The research on why this happens and what overcomes it is sufficiently specific to be practically useful. The failure modes are not random — they follow predictable psychological patterns that Schein’s model predicts in advance, Kotter’s analysis of 100 transformations documented in retrospect, and the social norms research explains mechanically.

Why the assumptions layer resists and how it reverts

Schein’s (1985) three-level model predicts the primary culture change failure mechanism precisely. Initiatives that address artefacts — new values posters, culture workshops, updated mission statements, leadership principles published on the intranet — while leaving the underlying assumptions untouched produce surface-level change that reverts to the original pattern as soon as the initiative’s energy dissipates.

The assumptions layer resists change because it is the cognitive default system. The schema that interprets ambiguous situations, generates automatic responses to familiar cues, and determines what feels right and wrong operates below conscious deliberation. When a new stated value conflicts with an existing assumption, the assumption governs behaviour — because behaviour is not primarily generated by stated values but by the automatic interpretation systems that the assumptions constitute.

Schein identified the psychological conditions required for genuine assumption-level change. First, “survival anxiety” — the recognition that the current assumptions are producing outcomes the organisation cannot tolerate. This recognition must be specific enough to make the status quo feel genuinely costly. Second, “learning anxiety” — the anxiety that the alternative assumptions will require capabilities, behaviours, and self-concepts that the current membership does not have. Third, sufficient psychological safety to reduce learning anxiety enough that people are willing to engage with the discomfort of genuine change. Without psychological safety, survival anxiety does not produce learning — it produces defensive rigidity.

The specific implication is that culture change requires generating a credible case that the current assumptions are producing unacceptable outcomes (not just suboptimal ones), and simultaneously creating the conditions under which the anxiety of changing can be tolerated rather than defended against.

The eight failure modes and what they reveal about why change doesn’t stick

Kotter’s (1995) analysis of 100 large-scale organisational transformations identified eight failure modes that recur across organisations, sectors, and stated change objectives. Each represents a specific psychological mechanism.

Insufficient urgency means the cognitive and emotional case for change does not overcome the comfort and familiarity of the status quo. The loss aversion and status quo bias research predicts this: the perceived costs of changing are weighted more heavily than the perceived benefits, and the case for change must exceed this asymmetric weighting before behaviour begins to shift.

A weak guiding coalition means the social proof of existing authority structures continues to signal the old norms. Culture change requires that enough high-status, visible people are genuinely embodying the new behaviours that the descriptive norm begins to shift. A change declared by one senior leader who does not change their own behaviour is producing the injunctive norm (what is supposed to happen) without the descriptive norm (what actually happens) — and the descriptive norm is what governs.

Failure to remove obstacles means the systems and structures that reward old behaviours continue operating while the new behaviours are being declared. The operant conditioning mechanism is precise: if the reward system continues to reinforce the old pattern, the old pattern persists regardless of what is stated. The performance review that still rewards the behaviours the culture change was designed to replace is communicating, through the reward system, that the change is rhetorical.

The absence of short-term wins means the reward schedule for new behaviour is too delayed to sustain motivation through the change period. The behavioural psychology of habit formation requires feedback loops that are close enough to the behaviour to reinforce it — which is why change initiatives that promise results in three to five years typically lose the behavioural compliance they need to get there.

Premature declaration of victory means the new behaviours have not yet become habitual and self-sustaining — they are still deliberate compliance with an injunctive norm rather than automatic expression of an embedded assumption. Deliberate norm compliance reverts under pressure. Habitual norm compliance does not. The culture change that is declared complete before the new behaviours have become automatic will revert at the first significant organisational stress event.

How norm cascades actually produce culture change

Cialdini, Reno and Kallgren’s (1990) social norms research provides the specific mechanism through which culture change succeeds when it does. Behaviour is governed primarily by descriptive norms — what people observe others actually doing in this environment — not by injunctive norms, which are what they are told they should do. Failed culture change attempts to shift the injunctive norm while leaving the descriptive norm intact. Successful culture change creates enough visible behavioural change that the descriptive norm shifts, at which point the social proof mechanism produces adoption across the remaining population through conformity pressure.

The informational cascade mechanism (Bikhchandani, Hirshleifer and Welch, 1992) predicts the threshold effect: once enough visible, high-status people in the organisation are genuinely exhibiting the new behaviours, adoption cascades through the remainder of the population through rational updating — if the people I most respect are behaving this way, this is apparently how things are done here. The cascade does not require convincing everyone; it requires convincing enough visible people that their behaviour changes the descriptive norm, at which point the cascade mechanism does the remaining work.

