In 1968, Robert Rosenthal and Lenore Jacobson published an experiment that should have changed how every organisation in the world hires, manages, and develops people. It didn't — which makes it more valuable, not less, to anyone paying attention.
The experiment was deceptively simple. At the start of the school year, researchers administered an IQ test to all students at an elementary school in San Francisco. They then told teachers that certain students — roughly 20% of each class — had been identified as "intellectual bloomers" who would show unusual academic gains over the coming year. The teachers received the names.
The critical detail: the "bloomers" had been selected entirely at random. The test scores were irrelevant. There was nothing special about the identified students. The only variable was the teacher's belief.
Eight months later, the researchers returned and retested every student. The randomly selected "bloomers" had gained significantly more IQ points than the control group. First and second graders showed the most dramatic effects — bloomers gained 12.2 and 8.4 more IQ points, respectively, than their peers. The teachers hadn't been given a curriculum. They hadn't been given teaching techniques. They'd been given a belief — and the belief had altered reality.
Rosenthal and Jacobson called it the Pygmalion Effect, after the Greek myth of a sculptor who carved a statue so beautiful he fell in love with it, and Aphrodite brought it to life. The myth captures the mechanism precisely: the creator's belief in the creation's worth made it real. George Bernard Shaw had dramatised the same principle in his 1912 play Pygmalion — later adapted as My Fair Lady — where Professor Higgins transforms a Cockney flower girl into someone who passes for a duchess. Eliza Doolittle herself identifies the mechanism in the final act: the difference between a lady and a flower girl is not how she behaves, but how she's treated.
The mechanism Rosenthal documented operates through four channels. Teachers who expected more from certain students provided warmer emotional climates — more smiling, more nodding, more eye contact. They gave those students more input — more material, more complex ideas, more time on task. They provided more opportunities to respond — calling on them more frequently, waiting longer for answers, prompting rather than moving on. And they delivered more differentiated feedback — specific praise for correct answers, specific correction rather than dismissal for wrong ones.
None of this was conscious. The teachers didn't know they were treating students differently. When asked, they denied it. Observational studies confirmed the behavioral differences were real and measurable. The teachers' expectations had rewired their behavior without their awareness, and that rewired behavior had produced the outcomes the expectations predicted.
The reverse operates with equal force. When teachers, managers, or leaders hold low expectations for someone, they provide less attention, less challenge, less feedback, and less patience. The target receives fewer opportunities to develop and more signals that development isn't expected. Performance declines. The low expectations are confirmed. Rosenthal termed this the Golem Effect, after the Jewish folklore creature of clay — powerful but ultimately lifeless, shaped by its creator's limitations rather than its own potential.
The implications for founders and leaders are direct and uncomfortable. Every hiring decision, every performance review, every team assignment, every one-on-one meeting carries an implicit expectation. That expectation — communicated through attention allocation, challenge level, feedback quality, and emotional tone — shapes the outcome it predicts. A leader who decides a new hire "isn't quite strong enough" will unconsciously provide less mentoring, less challenge, and less patience. The hire will underperform. The leader will point to the underperformance as evidence of their correct assessment.
The prophecy fulfils itself, and the prophet never knows they wrote the script.
Section 2
How to See It
The Pygmalion Effect is invisible to the person creating it. Leaders who shape their team's performance through expectations rarely recognise the mechanism — they experience the outcome as evidence of their judgment, not their influence. Look for the fingerprints, not the hand.
Leadership
You're seeing the Pygmalion Effect when a manager's "star" hires consistently outperform while their "risky" hires consistently wash out — at a rate that exceeds statistical probability. If a leader's initial assessment predicts outcomes with near-perfect accuracy, the assessment is likely creating the outcome rather than measuring it. The diagnostic question: do this manager's initial impressions ever prove wrong? If not, the Pygmalion Effect is operating.
Culture
You're seeing the Golem Effect when certain teams or divisions are labelled "B-team" or "non-strategic" and consistently underperform expectations, even after leadership changes. The institutional expectation — embedded in resource allocation, executive attention, and promotion pathways — produces the mediocrity it predicts. IBM's hardware divisions in the 2000s exhibited this pattern: labelled as legacy businesses, they received declining investment, lost their strongest engineers to growth divisions, and delivered exactly the stagnation that justified further disinvestment.
