Unscaled

1

The Great Unscaling

Throughout the twentieth century, technology and economics drove a dominant logic: bigger was almost always better. Around the world the goal was to build bigger corporations, bigger hospitals, bigger governments, bigger schools and banks and farms and electric grids and media conglomerates. It was smart to scale up—to take advantage of classic economies of scale.

In the twenty-first century, technology and economics are driving the opposite—an unscaling of business and society. This is far more profound than just startups disrupting established firms. The dynamic is in the process of unraveling all the previous century’s scale into hyperfocused markets. Artificial intelligence (AI) and a wave of AI-propelled technologies are allowing innovators to effectively compete against economies of scale with what I call the economies of unscale. This huge shift is remaking massive, deeply rooted industries such as energy, transportation, and healthcare, opening up fantastic possibilities for entrepreneurs, imaginative companies, and resourceful individuals.

If you feel that work, life, and politics are in disarray, this transformation is why. We are experiencing change unlike any since around 1900, when, as I will detail later, a wave of new technologies, including the car, electricity, and telecommunication, transformed work and life. Right now we are living through a similar ground-shaking tech wave, as AI, genomics, robotics, and 3D printing charge into our lives. Artificial intelligence is the primary driver, changing almost everything, much like electricity did more than one hundred years ago. We are witnessing the birth of the AI century.

In an economy driven by AI and digital technology, small, focused, and nimble companies can leverage technology platforms to effectively compete against big, mass-market entities. The small can do this because they can rent scale that companies used to need to build. The small can rent computing in the cloud, rent access to consumers on social media, rent production from contract manufacturers all over the world—and they can use artificial intelligence to automate many tasks that used to require expensive investments in equipment and people.

Because AI is software that learns, it can learn about individual customers, allowing companies built on rentable tech platforms to easily and profitably make products that address very narrow, passionate markets—even markets of one. The old mass markets are giving way to micromarkets. This is the essence of unscaling: technology is devaluing mass production and mass marketing and empowering customized microproduction and finely targeted marketing.

The old strategy of beating competitors by owning scale has in many cases become a liability and burden. Procter & Gamble, with all its magnificent resources, finds itself vulnerable to a newcomer like the Dollar Shave Club, which can rent much of its capabilities, get to market quickly, target a narrow market segment, and change course easily if necessary. General Motors finds itself chasing Tesla. Giant hospital chains don’t know how to respond to AI-driven apps that target patients with a specific condition such as diabetes. The economies of unscale are turning into a competitive edge.

In my work investing in startups as a venture capitalist, unscaling has become my central investment philosophy. I fund or help build companies that can take advantage of AI and other compelling new technologies such as robotics and genomics to peel away business and customers from scaled-up incumbents. By adhering to the philosophy of unscale, our firm has invested early in groundbreaking companies such as Snap, Stripe, Airbnb, Warby Parker, and The Honest Company. Unscaling has also led me to help nonprofit organizations such as Khan Academy and Advanced Energy Economy, which are reimagining the institutions of education and electric utilities, respectively. My activities have given me both a broad and a deep view into unscaling, helping me to see the big picture.

The story of one of my companies, Livongo, provides insight into the dynamics set in motion by AI and unscaling. Livongo (“life on the go”) points to the way unscaling can drive down the costs of healthcare while increasing effectiveness. The United States spends more on healthcare than any other nation—$3.5 trillion annually, about 18 percent of gross domestic product. Citizens, corporate leaders, and politicians desperately want to get those costs under control but don’t want to lose any quality of our healthcare. AI and unscaling can help in part by turning healthcare more toward personalized medicine that can help prevent more people from getting sick in the first place.

