Why Organisational Change Still Fails
Reading Time: 10 Minutes
Image created by the Superhumxn team.
Change
Jeff Immelt spent eleven years trying to transform GE. He invested billions into GE Digital, rebranded the company as a "digital industrial" business, and made innovation the centrepiece of his tenure. By most external measures, the ambition was serious and the resources were enormous. By the time he left in 2017, GE's stock had fallen 30% and the digital transformation had largely failed. The strategy was not wrong. The execution never caught up with it because the cultural and operational conditions surrounding 300,000 employees did not change at the pace the vision demanded.
GE is not an outlier. Between 60 and 70% of change initiatives fail. A 2024 Gartner survey found that 73% of employees are fatigued from change and 74% of managers do not feel equipped to lead it. The failure rate has not meaningfully improved in decades despite significant investment in change management and massive technological progress.
This doesn’t mean leaders lack the ambition or intelligence to lead change, but most transformation programmes are built around an assumption that explaining the change clearly enough will encourage people to want it. Human behaviour is not driven by information. Instead, emotions, incentives, and the effort required to sustain any given change all contribute.
With 95% of organisations now managing more than two major transformations over a 3-year span, and AI accelerating the pace of technological change exponentially, managing change is one of the most urgent leadership challenges of the decade.
McKinsey's next-generation operating model argues that most transformation efforts fall short because they run as uncoordinated initiatives in separate silos rather than as an integrated system. It's a useful frame for understanding why even well-resourced change programmes so often underdeliver.
What can I expect from a subscription?
Real-world strategies for navigating the future of work.
AI AGENTS AT WORK
Everything you need to know about how AI agents are changing teams and ways of working.
SUPERHUMXN IRL
Real insights from investors and operators.
MEMOS ON PEOPLE, CULTURE & AI AT WORK
The latest use cases and strategies for change and transformation.
Longevity
Change fatigue is a workplace risk
In 2022, the average employee experienced 10 major organisational changes, including restructures and technology system replacements. This is up from just two in 2016. That fivefold increase happened before AI accelerated the pace of change even more. For most teams, the question is no longer how to manage change as a scheduled event, but how to work within a constant and permanent state of change.
Neuroscience warns us of the risks when organisations get this wrong. Neuroscientist Bruce McEwen's concept of allostatic load describes the cumulative biological cost of chronic stress. Each change has a physiological price to pay. When there is enough recovery between demands of any given change, the body can reset. When there is no recovery time, the stress accumulates. The prefrontal cortex, the part of the brain responsible for decision-making, planning, and complex judgement, is particularly vulnerable to this kind of chronic pressure. Sustained change without adequate recovery physically impairs the capacity for exactly the kind of thinking high-performing organisations rely on.
The business consequences are measurable. A 2025 Gartner study found that 79% of employees have low trust in change initiatives, and that organisations with better than average healthy change adoption report two times higher year-over-year revenue growth. For large companies, Gartner estimates that gap at up to $2.2 billion annually. Change fatigue absolutely bears a cost.
Gartner's research consistently identifies two factors that most reliably increase employees' capacity to absorb change: trust and team cohesion. Employees with high trust in their leadership have 2.6 times the capacity for change compared to those with low trust. Employees in high-cohesion teams have nearly twice the capacity. The data says it all. Before layering on more change, companies must invest in the foundational infrastructure that supports employees.
Image created by the Superhumxn team.
Lasting
The 3 sides of change
BJ Fogg's behaviour model, developed at Stanford, offers useful data for organisations working through change. Behaviour change only sticks when three elements converge at the same moment: motivation, ability, and a reminder, or prompt. Remove any one of these conditions and the behaviour change will not last.
That is why a workshop can generate genuine enthusiasm and produce almost no lasting change. When employees return to an environment still rewarding the old behaviours, over time any new habits will fall away.
Here are a few IRL examples to take inspiration from:
Motivation: Satya Nadella at Microsoft connected the shift to a growth mindset directly to a compelling narrative about the company's survival and purpose, rather than performance goals only.
Ability: Google's Project Oxygen took its manager research findings and embedded them into the tools managers already used, making the new behaviours easier to adopt.