Hiring
You're seeing the Pygmalion Effect in a system when an organisation's performance data shows that candidates from prestigious backgrounds consistently outperform — but only within that organisation. If Stanford graduates outperform at your company but not across the industry, your company's expectations of Stanford graduates (more mentoring, faster promotion tracks, higher-profile projects) are likely producing the performance differential, not the degree.
Product
You're seeing the Pygmalion Effect in product teams when a founder's personal enthusiasm for a project predicts its success more reliably than market data does. Steve Jobs's conviction that the Macintosh team could build something insanely great produced behaviors — obsessive attention, unlimited resources, direct CEO involvement — that made greatness more likely. His indifference to other projects produced the opposite resource allocation. The belief preceded the outcome and produced it.
Section 3
How to Use It
Decision filter
"Before evaluating someone's performance, ask: what expectations did I set — explicitly or implicitly — and how did those expectations shape the conditions they performed in? If you can't separate your judgment from your influence, you're measuring your own prophecy, not their capability."
As a founder
Your expectations for each team member create the ceiling and floor of their performance. This isn't motivational rhetoric — it's the consistent finding across decades of research in classrooms, military units, and corporate settings.
The operational implication: be deliberate about what you expect from people, because you'll get it either way. When you assign someone a stretch project with the genuine belief they can handle it, you'll unconsciously provide the support, patience, and attention that makes success more likely. When you assign someone a stretch project while privately doubting they're ready, you'll unconsciously withhold exactly those resources.
Audit your own behavior. Track who gets your time, your follow-up questions, your benefit of the doubt. If you're spending 80% of your one-on-one attention on the three people you already consider top performers, you're running a Pygmalion Effect that widens the gap between your best and worst people — and then interpreting that widening gap as validation of your original assessment.
As a leader
The most powerful Pygmalion intervention is changing what "high performance" signals to your team. If performance reviews reward demonstrated competence — what people already know — you signal a fixed expectation. If reviews reward learning velocity — how fast people develop new capabilities — you signal an expandable expectation. The second frame activates the Pygmalion mechanism: people who believe their leader expects growth will seek opportunities to demonstrate growth.
Structure your feedback to carry the expectation of improvement. "This analysis missed three critical factors" is a dead-end statement. "This analysis missed three critical factors — here's how to build that capability, and I expect to see it in the next iteration" carries an implicit expectation of development that shapes the recipient's effort allocation. The information content is identical. The expectation payload is different.
As a decision-maker
The Pygmalion Effect has a structural implication for capital and resource allocation. Every budget decision, team assignment, and strategic priority carries an embedded expectation. When a board designates a business unit as "core" versus "non-core," the designation reshapes resource flows, talent allocation, and management attention in ways that fulfil the classification. Non-core units receive less investment, lose their best people to core units, and deliver the declining returns that justify their non-core designation.
The diagnostic: before labelling any unit, team, or initiative as high or low potential, ask what would happen if you allocated resources as though the opposite were true. If a struggling division received the same investment, executive attention, and talent pipeline as your star division, would the performance gap persist? If you're not sure, you may be confusing your expectations with reality.
Common misapplication: Treating the Pygmalion Effect as a justification for naive optimism. High expectations produce superior outcomes only when paired with genuine support — the four channels Rosenthal identified (warmth, input, response opportunity, differentiated feedback). A leader who announces "I have high expectations" but provides no additional mentoring, no stretch assignments, and no specific feedback has adopted the label without the mechanism. Expectations without support produce frustration, not growth.
Second misapplication: Assuming the effect operates with equal strength across all contexts. Rosenthal's original study showed the strongest effects in younger children and early-tenure employees — people whose self-concept is still forming and who are most responsive to authority figures' expectations. Senior professionals with established track records and strong self-models are less susceptible to external expectations, though not immune.
Section 4
The Mechanism
Section 5
Founders & Leaders in Action
The Pygmalion Effect is most visible in leaders who consistently extract performance from teams that other leaders had written off — and in the structural mechanisms they build to scale their expectations across an organisation. The pattern is not "be positive." The pattern is: hold a specific, genuine belief in someone's capacity, then behave in ways that reflect that belief through resource allocation, attention, and challenge.