The practical implication is that culture change is a social norm cascade problem, not a communication problem. More communication about what the culture should be does not change the descriptive norm. High-status leaders genuinely changing their behaviour changes the descriptive norm — because behaviour is the descriptive norm, and the descriptive norm is what the social proof mechanism causes others to adopt.

What the Microsoft case and the small wins research establish

Microsoft’s documented culture transformation under Satya Nadella illustrates all three necessary mechanisms operating simultaneously. The reward system was changed: stack ranking, which had been producing the zero-sum, knowledge-hoarding behaviours that characterised the old culture, was eliminated. The leadership modelling was genuine: Nadella’s public acknowledgment of his own learning, his uncertainty, and his mistakes provided the visible high-status behaviour that the descriptive norm cascade required. The change was sustained rather than declared: the new behaviours were reinforced consistently through multiple cycles rather than being announced and then left to self-sustain.

Weick’s (1984) small wins research adds the mechanism that explains why large-scale culture change declarations typically fail where incremental behaviour change succeeds. Large-scale change is psychologically overwhelming and activates the defensive rigidity that threat-rigidity research predicts. Small, specific, achievable behaviour changes — one meeting structure modified, one reward system adjusted, one piece of feedback delivered differently — accumulate into substantial cultural transformation over time without triggering the resistance that sweeping declarations produce.

The Wood and Rünger (2016) habit formation research confirms why the accumulation matters: the new cultural norms must be practiced until they become habitual rather than deliberate, because deliberate norm compliance requires cognitive resources that are unavailable under pressure. The culture that reverts at the first significant challenge has not yet habituated the new norms; the culture that holds under pressure has. The transformation from deliberate to automatic is not declared — it is accumulated through repetition across varied contexts and conditions.

What this means for the founder who wants to change the culture they inadvertently created

The specific application for early-stage companies is the founder who recognises that the culture their behaviour has installed is not the culture they want to sustain at scale. The recognition is necessary but insufficient. The change requires that the founder’s own behaviour changes first and visibly — not that they announce a new culture. The descriptive norm that the founder’s behaviour has established cannot be revised by the founder’s communication; it can only be revised by the founder’s behaviour.

This is the mechanism that makes culture change so personally demanding for founders. The culture is not a separate organisational entity that can be modified through policy. It is the accumulated expression of the founder’s behaviour over time, and changing it requires the founder to change — specifically, consistently, and visibly enough that the descriptive norm they establish reverts to the new behaviour pattern rather than the old one at every opportunity the organisation has to observe which pattern is operating.

Books worth reading on this

The Heart of Change by John Kotter and Dan Cohen. Kotter and Cohen’s follow-up to the original Leading Change research — addressing the specific emotional and behavioural mechanisms required at each stage of organisational transformation rather than the structural stages alone — provides the most practically detailed available account of what it actually looks like to navigate the eight failure modes this article identifies. Their specific treatment of how to build the guiding coalition, sustain urgency through the middle of a change initiative, and anchor new behaviours before declaring victory is the most directly applicable available complement to the research framework.

If the 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: Schein, E.H. (1985, 5th ed. 2017), Organizational Culture and Leadership, Jossey-Bass. Kotter, J.P. (1995), Leading Change: Why Transformation Efforts Fail, Harvard Business Review, 73(2), 59–67. Cialdini, R.B., Reno, R.R. & Kallgren, C.A. (1990), A Focus Theory of Normative Conduct, Journal of Personality and Social Psychology, 58(6), 1015–1026. Bikhchandani, S., Hirshleifer, D. & Welch, I. (1992), A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades, Journal of Political Economy, 100(5), 992–1026. Weick, K.E. (1984), Small Wins: Redefining the Scale of Social Problems, American Psychologist, 39(1), 40–49. Wood, W. & Rünger, D. (2016), Psychology of Habit, Annual Review of Psychology, 67, 289–314. Edmondson, A.C. (1999), Psychological Safety and Learning Behavior in Work Teams, Administrative Science Quarterly, 44(2), 350–383. Heath, C. & Heath, D. (2010), Switch: How to Change Things When Change Is Hard, Crown Business. Kotter, J.P. & Cohen, D.S. (2002), The Heart of Change, Harvard Business School Press.