In 2014 I helped Glen Tullman get Livongo off the ground, and he’s been the driving force leading the company as its CEO. Tullman was born near Chicago and in college studied economics and anthropology, a somewhat unusual background for a CEO who has made his mark in technology. After completing his education Tullman went on to run a couple of software companies and then in 1997 landed the job of turning around a struggling company called Allscripts. Founded in 1982 as Medic Computer Systems, Allscripts had bounced around for more than a decade as a maker of software for medical practices. Tullman and his team refocused Allscripts on software that lets physicians securely write prescriptions electronically. After spending two years improving Allscripts, Tullman took the company public at a $2 billion valuation, remaining as CEO until 2012.

I got to know Tullman when we both independently invested in a medical data analytics company called Humedica, which United Health Group bought in 2013. After the sale I wanted to continue working with Tullman, so we started hunting for an idea in the healthcare space.

Tullman was particularly interested in diabetes. For starters, it’s the fastest-growing disease in the world, and there are over 30 million people with diabetes in the United States alone. We also knew that diabetes is a manageable disease—people who are careful can live pretty normally. Still, for Tullman the disease is personal. “My youngest son was diagnosed with type 1 diabetes when he was eight,” Tullman says. “My mom has type 2. I’ve been surrounded a better part of my life by diabetes, and I was fascinated by how hard we made it for people to stay healthy.” People with diabetes typically need to buy expensive test strips, pricking their fingers several times a day and using the strips to analyze their blood sugar. And then it’s up to them to act on the reading. The whole procedure is problematic. The strips are expensive, people don’t like to poke themselves, and then, if blood sugar spikes or dives, the person can pass out or have a seizure. Longer term, the disease leads to other comorbidities like retinal blindness, kidney disease, and heart disease because people have a difficult time taking care of themselves effectively.

Tullman and I brainstormed ways to fix this, framing it in the following way: What if we could figure out a way to eliminate the hassle, to have people with diabetes spend less time, not more time, on their disease, to use all the wonderful innovations that we get from Silicon Valley, but do it in a way that the healthcare system could absorb? We ran across an inventor who had come up with a wireless glucometer—a way to measure blood sugar using a device that could send the results to medical professionals over wireless networks. Tullman acquired that technology, and we launched Livongo in 2014 to build a service for people with diabetes.

Livongo’s approach is simple and focused: it sends you a small mobile device that is both a glucose meter and pedometer (so it can track your exercise). It leverages cellular networks to communicate through the cloud back to Livongo software. As a patient tests glucose levels, and the Livongo device sends back data, Livongo’s AI-driven system gets to know that patient. If the system starts to see readings that point to a problem, it sends the patient a message to eat something, or to take a walk, or whatever might help. If the system determines there’s a serious problem, the patient gets a call from a health professional within a few minutes of checking his or her blood glucose.

As you might imagine, Tullman signed up his son, Sam, for the service, so the senior Tullman has personal experience. Sam, at this writing, is twenty-one and plays football for the University of Pennsylvania. Tullman recalls when he recently met Sam before one of Penn’s football games: “When I got there Sam said, ‘Hey Dad, I’ve got something great to tell you.’ I assumed it was about football or sports or girls. Instead he told me, ‘I had my first Livongo moment.’ I said, ‘That sounds good. What does that mean?’ Then Sam told me, ‘It was four a.m. last night, and I woke up. My blood sugar was thirty-seven.’ He’s six-foot-three and 240 pounds. He knows he can’t even stand up with blood sugar at that level. He said, ‘I didn’t know what to do. My roommate was out. I didn’t know whether to call 911. The phone rang, and it was Kelly.’ I said, ‘Who’s Kelly?’ ‘She works for you,’ Sam said. ‘Kelly is one of your CDEs [certified diabetes educators]. When we worked through it, Kelly had me crawl over to the refrigerator. If I passed out, she said that she would call 911, but it all worked out great.’ Sam then said, ‘I realized you weren’t in the business for yourself—you were in the business of making sure people don’t feel alone anymore.’”

Livongo created a new way for people to manage diabetes—one that would never have come out of the traditional medical field. It doesn’t replace the doctor, but it can help people with diabetes manage their lives so they need far less care from doctors or hospitals, which ultimately saves lots of money for individual patients—and in healthcare spending in society overall. But how is this unscaling?