The leaders below didn't just happen to have talented teams. They created the conditions under which talent emerged — often from people that previous leaders had overlooked or dismissed.
The "reality distortion field" — the term Apple engineers used to describe Jobs's insistence that impossible deadlines and specifications were achievable — is the most documented case of the Pygmalion Effect in technology history. The mechanism was not charisma alone. It was that Jobs genuinely believed his teams could accomplish what they considered impossible, and that belief rewired his behavior in ways that made the impossible more likely.
During the development of the original Macintosh in 1983–1984, Jobs told the team they would build a computer that would "put a dent in the universe." Engineers who protested that specific technical targets couldn't be met on schedule were told they could, and then received the resources, attention, and protection from bureaucracy to prove it. Bill Atkinson, the engineer who created MacPaint, recounted that Jobs's absolute conviction that the team could build a graphical interface for a $2,495 computer — when comparable technology cost $10,000 — changed what the engineers believed was achievable. The expectations didn't suspend physics. They suspended the engineers' assumptions about what was possible within the constraints.
The dark side was equally powerful. Jobs's low expectations for people or projects produced Golem effects of matching intensity. Teams and executives who fell out of his favor experienced an abrupt withdrawal of attention, resources, and patience. The MobileMe team in 2008 received the Golem treatment publicly — Jobs excoriated the entire group after the botched cloud service launch. The pattern was consistent: the intensity of his belief — positive or negative — shaped the conditions that confirmed it.
Apple's market capitalisation was $3 billion when Jobs returned in 1997. By the time of his death in 2011, it was $350 billion. The iMac, iPod, iPhone, and iPad were each built by teams that Jobs believed could do something the market considered impractical. His expectations didn't guarantee success. They altered the probability distribution.
Nadella's transformation of Microsoft was, at its root, a Pygmalion intervention applied to 127,000 people. The "know-it-all to learn-it-all" reframe — borrowed from Carol Dweck's growth mindset research — was a deliberate reset of organisational expectations. Under Steve Ballmer's stack-ranking system, the institutional expectation for roughly 20% of employees was that they were underperformers. The label shaped resource allocation (less mentoring, fewer growth opportunities), which shaped performance (stagnation or decline), which confirmed the label.
Nadella replaced the system with one that asked: "How have you grown? How have you helped others grow?" That structural change carried a new expectation — every employee was capable of development. The expectation was encoded in incentive design, not just rhetoric. Managers who developed their reports were rewarded. Teams that collaborated across divisions were recognised. The expectation of growth became the institution's default assumption about its people.
Azure grew from roughly $4 billion to over $80 billion in annual revenue between 2015 and 2024. That growth required thousands of engineers to develop capabilities in cloud computing, artificial intelligence, and enterprise sales that they hadn't possessed when Nadella took over. They developed those capabilities because the organisation expected them to, resourced them to, and rewarded them for doing so.
The counterfactual is instructive: under the previous system, many of those same engineers had been classified as adequate or underperforming. The humans hadn't changed. The expectations had.
Welch's "differentiation" system at GE is the most instructive case of the Pygmalion and Golem effects operating simultaneously at industrial scale. The system classified employees into three tiers: the top 20% ("A players"), the middle 70% ("B players"), and the bottom 10% ("C players"). A players received disproportionate compensation, development opportunities, stock options, and executive attention. C players were counselled out. The system operated annually across 300,000 employees for two decades.
The Pygmalion Effect on the top tier was deliberate and powerful. Welch personally reviewed the development plans of GE's top 500 leaders. He spent roughly 40% of his time on talent assessment and development. The message was unambiguous: GE expects extraordinary performance from you, and it will invest extraordinary resources to get it. Leaders identified as A players received stretch assignments, executive mentoring, and fast-track promotions that developed capability at an accelerated rate.
The Golem Effect on the bottom tier was equally powerful and less discussed. Employees classified as C players received the institutional equivalent of Rosenthal's control group: less attention, fewer resources, lower-quality feedback, and an ambient expectation of departure. Whether they were genuinely the lowest performers, or whether the classification itself contributed to the underperformance through reduced investment, is the central question the Pygmalion Effect raises about any forced-ranking system.