Prompt: Humu and Perceptyx deliver prompts, or nudges, inside Slack and Teams. These happen in real-time during work, not in a training room six months earlier.
Psychological safety is often misunderstood
When Boeing's new CEO Kelly Ortberg addressed employees in early 2025, one of the challenges he highlighted wasn't related to their technology or manufacturing. It was communication. The company needed a culture where people felt comfortable speaking up, sharing concerns, and communicating across teams more openly.
That's a surprisingly common challenge during periods of change. It’s natural for leaders to assume fewer questions or less pushback means people are aligned. In reality, employees may simply decide it's safer to stay quiet. This typically precedes a period of “quiet quitting,” where teams disengage from work and withdraw any discretionary effort.
Amy Edmondson and Michaela Kerrissey's research suggests this is backwards. Tracking more than 27,000 healthcare workers through the pandemic, they found that a one standard deviation increase in psychological safety cut burnout by 0.72 points and increased willingness to stay by 0.63 points, even among staff working with inadequate tools and short-staffed teams. The people who had felt safe enough to speak up before the crisis were more likely to still be there after it.
"Psychological safety, rather than being created by a policy, is built in a group, interaction by interaction." It does not require a mandate from the top. Any team, at any level, can work towards this.
The strongest adaptive cultures are therefore not usually the ones with the least tension. More often, they are environments where people can move through tension, disagreement, experimentation, and uncertainty without shutting down socially or withdrawing from participation altogether.
Hypocrisy
Why workplace change often fails
Picture a company leader that invests heavily in developing its values and an attractive careers page, and then pings everyone on Slack at 10pm expecting a response. Or one that tells managers to encourage experimentation, then punishes the employee whose experiment didn't work out. Humans are wired for pattern recognition, and people clock these kinds of discrepancies. Mismatched values and behaviour erode organisational trust.
MIT Sloan analysed 1.2 million Glassdoor reviews using AI to compare what companies say their values are against what employees actually experience day to day. The gap was consistent across almost every organisation studied. Official values rarely translate into behavioural guidance specific enough to act on. "Innovation" and "integrity" sound impressive on a pitch deck, but employees learn what really matters by watching what gets rewarded versus ignored.
BAU
Consistency is not enough
A common trap companies can fall into is locking in practices that do not support the modern pace of change. For example, preserving outdated values due to legacy.
Alignment between company values and behaviour in itself is not enough to guarantee success. Research on this dates as far back as 1992, when Kotter and Heskett found that strength of culture alone, meaning how widely shared and deeply held the values are, did not predict performance. What predicted performance was whether the culture could keep adapting.
Adaptability is a cultural aspect that enables a company to respond to changing environmental conditions.
Cognizant, which employs around 360,000 people, is trying to tackle this challenge head-on. As AI changes how work gets done, the company announced they are moving away from a traditional top-down structure so teams can adapt and make decisions faster, rather than waiting for instructions to filter through layers of management.
That requires a different mindset. Ganesh Ayyar, who leads automation at Cognizant, says the company is actively trying to build a culture of experimentation. The expectation isn't that every idea will work. It's that people test things, learn quickly, and share what they discover. As he puts it, organisations need to be willing to "celebrate responsible failures" because some experiments simply won't succeed.
What makes his comments interesting is that he doesn't pretend leaders have all the answers. Speaking about AI, he admitted: "Even I'm scared. There is fear, fear of uncertainty, fear of the unknown, fear of ambiguity." The challenge isn't eliminating uncertainty, but instead learning how to keep moving regardless.
Leverage
Cognizant isn’t pretending to have it figured out. "Even I'm scared," Ayyar said, while leading automation for a company of 360,000 people. That admission is rare but refreshing, and certainly not a weakness by any measure.
Donella Meadows, who helped popularise systems thinking, believed the most powerful changes often happen in places people overlook. She calls these places leverage points, defined as points within a complex system “where a small shift in one thing can produce big changes in everything."
Most attention goes to parameters (like budgets and targets), which she calls "dead last" in terms of leverage. Higher leverage sits in "the structure of information flows" and "the rules of the system” like incentives, punishments, and constraints.
