第1章
When Scaling Breaks the Voltage: The Surprising Science Behind Success
Imagine a world where every great idea reached its full potential. Where innovative startups never failed, promising social programs always delivered, and transformative technologies changed lives as intended. Now wake up to reality. For every Netflix or iPhone that revolutionizes an industry, countless brilliant concepts fizzle when attempting to grow. This phenomenon-what economist John List calls "the voltage effect"-explains why scaling breaks most ideas, regardless of their initial promise. Having advised the White House, Uber, Lyft, and numerous Fortune 500 companies, List has become Silicon Valley's secret weapon for predicting which innovations will maintain their voltage when expanding. His research has saved companies billions by identifying scaling failures before they happen. In a world obsessed with growth, List's counterintuitive insight might be the most valuable: sometimes the most brilliant move is knowing when not to scale at all.
第2章
The Voltage Effect: When Great Ideas Lose Power at Scale
The concept of "scaling" has become ubiquitous in our growth-obsessed culture, yet it remains poorly understood. Whether you're expanding a business, rolling out a successful pilot program nationwide, or trying to spread a social movement, the fundamental challenge is the same: maintaining an idea's "voltage" (its impact and effectiveness) as it grows from small to large.
My journey into scaling science began unexpectedly in the late 1980s when I studied baseball card markets as an undergraduate. This early fieldwork taught me to see the world as my laboratory-a perspective that would later take me from Chicago Heights preschools to the White House, and eventually to rideshare giants Uber and Lyft. Along the way, I discovered that most ideas, regardless of their initial promise, lose voltage when scaled.
This voltage drop isn't random or inevitable-it follows predictable patterns. Some ideas appear successful in small tests but contain fatal flaws that only emerge at scale (false positives). Others work brilliantly in controlled environments but fail when confronting real-world complexity. The most heartbreaking are those that genuinely work but collapse under the weight of implementation challenges or runaway costs.
The stakes couldn't be higher. From climate change to educational inequality, our most pressing problems require solutions that work at scale. Yet most organizations approach scaling with dangerous naivety, assuming that what works small will naturally work large. This fundamental misunderstanding wastes billions of dollars and countless hours on doomed initiatives while truly scalable ideas languish unrecognized.
By understanding the science of scaling-identifying which ideas have genuine voltage and which are merely mirages-we can dramatically improve our success rates. Whether you're a startup founder, corporate executive, policymaker, or simply someone with an idea worth spreading, recognizing the five vital signs of scalability can mean the difference between transformative impact and frustrating failure.
第3章
The False Positive Trap: When Data Lies
Remember D.A.R.E., the anti-drug program that became the crown jewel of Nancy Reagan's "Just Say No" campaign? It seemed brilliantly effective in early studies, eventually expanding to 75% of American school districts and 52 countries. There was just one problem: it didn't work. Later research showed D.A.R.E. had no impact on teen drug use-and in some cases actually increased it. This wasn't just a waste of resources; it actively diverted support from potentially effective alternatives.
D.A.R.E. exemplifies the most dangerous scaling pitfall: the false positive-interpreting evidence as proof when it isn't. False positives appear everywhere, from manufacturing (misidentifying products as defective) to criminal justice (wrongful convictions) to medical testing. But they're particularly devastating when they lead organizations to scale ideas that were never actually effective.
I witnessed this firsthand at Chrysler, where initial data suggested a wellness program significantly improved employee health at one plant. Before rolling it out company-wide, my team conducted additional tests that revealed the initial results were statistical noise. This saved Chrysler from an expensive implementation of an ineffective program.
Why do false positives occur? Sometimes it's innocent statistical error-small samples aren't always representative of entire populations. Other times, cognitive biases like confirmation bias lead us to see what we expect to see, gathering and interpreting information that conforms to existing beliefs while ignoring contradictory evidence.
The bandwagon effect compounds this problem. Solomon Asch's famous conformity experiments showed how easily our judgment can be swayed by others' opinions. When influential people endorse bad ideas, they become contagious through social signaling. This explains why venture capitalists often pile into doomed startups or why policymakers champion ineffective programs with evangelical fervor.