Over the past four or five decades carbohydrate-heavy diets—pushed by mass-market production and mass marketing of cereals and drinks laced with high-fructose corn syrup—created an epidemic of obesity and, ultimately, diabetes. The medical profession lumped most people with diabetes into one of two categories of the disease—type 1 is genetic and type 2 is diet related—and prescribed a standard treatment. It was a classic mass-market medicine approach. So the healthcare industry scaled up to meet demand. It built diabetes centers and more hospitals and ran every patient, assembly-line style, through the same tests the few times a year they’d be able to visit an endocrinologist, whose schedule was packed. Yet for patients, sugar levels in between appointments can change, rising and falling to dangerous levels, and the disease can progress, adding more costs and more visits to bigger hospitals. People suffering from diabetes end up costing the healthcare system $300 billion a year in the United States alone. (It’s only going to get worse globally: within a decade China will likely have more people with diabetes than the entire US population.) The scaled approach can’t keep up with the growing number of people with the condition, and it fails to give people with diabetes what they really want: a healthy life.

In reality every person who has diabetes suffers from it differently, and the best way to treat it is different for everybody. So Livongo, a startup, was able to quickly build a product and offer it nationwide—and, eventually, worldwide—by leveraging tech platforms such as smartphones and cloud computing. The software and data from patients allows Livongo to offer more personalized care, making patients feel like they are a market of one, not one insignificant person in a mass market—and that makes for happier customers. The technology allows Livongo to nimbly compete against the diabetes-related offerings of giants such as Johnson & Johnson and UnitedHealth Group, winning a fast-growing subset of their customers and serving them at a profit.

Personalized AI-driven care can reduce the amount Americans spend caring for diabetes by as much as $100 billion just by keeping more people with diabetes well more of the time. Unscaled solutions can change the game and reduce healthcare costs by keeping people well. The nation can save money while at the same time making citizens healthier, happier, and more productive.

Livongo is one small example of what’s happening in sector after sector all over the world.

For more than a century size mattered. Economies of scale reigned as a competitive advantage. They worked like this: if a company spent a billion dollars to develop a physical product and build a factory, the amortized cost, at the extreme, would be a billion dollars to make one unit but only one dollar for each unit if the company produced a billion of them. So scale gave a company a cost advantage over competitors. It also brought other advantages, like an ability to negotiate for lower prices from suppliers and the money to blanket mass media with advertising. Once a company built massive scale and accumulated all its advantages, that scale became a huge barrier against competitors. A newcomer would need to build that scale—at great cost—to effectively take on a highly scaled incumbent.

In many ways scale was a net good for society for a long time. Scale was how the world achieved great things like global banking, air travel, widespread healthcare, and the internet. Scaled industries lifted more people out of poverty in the past fifty years than over the previous five hundred years.

The world we’re creating now will work differently. Small entrepreneurial companies routinely befuddle corporate giants. Serving niche markets of passionate customers now beats addressing mass markets of merely satisfied customers—because who wouldn’t prefer a product or service tailored just for them? We see this in now-familiar instances like when Uber upended long-established taxi companies or Airbnb out-innovated even smart hotel companies such as Marriott. We’ve known for a while that big companies and entrenched enterprises, which got accustomed to being business superpowers, need to fear two-person garage startups. But now unscaling is becoming systemic, taking apart whole sectors of the economy. The relationship between scale and success is flipping, as I’ll describe throughout this book. The winners will be those who exploit the economies of unscale, not the old economies of scale. This is a trend that began playing out around 2007 and will continue for another two decades.

Whether the kind of world that comes out of unscaling will be beneficial for most people depends on the choices we make, starting now. These will be big and difficult choices about the accountability of technology, the role of education, the nature of work, and even the definition of a person. We’ll need to make sure the unscaling revolution benefits society broadly, not just the wealthy or the technologically advanced. Those are huge responsibilities.

