistakes.

This story is almost always simpler than reality. Real failures in complex systems involve interactions between multiple factors across multiple time scales, many of which were individually reasonable. But "five overlapping system factors created a failure that no individual could have predicted" is an unsatisfying narrative. "The deployment engineer didn't follow the checklist" is satisfying, assignable, and actionable.

The satisfying story, however, is often wrong — and it ensures the same failure will occur again, because the systemic conditions that enabled it were never addressed.

"In a just culture, the question after an incident is not 'who made the mistake?' but 'what conditions made the mistake easy to make?' The goal of the post-mortem is to change conditions, not to punish individuals." — Sidney Dekker, The Field Guide to Understanding Human Error

The Blameless Post-Mortem

Software engineering, borrowing from aviation and healthcare safety, has developed the blameless post-mortem as a structured intervention against these distortions. The core principles:

  • Assume all participants acted with good intentions and the information they had at the time
  • Focus on what the system — processes, tooling, communication — made easy or difficult
  • Invite everyone involved to share their actual experience, without fear of blame
  • Publish findings openly to enable organizational learning

Companies like Google, Etsy, and Netflix have published accounts of how blameless post-mortems changed their engineering cultures. Etsy's engineering team, in particular, produced detailed public documentation in the early 2010s showing how blameless retrospectives increased the rate of incident reporting (because people weren't afraid to surface problems) while decreasing the time to resolution (because root causes were more accurately identified).

The healthcare sector has developed analogous approaches. The After Action Review (AAR) methodology, developed by the U.S. Army and subsequently adapted for medical training and clinical practice, provides a structured conversational format for reviewing critical events: What was planned? What actually happened? Why did it differ? What will we do differently? The format explicitly brackets blame and focuses on understanding the gap between plan and execution — a design choice that makes genuine learning more likely.

The Social Dimension: Sensemaking as Collaborative Narrative

Retrospective sensemaking is not only individual — it is deeply social. When people describe events to others, they are not just reporting; they are collaboratively constructing meaning.

The Water Cooler Effect

Researchers studying organizational sensemaking have documented what might be called the "water cooler effect": informal conversations after significant events are where organizational meaning is actually made. The official narrative — what the all-hands meeting says about a leadership change, a strategic pivot, or an acquisition — competes with and is often overridden by the informal narratives constructed in hallways, chat channels, and lunch tables.

These informal narratives are shaped by each participant's identity, relationships, and prior experiences. They are also highly sensitive to social proof: the interpretation that emerges from early, influential voices tends to be the one that sticks, regardless of its accuracy.

This social construction of organizational history has been studied systematically by Weick and others through the concept of organizing — the ongoing process by which collective meaning is created and maintained. Organizations are not simply structures with explicit cultures; they are ongoing sense-construction projects in which members continuously create and sustain shared accounts of what is happening and why.

Organizational Memory and the Risk of Calcification

Organizations develop collective memories through the accumulation of sensemaking over time. These memories are stored in documents, procedures, cultural stories, and the tacit knowledge of long-tenured employees. They are useful: they allow organizations to act without re-deliberating every routine. But they are also subject to the same reconstructive distortions as individual memory.

Organizational memory can calcify around simplified narratives that no longer accurately represent why particular practices exist. A safety procedure developed in response to a specific incident is observed for years after the original reasons are forgotten — if new employees are told only that "we do it this way," the procedure's logic becomes invisible, and it may be abandoned when it becomes inconvenient, with the failure re-emerging in a new form. The root cause of the original incident, captured in the procedure, lives in the organizational narrative about the procedure's origin — and when that narrative degrades, so does the protection.

Narrative Identity

Psychologist Dan McAdams has argued that human identity is itself a retrospective narrative — what he calls a personal myth. We do not simply have experiences; we construct life stories that integrate experiences, relationships, and values into a coherent account of who we are (McAdams, 2001).

This means that retrospective sensemaking is not a peripheral cognitive glitch but a central process of selfhood. We are, in a meaningful sense, the stories we tell about ourselves.

The practical implication: major experiences are interpreted through the lens of the self-story we already have. An entrepreneur who sees herself as resilient will make sense of failures differently from one who fears he is fundamentally incapable. The factual failure may be identical; the retrospective meaning assigned to it will differ in ways that profoundly affect future behavior.