The antidote to false positives is rigorous replication-repeating tests multiple times with similar populations before scaling. This principle applies broadly in daily life, from dating (meeting someone multiple times before committing) to medical diagnoses (seeking second opinions). For businesses, testing features in limited markets provides crucial data before full implementation.
Unfortunately, even replication isn't foolproof. The scientific community currently faces a "replication crisis" after finding that only 39% of high-profile experiments could be reproduced. This highlights the "file drawer problem"-failed studies remain unpublished despite their value to scientific knowledge.
Most disturbing are cases where false positives cross into deliberate deception. Brian Wansink, once the celebrated director of Cornell's Food and Brand Lab, saw his career implode when nineteen studies were retracted for falsified results. Elizabeth Holmes raised $700 million for Theranos's non-existent blood-testing technology. These "dupers" exploit our desire for simple solutions to complex problems.
The solution lies not just in independent replication but in creating environments where truth-telling is valued over confirming leadership's desired outcomes. Before scaling any idea, ask yourself: Has this been independently verified? Am I seeing what I want to see? And most importantly: What happens if I'm wrong?
第4章
Know Your Audience: Why One Size Never Fits All
When I joined Lyft after leaving Uber, CEO Logan Green proposed creating a membership program inspired by Costco's successful model. While I agreed with his focus on differentiation and customer loyalty, I argued that a subscription model wouldn't scale in ride-sharing's unique market conditions.
For membership programs to work economically, you need more "JoGoods" (customers who increase purchases due to discounts) than "NoGoods" (frequent users who simply pay less for what they'd buy anyway). Despite my concerns that most subscribers would be existing frequent riders who would cost the company money through discounts, we agreed to test the concept with a pilot program offering six different pricing structures to 1.2 million users.
The data confirmed my suspicion-NoGoods were nearly three times more prevalent than JoGoods, making the initial membership models unscalable. After analyzing the data, we determined an optimal structure with a 7.5% discount at $19.99 monthly, though Lyft ultimately launched "Lyft Pink" with a 15% discount plus additional perks.
This experience highlights a crucial scaling principle: different people respond differently to products and services across cultures, geographies, and socioeconomic groups. Like comedians who must know their audience for jokes to land, businesses must identify exactly whom their idea serves to assess scaling potential.
Selection bias can devastate scaling efforts. Consider an iron-fortified salt program that worked for adolescent girls but failed when scaled to broader populations with different physiologies. In contrast, the Nurse-Family Partnership tested their maternal support program across three demographically diverse cities, ensuring their positive results would replicate at scale.
The social sciences face a fundamental challenge: many supposedly universal findings about human nature come from WEIRD people (Western, educated, industrialized, rich, democratic) and don't apply across cultures. To create scalable impact, we must test ideas on diverse populations through natural field experiments rather than self-selected samples.
When your scaling hits audience limitations, consider broadening your reach. Gopuff exemplifies this approach-starting as a college-focused delivery service for convenience items, founders Rafael Ilishayev and Yakir Gola built a business that resonated perfectly with their demographic. But recognizing they'd eventually saturate the college market, they diversified their offerings to appeal to older demographics-adding pharmacy items for older adults and baby supplies for parents.
If product expansion proves too complex, consider optimizing production, distribution, or finding better-suited markets-or be prepared to pivot entirely. For nonprofits, the same principles apply-when Sierra Club fundraising showed men responded to matching gifts while women didn't, alternative approaches were developed. Similarly, when Parent Academy in Chicago Heights worked only for Hispanic families, program tweaks like all-day preschool were developed to serve Black and white families too.
Remember: scaling isn't about forcing everyone into the same box-it's about understanding the unique needs of different audience segments and adapting accordingly.
第5章
The Chef or the Ingredients: What Makes Your Idea Special?