GE's revenue grew from $27 billion to $130 billion under Welch, and the company's market capitalisation increased from $14 billion to over $400 billion. Welch's defenders attribute this to rigorous talent selection. His critics argue that any system which invests disproportionately in people it has already labelled as winners will produce evidence that those people are winners — regardless of the accuracy of the initial classification.
Ed CatmullCo-founder & President, Pixar Animation Studios, 1986–2018
Catmull built Pixar's Braintrust as a structural Pygmalion mechanism — a system designed to hold high expectations for every creative project while providing the support channels that make those expectations productive rather than crushing.
The foundational belief, articulated in Creativity, Inc. (2014), was that "early on, all of our movies suck." That statement carried an embedded Pygmalion expectation: if every movie starts badly and the final product is excellent, then every creative team is capable of the transformation from bad to great. The expectation wasn't that people would produce brilliance on the first draft. It was that they would produce brilliance through iteration — and that every person in the building had the capacity for that iterative journey.
The Braintrust's structure operationalised all four Rosenthal channels. Climate: feedback was delivered with candor but without personal attack, creating psychological safety around creative risk. Input: directors received notes from the most experienced creative leaders at the studio — more complex, more challenging input than a standard review process. Response opportunity: directors were not required to implement any Braintrust suggestion, preserving their agency and the expectation that they could find the right solution themselves. Feedback: specific, actionable, and focused on the work rather than the person.
Between 1995 and 2015, Pixar released fourteen consecutive commercially successful films with cumulative worldwide box office exceeding $14 billion. The consistency suggests that the creative output was produced by a system, not by individual genius — a system whose core operating principle was that high expectations, properly supported, would reliably produce high-quality results from a range of different creative teams.
Section 6
Visual Explanation
Section 7
Connected Models
The Pygmalion Effect sits at the intersection of expectation, behavior, and system design. It provides the psychological mechanism behind several adjacent frameworks and generates productive tension with models that complicate its clean narrative. Understanding these connections reveals when the effect is most powerful and where it can mislead.
Reinforces
Feedback Loops
The Pygmalion Effect is a feedback loop with a human operator. High expectations → different leader behavior → better performance → confirmed expectations → sustained high behavior. The loop is self-reinforcing in both directions: positive expectations compound upward, negative expectations compound downward. Understanding feedback loop structure reveals why the Pygmalion Effect is so difficult to interrupt once established — the loop generates its own confirming evidence at each rotation. Rosenthal's four-channel mechanism (climate, input, response opportunity, feedback) maps precisely to the flows that sustain the loop. Breaking the cycle requires intervening at the expectation node, not the performance node — because performance is downstream of everything else.
Reinforces
Growth vs Fixed Mindset
The Pygmalion Effect is growth mindset projected outward. A leader with a growth mindset believes their team members can develop new capabilities. That belief activates the Pygmalion mechanism — more challenge, more support, more patience with the messy middle of skill development. A leader with a fixed mindset believes capability is innate and stable. That belief activates the Golem mechanism for anyone not already demonstrating high performance — less investment, less challenge, less patience. Nadella's transformation of Microsoft explicitly linked these: by shifting the institutional expectation from "prove what you know" to "demonstrate what you're learning," he activated a Pygmalion Effect across 127,000 employees. The growth mindset provided the belief. The Pygmalion mechanism provided the transmission channel.
Tension
Incentive-Caused Bias
Incentive structures can override or corrupt the Pygmalion Effect. A leader may genuinely expect great things from a team member, but if the incentive system rewards short-term metrics rather than development, the expectation gets overridden by the structure. Jack Welch's differentiation system illustrates the tension: he held high expectations for top performers (Pygmalion), but the forced-ranking incentive structure created a zero-sum competition that channelled the Golem Effect toward the bottom 10% — regardless of whether those individuals could have developed under different expectations. The tension is structural: individual expectations operate within institutional incentive systems, and when the two conflict, the incentive system usually wins.
Section 8
One Key Quote
"You see, really and truly, apart from the things anyone can pick up (the dressing and the proper way of speaking, and so on), the difference between a lady and a flower girl is not how she behaves, but how she's treated."