McAdams's research identifies two key narrative dimensions that shape how individuals retrospectively make sense of their lives: narrative coherence (how well the story holds together as an integrated account) and narrative complexity (how many perspectives, contradictions, and nuances the story accommodates). Research by McLean and colleagues (2007) found that individuals with high narrative complexity — those who could hold contradictions and ambiguities in their life stories without resolving them prematurely — showed better psychological wellbeing and more flexible responses to adversity than those whose self-narratives were highly coherent but simplistic.

The parallel for organizational sensemaking is direct: organizations that maintain complex, nuanced accounts of their history — including failures, contradictions, and unresolved tensions — are better positioned for genuine learning than those that maintain streamlined, heroic narratives of progress.

Practical Implications: Working With Sensemaking

Understanding retrospective sensemaking is not merely academically interesting — it has concrete implications for anyone who makes decisions, leads teams, or tries to learn from experience.

Keep Contemporaneous Records

Because memory is reconstructive and retrospective sensemaking operates on reconstructed memory, the most powerful intervention is capturing experience in real time. Field notes, decision logs, design journals, and project diaries preserve the actual state of knowledge and belief at each moment, making it harder to impose false certainty in retrospect.

Many software teams now practice architectural decision records (ADRs) — brief documents recording not just what was decided but what options were considered, what information was available, and what tradeoffs were accepted. These are invaluable when the decision looks questionable later, because they allow the team to evaluate it against the information available at the time, not against hindsight.

Clinical researchers use pre-registration of hypotheses on platforms like the Open Science Framework or ClinicalTrials.gov for precisely this reason: committing to predictions before data collection prevents the retrospective reframing of exploratory findings as confirmatory ones. This practice is slowly spreading from medical research to behavioral and social science in response to the replication crisis.

Conduct Pre-Mortems

The pre-mortem, developed by psychologist Gary Klein, is a structured technique that deliberately inverts the temporal problem of retrospective sensemaking. Before a project begins (or a decision is implemented), participants are asked to imagine it is one year in the future and the project has failed catastrophically. They then work backward to identify what could have caused that failure.

Because the pre-mortem is prospective, it bypasses hindsight bias. And because it licenses "negative" thinking in a constructive context, it surfaces concerns that might not be raised in a standard planning meeting where social pressure favors optimism. Klein's research found that pre-mortems consistently identified more failure modes than standard risk assessments, and identified failure modes that standard assessments systematically missed (Klein, 2007).

Use Structured Reflection

Not all sensemaking is distorting. Structured reflection — in the form of after-action reviews, learning histories, and narrative debriefs — can channel retrospective sensemaking toward genuine insight.

The key is to make the process explicit and collaborative:

  • What did each person expect to happen?
  • What actually happened?
  • Where did those expectations diverge?
  • What does that divergence reveal about our mental models?

By making the sensemaking process visible, teams can identify where their collective narratives diverged from experience and update their models accordingly.

Learning histories, a methodology developed by Art Kleiner and George Roth at MIT, extend this approach by interviewing multiple stakeholders after significant organizational events, compiling their accounts in a format that juxtaposes different perspectives, and using the compiled narrative as a reflective tool. The format makes explicit the fact that different actors experienced the same events differently — a direct challenge to the common assumption that there is a single correct retrospective account of what happened.

Deliberate Counterfactual Thinking

One research-supported practice for limiting the distortions of retrospective sensemaking is deliberate counterfactual thinking — asking "what if it had gone differently?" after an outcome is known. Research by Roese and Epstude (2008) suggests that actively entertaining counterfactuals reduces hindsight bias by restoring a sense of the contingency that was actually present before the outcome occurred. If it is easy to imagine how things could have gone another way, the current outcome feels less inevitable — and the false certainty of "I knew it all along" is harder to sustain.

Conclusion

Retrospective sensemaking is not a bug in human cognition; it is a fundamental feature of how meaning is made. We cannot experience events in real time with perfect comprehension — complexity is too high and time too short. We act, and then we understand what we were doing.

The problems arise when this process operates invisibly and unchecked — when we mistake our retrospective narratives for accurate memory, when we allow hindsight to make failures seem obviously avoidable, and when we construct self-serving stories that prevent genuine learning.

Understanding that meaning is made backward, that memory is reconstructed rather than retrieved, and that our experience of rational decision-making is often a post-hoc story about processes already completed — this understanding does not paralyze action. It invites a kind of epistemic humility: an acknowledgment that we understand our own history less accurately than we feel we do, and that the stories we tell about ourselves are more creative than we realize.

The goal is not to stop making sense. It is to make it better.