When scaling, you must determine whether your success depends on unique human talent (the chef) or replicable elements (the ingredients). Jamie Oliver's restaurant chain initially succeeded by scaling the seemingly unscalable-a celebrity chef's cuisine. Unlike most high-end chefs whose unique talents can't be replicated, Oliver's success stemmed from simple recipes using fresh ingredients that other chefs could easily reproduce.
Yet his empire eventually collapsed, revealing that certain "ingredients" were more critical than realized: his managing director Simon Blagden's expertise in selecting franchise partners and maintaining company culture, and Oliver's own oversight. When Blagden left and Oliver's brother-in-law took over without restaurant experience, expansion became reckless, employee retention plummeted, and food quality tanked.
For any enterprise to maintain voltage at scale, you must identify what drives your success and preserve those elements. First determine if your "secret sauce" is the "chef" (indispensable people) or the "ingredients" (replicable elements). If it's people-dependent, recognize your scaling limitations-like my family's trucking business that remained small because it relied on the personal touch of individual family members.
If your success combines both people and ingredients, determine their relative importance and identify your negotiables versus non-negotiables-those elements your enterprise absolutely cannot survive without-then assess whether these critical components can truly scale.
Maintaining fidelity to non-negotiables becomes particularly challenging when organizations must serve "two masters"-like nonprofits with for-profit arms (such as AARP's billion-dollar insurance business) or businesses pursuing both profit and social impact. Research shows this dual focus often leads to mission drift as resources become spread too thin.
To combat noncompliance and drift, organizations must address economic and psychological incentives by making benefits more immediate while reducing compliance costs. Having founders or original innovators on implementation teams helps maintain fidelity, as does ensuring implementers understand the "why" behind the mission.
Digital technologies appear inherently scalable since code is infinitely replicable, suggesting non-negotiables remain secure at scale. But human behavior complicates this assumption. When Opower launched their energy-saving smart thermostat in California, anticipated energy savings never materialized despite perfect engineering tests. Our analysis of nearly 200,000 households revealed why: customers gradually undid the default settings, returning to wasteful habits.
The key insight? Adoption doesn't equal compliance. Technologies must be designed with human limitations in mind through extensive beta-testing with average users, not just tech-savvy engineers. User-friendly technologies like Instagram scale more successfully because they require minimal instruction.
To avoid voltage drops, we must flip traditional product development by starting with the vision of success at scale, identifying non-negotiables early, and testing with representative users rather than cherry-picking ideal participants. This "backward induction" approach means understanding constraints in real-world settings before scaling.
第6章
The Ripple Effect: Unintended Consequences at Scale
In 1965, consumer advocate Ralph Nader's expose on automotive safety hazards sparked regulations requiring seatbelts in all vehicles. Yet a decade later, economist Sam Peltzman discovered these safety measures hadn't actually reduced highway deaths. Why? Drivers felt safer with seatbelts, so they took more risks-an unintended consequence that became known as the Peltzman effect.
This phenomenon illustrates risk compensation-we adjust our behavior based on perceived safety levels-and reveals how seemingly free choices are shaped by hidden effects. In scaling contexts, this manifests as "spillover effects"-unintended impacts one group's actions have on others.
One dangerous form of spillovers stems from "general equilibrium effects," where disruptions in one area of a system trigger adjustments throughout until equilibrium is restored. This economic principle demonstrates how a single change can ripple through an entire system in ways impossible to predict at smaller scales.
When social dynamics interact with economic incentives, surprising spillovers emerge. My fundraising experiment for the Flossmoor Firebirds youth baseball team revealed this perfectly. Initially, I found that solicitors paid $15/hour outperformed those earning $10/hour when mixed together. Logically, I paid everyone $15/hour the next summer until budget constraints forced me to pay some $10/hour-surprisingly, these lower-paid workers performed just as well.
Further experimentation revealed the truth: it wasn't the higher wage that motivated better performance, but rather knowledge of wage disparity that demotivated the lower-paid workers through "resentful demoralization." When unaware of pay differences, both groups performed equally well. This demonstrates how transparency about compensation can create unintended consequences through social comparison.