Although there are serious issues we must focus on, most of the news about unscaling and the technology behind it is positive. We are opening up new ways to solve some of the world’s great problems, including climate change and soaring healthcare costs. If we make the right choices, unscaling can reverse many of the ills mass industrialization has brought on, helping to create a future that works better than the past. But we’re just starting on this journey. To predict the full ramifications today of AI and unscaling would be like trying to predict the impact of personal computing back in the 1980s, when Microsoft pitched the then-outrageous idea of a computer on every desk and in every home. Yet unscaling is most certainly our future and the outcome of the development of powerful AI. To overlook or deny this would be irresponsible. Better to understand the coming outsized change, guide it, and reap its rewards.

The emergence of powerful artificial intelligence and the economic force of unscaling can trace their beginnings to 2007, when the Apple iPhone, Facebook, and Amazon Web Services—pioneering mobile, social, and cloud platforms—took wing at roughly the same time. As more of work and life moved online thanks to such platforms, the amount of data exploded. At first the explosion just seemed like more data that could inform business, and we even called it Big Data, as if that’s all there was to it. But Big Data turned out to have a higher purpose. It was the key to making AI, which previously had a long and tortured history of disappointment, into a force that will literally change the world. Other new technologies such as virtual reality, robotics, and genomics are also now breaking out, all driven by the power of AI. (Much more on all that in the next chapter.)

These technologies are becoming the foundations of global platforms. The world has been making platforms for generations—the interstate highway system, the internet, as well as mobile phone networks, cloud computing services, and social networks are all platforms. What is so important about platforms is that they do something so you don’t have to. A trucking company, for example, doesn’t first need to pave a road to transport a load of beer; an app maker doesn’t need to build a mobile network or app store to get its software to consumers. The more platforms we build, the less an individual company—or lone entrepreneur—needs to do by itself in order to create, produce, market, and deliver a product.

Now, for much of the twentieth century, even though some platforms, like the highway systems, were in place, most companies still had to build a lot of capabilities by themselves. That need gave rise to the vertically integrated corporation. Vertical integration means owning much of the “stack” that gets a product from an idea to a customer’s door. A corporation might own a lab to invent products, a factory that made parts for products, another factory that assembled the parts into a whole, a distribution system, and maybe the retail stores. It meant building huge scale, which took time and a lot of money. Once erected, these big-scale barriers to entry made it hard for newcomers to compete because it was supremely difficult to build all that scale.

By the 1990s, with the arrival of the widespread use of computers, the internet, and globalization, we began to see cracks in the foundations of vertically integrated corporations—the first intimations of unscaling. Companies discovered they could outsource entire functions and whole departments to other companies and even other countries—the connected outsourcing dynamic behind the sentiment that “the world is flat,” as author and New York Times columnist Thomas Friedman put it. The more platforms we built using new technologies, the more companies could rely on those platforms to do a job or task instead of doing it themselves. Barriers to entry kept falling. New entrants could be smaller and instead use platforms to seem big. Consider how upstarts like online eyeglass company Warby Parker or Jessica Alba’s consumer health and wellness goods company, The Honest Company, were able to quickly use the internet to sell to a global market to compete against established eyeglass makers and consumer products giants. The new era of startup-driven disruption took shape.

Around 2007 the creation of platforms accelerated. Smartphones and mobile networks allowed new services and products to reach almost anyone, anywhere. Social networks exploded and gave companies new ways to find people and advertise to them. Cloud computing meant any company could start a computing-intensive digital company without ever buying more than a laptop—just click a few settings on Amazon Web Services, enter a credit card number, and start selling to the world. At the same time, more businesses became digital—music, news, online retail, software as a service. Digital businesses especially could utilize platforms to instantly create, make, market, and deliver products anywhere in the world. As more business became digital, companies could collect more data about almost everything—customers, products, transactions, logistics—and that data made software and platforms smarter, creating an accelerating positive cycle. As this trend sped up—building more digital platforms, turning more business into digital business, and generating more data—we hit an inflection point. We started to reinvent the dynamics of business.