References

  • Weick, K. E. (1995). Sensemaking in Organizations. Sage Publications.
  • Weick, K. E. (1993). The collapse of sensemaking in organizations: The Mann Gulch disaster. Administrative Science Quarterly, 38(4), 628-652. https://doi.org/10.2307/2393339
  • Bartlett, F. C. (1932). Remembering: A Study in Experimental and Social Psychology. Cambridge University Press.
  • Loftus, E. F., & Palmer, J. C. (1974). Reconstruction of automobile destruction: An example of the interaction between language and memory. Journal of Verbal Learning and Verbal Behavior, 13(5), 585-589. https://doi.org/10.1016/S0022-5371(74)80011-3
  • Fischhoff, B. (1975). Hindsight is not equal to foresight: The effect of outcome knowledge on judgment under uncertainty. Journal of Experimental Psychology: Human Perception and Performance, 1(3), 288-299. https://doi.org/10.1037/0096-1523.1.3.288
  • Gazzaniga, M. S. (1985). The Social Brain: Discovering the Networks of the Mind. Basic Books.
  • Haidt, J. (2001). The emotional dog and its rational tail: A social intuitionist approach to moral judgment. Psychological Review, 108(4), 814-834. https://doi.org/10.1037/0033-295X.108.4.814
  • Dekker, S. (2006). The Field Guide to Understanding Human Error. Ashgate Publishing.
  • McAdams, D. P. (2001). The psychology of life stories. Review of General Psychology, 5(2), 100-122. https://doi.org/10.1037/1089-2680.5.2.100
  • McLean, K. C., Pasupathi, M., & Pals, J. L. (2007). Selves creating stories creating selves: A process model of self-development. Personality and Social Psychology Review, 11(3), 262-278. https://doi.org/10.1177/1088868307301034
  • Arkes, H. R., Wortmann, R. L., Saville, P. D., & Harkness, A. R. (1981). Hindsight bias among physicians weighing the likelihood of diagnoses. Journal of Applied Psychology, 66(2), 252-254. https://doi.org/10.1037/0021-9010.66.2.252
  • Klein, G. (2007). Performing a project premortem. Harvard Business Review, 85(9), 18-19.
  • Roese, N. J., & Epstude, K. (2008). The functional theory of counterfactual thinking. Advances in Experimental Social Psychology, 40, 283-306. https://doi.org/10.1016/S0065-2601(07)00006-X
  • Kleiner, A., & Roth, G. (1997). How to make experience your company's best teacher. Harvard Business Review, 75(5), 172-177.

Tags

sensemaking psychology cognitive bias memory organizational learning

Frequently Asked Questions

What is retrospective sensemaking?

Retrospective sensemaking is the process by which people construct meaning from past events after the fact. Coined by organizational psychologist Karl Weick, the concept describes how we impose a coherent narrative on experiences that were, while happening, ambiguous and confusing. We understand what we did only by looking back at it — action precedes interpretation.

How does retrospective sensemaking differ from memory?

Memory stores information; sensemaking creates meaning. Memory research shows that human recall is reconstructive, not reproductive — we do not replay stored recordings but rebuild memories each time we retrieve them. Retrospective sensemaking is the layer of interpretation and narrative we add to these reconstructed memories, shaping them into coherent stories about who we are and why we act as we do.

What is post-hoc rationalization and how is it related?

Post-hoc rationalization is the most familiar form of retrospective sensemaking: constructing a rational explanation for a decision that was actually made intuitively or emotionally. Neuroscientist Michael Gazzaniga's 'interpreter' hypothesis suggests the left brain hemisphere is specifically evolved to generate plausible explanations for actions initiated by other brain systems, creating the sense that we are always in rational control.

Why does retrospective sensemaking make it hard to learn from failure?

When we look back at failures, we tend to construct narratives that preserve our self-image, attribute causes externally, and make the outcome seem inevitable in hindsight (the hindsight bias). These narrative distortions mean we often fail to identify the real causes of errors. Blameless post-mortems in engineering and healthcare are specifically designed to counteract this tendency by focusing on system factors rather than individual blame.

Can retrospective sensemaking be a useful process?

Yes. Sensemaking is not only a source of distortion — it is also how we integrate experience into knowledge. Structured sensemaking practices, like after-action reviews, narrative debriefs, and reflective writing, can help individuals and organizations extract genuine learning from experience rather than simply generating self-serving stories. The key is to make the sensemaking process explicit and collaborative rather than private and automatic.

Share this article