Our peers profoundly influence our behavior in unexpected ways. At our Chicago Heights Early Childhood Center, we witnessed powerful spillover effects. Children in our treatment program, who received special curriculum for cognitive and non-cognitive skills, unconsciously transferred these skills to control group children through daily interactions and play. This effect intensified among neighbors, creating a rising-tide-lifts-all-boats phenomenon.
Within treatment groups, advanced children accelerated their peers' development, creating exponential benefits. Meanwhile, parents in the treatment group shared techniques with control group parents, who then sought additional educational opportunities for their children. These positive spillovers expanded our program's impact tenfold.
Network effects like these-whether in education, social media platforms like Facebook, or vaccination programs-demonstrate how interconnected systems can create parabolic growth at scale. Understanding these ripple effects is crucial for scaling successfully, as they can either amplify your impact beyond expectations or undermine your core objectives in ways you never anticipated.
第7章
The Cost Trap: When Economics Kills Great Ideas
Arivale promised to revolutionize healthcare through "scientific wellness"-a personalized approach combining genetic testing, blood work, microbiome evaluation, and one-on-one health coaching. Founded by pioneering scientist Leroy Hood and CEO Clayton Lewis, the company offered customers insights into their biological vulnerabilities along with tailored lifestyle recommendations.
Despite checking all four boxes necessary for scaling-evidence-based science, broad appeal, clear non-negotiables, and no negative spillovers-the company ultimately couldn't overcome the fundamental cost problem that doomed it. Initially charging customers $3,500 annually, Arivale reflected high overhead from laboratory testing and coaching salaries. Despite the potential long-term health benefits, this steep price severely limited demand.
After three years of struggling to attract customers, the company slashed prices to $1,200 annually in 2018. Even at this more accessible rate, they only enrolled 2,500 clients-not enough to achieve profitability. By April 2019, Arivale announced its closure, explaining that "the cost of providing the program exceeds what our customers can pay for it."
Adam Smith's concept of economies of scale explains why some businesses thrive while others fail. When average costs decrease as output increases (like with iPhones, movies, electricity, or pharmaceuticals), companies can charge lower prices while remaining profitable. Conversely, diseconomies of scale occur when costs increase with production, often due to resource scarcity.
Cost problems plague scaling efforts in both business and social sectors. While companies like SpaceX find creative ways to achieve economies of scale (like reusable rockets cutting costs by a factor of eighteen), social programs face unique challenges. The polio vaccine succeeded at scale because it was inexpensive to produce and easy to deliver, while a Zambian hovercraft-based vaccination program failed due to prohibitive costs.
For social interventions to scale successfully, benefits must outweigh financial costs. Even effective programs with high per-person costs ($50,000+) rarely scale, as funders inevitably choose interventions helping the most people per dollar spent.
When designing the Chicago Heights Early Childhood Center, my team initially wanted only the best teachers. We quickly realized this approach wouldn't scale-specialized workers become more expensive at scale, not cheaper. Using backward induction, we designed our curriculum assuming teachers with average abilities would implement it. By hiring teachers the same way Chicago Heights public schools would, with the same candidate pool and salary caps, we ensured our program could realistically scale.
"Perfection is the enemy of scale"-when replicating non-negotiables across various contexts, you sacrifice perfection but gain real-world viability. This approach allowed us to test a program that could eventually expand beyond our initial implementation.
No matter how brilliant an idea might be, if returns don't exceed costs, it loses voltage and becomes unscalable. Even with a proven concept, large audience, fidelity to core principles, and positive spillovers, runaway costs will kill scalability every time.
第8章
The Power of Incentives: Engineering Human Behavior at Scale
When scaling enterprises, we often overemphasize individual leadership characteristics while underestimating situational factors-a "correspondence bias." Getting incentives right is more important than individual character for scaling success. Well-designed incentives can shape behavior regardless of who is involved, making them infinitely scalable compared to relying on specific people.