— George Bernard Shaw, Pygmalion, Act V (1913)
Section 9
Analyst's Take
Faster Than Normal — Editorial View
The Pygmalion Effect is the most uncomfortable model in this collection — not because it's controversial, but because it implicates the reader directly. If your team is underperforming, this model says the first place to look is the mirror.
The research base is robust and the effect is moderate, which is exactly what makes it strategically significant. A meta-analysis of 479 studies shows an average effect size of d = 0.30. That doesn't sound dramatic until you compound it across every interaction, every week, for years. A manager who holds subtly higher expectations for one employee than another creates a performance gap that widens with each feedback cycle, each assignment decision, each allocation of mentoring time. Over a two-year period, the accumulated differential can be career-defining — not because the manager made a single biased decision, but because they made hundreds of small, unconscious ones that all pushed in the same direction.
The most common failure mode is not holding low expectations. It is holding accurate expectations that prevent higher ones. A leader who correctly assesses that an employee is performing at a B level, and then provides B-level challenge, B-level support, and B-level investment, has created a ceiling disguised as an observation. The Pygmalion research asks a disruptive question: what if you're right about current performance and wrong about capacity? What if the employee is producing B-level work because B-level expectations are all the system provides?
The organisational design implications are specific and actionable. Every talent system embeds expectations: who gets stretch assignments, who gets executive mentoring, who gets the high-visibility projects. Audit those allocations against your stated belief that talent can be developed. If 80% of development resources flow to people you've already labelled as top performers, you're running a Pygmalion Effect that confirms your classifications rather than testing them. The structural fix is deliberate: allocate a meaningful percentage of development resources to people you haven't yet classified as stars, and track whether the investment changes the trajectory. If it does, your previous classifications were measuring your expectations, not their capability.
Steve Jobs's reality distortion field is the extreme case, and it reveals the model's upper boundary. Jobs's expectations didn't just improve performance incrementally. They shifted what entire teams believed was technically possible. That's the Pygmalion Effect operating at maximum intensity — where the expectation changes not just effort but the target's own model of what's achievable. The boundary condition is that Jobs's expectations were specific and tied to concrete engineering targets, not abstract motivation. "You will ship this in six months" is a Pygmalion trigger. "I believe in you" without specificity is not.
Section 10
Test Yourself
These scenarios test whether you can identify the Pygmalion Effect when it's operating — and distinguish it from simple talent selection, competent management, or post-hoc rationalisation. The mechanism is subtle precisely because it produces outcomes that look like evidence of accurate judgment rather than evidence of expectation-driven influence.
Is this mental model at work here?
Scenario 1
A venture capital partner consistently backs first-time founders and provides them with intensive mentoring, board introductions, and operational support. Her portfolio outperforms the firm average by 3x. She credits her 'eye for talent.' A colleague notes that other partners who invest in first-time founders without the same support infrastructure see average returns.
Scenario 2
A new engineering director inherits a team previously rated as the lowest-performing in the company. She conducts a skills audit, finds significant capability gaps, and implements a six-month training programme. She explicitly tells the team: 'I think you can be the top team here within a year.' She assigns stretch goals, provides weekly coaching, and requests budget for external training. Twelve months later, the team ranks second out of eight.
Scenario 3
A CEO hires a new VP of Sales from a competitor, announces to the board that this is 'the best sales leader I've ever recruited,' and provides the VP with a larger team, bigger budget, and more executive access than the previous VP had. The new VP's first-year results exceed the predecessor's by 40%. The CEO presents this as vindication of the hire.
The foundational text. Rosenthal and Jacobson document the San Francisco experiment with full methodological detail, including the IQ data, teacher behavior observations, and statistical analyses that critics spent decades debating. The prose is academic rather than accessible, but the experimental design is elegant and the findings remain the empirical bedrock of everything built on the concept since. Read the original before the popularizations.
Rosenthal's earlier work on experimenter expectancy effects — showing that researchers' hypotheses influenced their data even with animal subjects — established the theoretical framework that Pygmalion in the Classroom extended to education. The rat maze studies remain among the most vivid demonstrations that expectations alter behavior below the threshold of conscious awareness. Essential for understanding the mechanism before the application.