This principle was demonstrated through my experience at Uber, where implementing a tipping system became necessary to rebuild driver trust after the #DeleteUber campaign damaged the company's reputation. Despite CEO Travis Kalanick's initial resistance to tipping, the need to retain drivers-a non-negotiable for Uber's business model-ultimately made the change necessary.
When Uber implemented tipping, the results surprised executives. While drivers initially appreciated the feature, their gratitude turned to disappointment when they discovered that although tips increased their earnings per ride, overall wages didn't improve because the new tipping option attracted so many new drivers that each got fewer rides.
Most shocking was that only 1% of passengers tipped on every trip, while 60% never tipped at all. This behavior revealed a fundamental insight: without social observation, people avoided tipping because there was no potential for reputation loss. Unlike traditional tipping scenarios where others can observe your generosity (or lack thereof), Uber's private tipping system removed the fear of social judgment that typically motivates tipping behavior.
Humans hate losses more than they value equivalent gains-a fundamental principle of behavioral economics established by Kahneman and Tversky. This evolutionary adaptation made sense when losing food might mean death while gaining extra food merely meant an easier tomorrow. Loss aversion extends beyond material resources to social standing. As inherently social creatures who evolved needing tribal cooperation for survival, humans monitor how others perceive them.
I harnessed this insight when the Dominican Republic faced widespread tax evasion (62% of companies and 57% of individuals). The government partnered with me to run an experiment targeting 28,000 self-employed individuals and 56,000 companies. Half received messages about jail time for tax evasion, while half were informed that tax offenders' names would be publicly disclosed. Both messages worked, generating over $100 million in additional tax revenue (0.12% of GDP).
The clawback approach proved remarkably effective at Wanlida, with productivity increasing by over 1% in the loss treatment group-a persistent effect throughout the six-month experiment. This approach works because of loss aversion: people are motivated by the discomfort of losing something they already possess, even when the bonus amount is relatively small. I replicated these results across different cultures and contexts, including with bean sorters in Uganda where productivity increased by a staggering 20%.
The clawback approach works beyond business settings. In Chicago Heights schools, where only 64% of students met minimum state standards, teachers in the "loss group" (who received $4,000 upfront but would return money if performance fell below average) saw tremendous gains in student test scores, with improved teaching performance persisting for five years after the experiment ended.
While some worry external rewards undermine intrinsic motivation, research shows that in communities with already low motivation, rewards can actually build intrinsic satisfaction without long-term downsides. When designing incentives for scale, focus on loss aversion, social norms, and immediate rewards rather than distant benefits.
第9章
Marginal Thinking: The Revolution That Changes Everything
In 2002, I joined the Bush administration as a senior economist, working in the Eisenhower Executive Office Building near the White House. My primary responsibility involved benefit-cost analysis of implementing large-scale policies, evaluating the approximately 50-100 "economically significant" proposals (those exceeding $100 million yearly in benefits or costs) among the 4,500 new rulemaking notices issued annually by federal agencies.
There I encountered a fundamental problem: government agencies were making decisions based on averages rather than examining the impact of the last dollar spent. Some EPA programs spent tens of millions to save one life while others achieved the same outcome for tens of thousands. This insight revealed that certain policies became dramatically less effective as they scaled.
The late nineteenth century "Marginal Revolution" transformed economics through the work of William Stanley Jevons, Carl Menger, and Leon Walras. They introduced utility theory to explain value, arguing that satisfaction from goods isn't static-the value of each additional unit (marginal utility) typically diminishes. This explains paradoxes like why diamonds cost more than water despite water being essential for survival.
After leaving government work, I found myself confronting the same marginal thinking problems at Lyft. During an executive meeting reviewing marketing expenditures, I noticed glaring inefficiencies-the last dollars spent on Facebook ads yielded only 1/50th the return of Google ads. My team's investigation revealed similar misallocations across the company's operations.
Unlike my government experience, where bureaucracy resisted efficiency, Lyft embraced these insights. CEO Logan Green implemented this approach company-wide, examining the impact of the last dollar spent in every department. When COVID-19 hit, this marginal thinking became crucial for survival, and later guided efficient resource allocation during recovery.
I witnessed the failure of marginal thinking firsthand as a teenager working at Wisconsin Cheeseman. As a forklift driver, I observed that when the company doubled its assembly workers, productivity didn't correspondingly increase. Management had budgeted based on average productivity rather than recognizing diminishing marginal returns-newer workers were less productive, and assembly lines moved only as fast as their slowest worker. This oversight contributed to the plant's eventual closure in 2011.
To avoid this cost trap, organizations should look for "weak spots" in areas with multiple investment or production levers. The key is collecting granular data across different strategies and time periods, rather than relying on aggregated averages. Effective marginal thinking requires experimentation-testing different combinations of resources and comparing results across diverse situations and populations.
When applying marginal thinking, we must avoid the sunk cost fallacy-our irrational commitment to resources already spent. I illustrate this with my experience helping the University of Chicago's fundraising department, which had abandoned their phone banking system because they incorrectly calculated costs by including the initial investment rather than focusing only on ongoing operational expenses. When I helped them recalculate based on marginal costs, they discovered phone calls were not only cheaper but more effective than mailers.
The bygones principle teaches us that past expenditures should have no influence on rational present decisions. Organizations must create cultures where people feel safe admitting mistakes and institute mechanics like rotating portfolios between employees to provide fresh perspectives. My advice is simple but difficult: cut your losses, let sunk costs sink, and be willing to upset your past self-even if that means quitting.
第10章
The Art of Quitting: Why Knowing When to Stop Is Essential for Success
I open with my personal journey as a talented collegiate golfer who dreamed of joining the PGA tour. After recognizing my skills wouldn't scale professionally, I redirected to economics despite facing rejection from 149 of 150 job applications after earning my PhD. Unlike golf, economics was where I knew I could make a mark. This pivotal career shift taught me that getting good at quitting is essential for scaling successfully, as people and organizations typically don't quit enough or soon enough.
Opportunity cost-the gains missed when choosing one option over another-explains why quitting at the right time is crucial. I illustrate this with my son Mason's baseball equipment purchase, where choosing a $200 bat over a $325 one allowed him to also buy a glove with the remaining $125. We often neglect opportunity costs until alternatives become visible, as psychological research shows our judgments typically focus only on explicitly presented information.
This "opportunity cost neglect" affects individuals, policymakers evaluating programs, and businesses alike. The most precious resource at stake isn't just money but time-when we spend time scaling one idea, we can't spend it on another. For people with ambitious scaling goals, opportunity costs are especially significant, as more time invested in an unscalable idea means more promising paths not taken.
When an idea shows diminishing returns as it scales, it's often time to quit or pivot. But we must consider not just whether the idea is scalable, but whether we're the right people to scale it. David Ricardo's 1817 concept of comparative advantage illustrates this principle: countries should focus on producing goods they can make most efficiently. This economic principle remains relevant today-Japan efficiently makes cars, Saudi Arabia produces oil, and the US excels in technology.
The principle extends beyond international trade to our individual pursuits. We should build careers doing what we excel at, but often we dedicate ourselves to goals where we lack comparative advantage. Many enterprises fail because they launch without understanding their comparative advantage or haven't developed one at all. The key is quitting at the right time to pivot toward your true strengths, as Twitter did when emerging from podcast platform Odeo, and as PayPal did when shifting from PalmPilot payments to online transfers.
Optimal quitting is extraordinarily difficult despite having the mental toolkit to rationally evaluate opportunity costs and comparative advantages. We resist quitting because we want to avoid the heartbreak of failure, especially after investing significant time and effort. Fear of the unknown also keeps us clinging to the status quo, though research shows people who make major life changes typically report being happier afterward.
A Freakonomics experiment involving over 20,000 virtual coin flips demonstrated that those who made significant changes were happier both two and six months later compared to those who maintained the status quo. Our cognitive bias toward ambiguity aversion makes us favor the known over the unknown, even when the known leads to disaster.