Catmull's account of Pixar's creative process is the best case study of the Pygmalion Effect embedded in organisational architecture. The Braintrust chapters show how to build systems that hold high expectations for creative output while providing the support structure that makes those expectations productive. His principle that "early on, all of our movies suck" is a Pygmalion norm — establishing that every team can reach excellence through iteration, not that some teams are inherently excellent and others are not.
Nadella's account of Microsoft's cultural transformation documents the Pygmalion Effect applied at organisational scale. His reframe from "know-it-all" to "learn-it-all" changed the institutional expectation for 127,000 employees — from proving existing competence to developing new capability. The specific structural changes he describes (performance reviews, incentive systems, collaboration norms) show how expectations are transmitted through organisational design, not just interpersonal interaction.
Merton coined the term "self-fulfilling prophecy" in this Antioch Review essay, two decades before Rosenthal's classroom experiment. His analysis of bank runs — where the belief that a bank will fail causes the withdrawals that make it fail — established the broader theoretical framework that the Pygmalion Effect instantiates in interpersonal settings. Short, precise, and the clearest statement of how false beliefs create true outcomes through the behavioral responses they trigger.
The Pygmalion Effect — how expectations create self-reinforcing cycles that produce the performance they predict
Tension
[Narrative](/mental-models/narrative) Fallacy
The Pygmalion Effect creates outcomes that look, in retrospect, like evidence of innate talent — which is exactly the story the narrative fallacy prefers. A leader invests heavily in an employee they expect will succeed. The employee succeeds. The post-hoc narrative: "I always knew she was special." The actual mechanism: the leader's expectation shaped the conditions that produced the success. The narrative fallacy obscures the Pygmalion mechanism by rewriting the causal chain from "expectations created the outcome" to "I correctly identified pre-existing talent." This tension is dangerous because it makes the Pygmalion Effect invisible to the person generating it — they experience their prophecy as perception rather than creation.
Leads-to
Self-Fulfilling Prophecy
The Pygmalion Effect is a specific instance of the broader self-fulfilling prophecy — the pattern where a belief creates the conditions for its own confirmation. Understanding the Pygmalion mechanism in interpersonal settings leads directly to recognising self-fulfilling prophecies in markets (bank runs where fear of insolvency causes insolvency), technology (platforms declared "winners" that attract the developer investment that makes them win), and geopolitics (arms races where each side's defensive buildup is interpreted as offensive threat). The Pygmalion Effect is where most people first encounter the self-fulfilling dynamic. The broader pattern is where the concept becomes a systems-level analytical tool.
Leads-to
Extreme Ownership
If your expectations shape your team's performance, you own that performance — including the underperformance. The Pygmalion Effect makes extreme ownership unavoidable for honest leaders: the team's output is downstream of the leader's expectations, and the leader's expectations are downstream of the leader's choices about who to believe in, who to invest in, and who to challenge. A leader who complains about underperforming team members while providing those members with less attention, fewer resources, and lower-quality feedback has created the problem they're diagnosing. Recognising the Pygmalion mechanism leads directly to the ownership stance: you built the expectation environment. You own what it produces.
The Golem Effect deserves more attention than it typically receives. For every team elevated by a leader's high expectations, there's a team suppressed by institutional low expectations. Non-core business units, "legacy" teams, support functions — these groups often receive the Golem treatment at an organisational level: fewer resources, less executive attention, lower expectations for innovation. The underperformance that follows gets attributed to the team rather than to the expectation environment the organisation created. The most valuable application of this model may be identifying where in your organisation the Golem Effect is silently operating — and running the experiment of treating those teams as though you expected excellence.
One pattern recurs across every case study in this article: leaders who produce consistently strong teams are not better talent evaluators. They are better expectation-setters. They don't find better people. They create environments where people become better. The distinction is structural, not semantic — and it changes how you should think about every hiring decision, performance review, and resource allocation in your organisation.
Scenario 4
A teacher reads a student's file, sees low test scores and disciplinary notes from the previous year, and assigns the student to the remedial track. The student receives simplified material, less challenging assignments, and fewer opportunities to participate in class discussions. At year-end, the student's test scores are roughly the same as the previous year. The teacher cites this as confirmation that the track placement was correct.