Thomas Edison exemplifies the power of giving up, discarding 10,000 low-voltage ideas to eventually produce high-voltage inventions like the lightbulb. Optimal quitting should be a strategic component of scaling rather than a last resort. Sometimes the wisest decision is to walk away from an idea-no matter how promising it initially seemed-to pursue something with greater potential for lasting impact.
第11章
Scaling Culture: The Hidden Foundation of Sustainable Growth
On Brazil's Atlantic coast lies Cabucu, a fishing community where men work in teams of three to eight due to choppy waters and large fish requiring collaborative effort. Fifty kilometers inland along the Paraguacu River, Santo Estevao's fishermen work alone, catching smaller fish from calm lake waters with lighter equipment.
Through field experiments testing trust, cooperation, and fairness, researchers found that Cabucu fishermen trusted others significantly more, were more trustworthy, proposed more equal offers in ultimatum games, contributed more to collective interests, and donated more to charity. Their daily teamwork had instilled more prosocial behaviors that carried over into other areas of decision-making-their culture scaled effectively.
This stands in stark contrast to my experience at Uber in 2016, where I noticed an employee holding back tears while colleagues ignored her-my first hint that something was amiss despite the company's "Data is our DNA" credo. By early 2017, Uber's culture unraveled spectacularly through scandals involving sexual harassment, stolen trade secrets, CEO Travis Kalanick berating a driver on camera, and software designed to evade regulators.
Though Kalanick appeared dedicated and engaged with employees, Uber meetings were "fast, fierce, and gladiatorial" where only the loudest voices prevailed. This "meritocracy" initially fueled Uber's expansion to 70 countries, but as the company scaled, the culture became toxic. True meritocracy requires trust that contributions will be fairly assessed, but Uber's version made it acceptable to "run people over" for profit, silencing introverts and driving away talent.
While meritocracy focuses on individual achievement, successful scaling requires trust and teamwork. At Uber, the incentive structure rewarded individual innovation rather than collaboration, creating siloed departments that couldn't effectively work together. As companies scale, the opportunity cost of not collaborating increases dramatically-in a five-person company, going alone often makes sense, but in a five-thousand-person organization, potential partners with complementary skills exist around every corner.
Netflix exemplifies "coopetition"-balancing competition with cooperation-through its culture of "freedom and responsibility" where employees aren't micromanaged but are trusted with significant autonomy. Rather than individual bonuses, Netflix ties compensation to company-wide success through equity options. While employees challenge each other's ideas, CEO Reed Hastings doesn't tolerate "brilliant jerks."
Diversity-across race, sex, age, ethnicity, religion, class, sexual orientation, gender identity, and neurotype-drives organizational success through better decisions, complex thinking, innovation, resilience, and higher profits. However, achieving diversity at scale presents challenges beyond just conscious or unconscious bias. My research with Andreas Leibbrandt revealed that well-intentioned Equal Employment Opportunity (EEO) statements in job postings actually decreased minority applications by up to 30%, particularly in less diverse cities and among highly educated candidates.
Even with the best culture, mistakes are inevitable-especially at scale where more employees, customers, and communities create more opportunities for error. After experiencing a frustrating Uber ride that left me late for a keynote speech, I conducted research revealing that riders who had bad experiences spent 5-10% less on the platform over the next 90 days. To address this, I designed an experiment testing different apology approaches with 1.5 million customers who'd had bad experiences. The results showed three key findings: how an apology is delivered matters (more contrite apologies were more effective); money speaks louder than words (apologies with coupons worked best); and too many apologies can backfire (apologizing three or more times in a short period was worse than not apologizing at all).
The implications of workplace culture reach far beyond organizational success. Whether in leadership positions or not, individuals have the power to shift culture toward trust and cooperation or toward distrust and selfishness. Research shows that organizational culture affects attitudes and choices outside the workplace, with evidence suggesting these norms correlate with economic growth and democratic quality. When scaling an enterprise, you're inevitably scaling values that can bleed into society and shape the lives of people you'll never meet.