By 2017, ten years after the iPhone, platforms could do almost everything a business might need. One person could start a global company in her basement and compete against giants just by renting everything that major corporations used to need to build for themselves. Warby could rent computing power on a cloud service, rent ways to reach consumers via social networks and search engines, rent production from contract manufacturers, rent distribution of its glasses through FedEx and UPS, and so on. This is the essence of unscaling: Companies can rent scale. They no longer need to own it. And that changes everything.

Unscaling, it is important to note, is only beginning. As AI and other new technologies emerge and get developed into platforms, tiny entrepreneurial companies that have yet to be founded can serve customers in ways that big, mass-market companies could never imagine. Entrepreneurs will increasingly plug into platforms to build super-focused products that greatly appeal to niche markets, then find passionate customers and sell to them anywhere in the world—and do it all at profit margins that once only came with the old economies of scale. Big companies, bogged down by their own scale, will find it increasingly challenging to win against highly specialized, fast-changing products and services. That’s why the forces of AI and unscale are taking the twentieth-century economy apart and reassembling it in an entirely different way.

The emergence of the AI engine underneath unscaling is a grand technology story. In 2007 Apple introduced the iPhone. There had been smartphones before with brands like Blackberry and Nokia, but they didn’t have anything close to the iPhone’s capabilities. More importantly, Apple introduced the concept of the app. Over the following decade the mobile device moved from being an accessory to becoming the main way most people use software, data, and connected services—which, significantly, were hosted in the cloud. Before 2007—heck, even in 2010—cloud computing was a nerdy tech concept most people didn’t comprehend. Now most people know it as a handy reference for the fact that most of our data and the software we use sits on some computer in a gigantic data center somewhere, and we connect to it through wireless networks.

A number of other important technology platforms emerged around 2007 and took hold in the years after. When Amazon.com, which had already moved commerce online, launched Amazon Web Services (AWS) in 2006 it gave every software developer the power to launch a cloud-based software product and become an entrepreneur. Facebook was founded in 2004, but it wasn’t until 2007 that it turned into a platform, opening up so developers could build applications on it. Added together, 2007 can be called the origin point of an AI revolution, made possible by the combination of mobile computing, cloud computing, and social networking. In 2007 a little more than 1 billion people were on the internet; by 2016 it was 3 billion. Smartphone use had grown from a tiny sliver of society in 2007 to more than 2.5 billion people in 2016.

The new platforms made it possible for a new generation of entrepreneurs to reimagine how we do things and to build disruptive new apps. At first the platforms gobbled up cameras, flashlights, maps, publishing, music—all now on your phone or in the cloud, generating data. Because of the smartphone and cloud, Travis Kalanick and Garrett Camp could turn their frustration over waiting for a taxi in Paris into something productive by reimagining ride hailing through an app—giving birth to Uber. The concept of a cloud-based social graph gave the founders of Airbnb a way to connect people with places and to build in a system of trust. Two young brothers from Ireland, John and Patrick Collison, saw a way to use the cloud to offer developers a way to take payments anywhere in the world and founded Stripe. Evan Spiegel could reimagine communications as something more ethereal than it had been on the internet and started Snapchat.

In ten years a few important technology platforms completely transformed the way 3 billion people work and live. As content, community, and commerce continue to move online, we’re collecting data we never had before—data about what you buy, what you read, who you know, where you go. That data gives companies exciting new insights that can lead to even more new products and services. And it feeds the machine learning of AI software, which constantly gets better the more it is used because every interaction teaches it more about whatever the software is programmed to do.

I didn’t fully understand what was happening in 2007 and nearly missed it. Let me explain why—and how I came back around to understand the force of unscaling. The story goes back to New Delhi, where I grew up.

My parents were smart enough to recognize that there was no level playing field for a family of our means in India. They didn’t have the resources to provide a world-class education for me and my sister. So when my uncle sponsored us for a green card in the United States, my parents gambled everything to give my sister and me the opportunity to thrive in a more egalitarian society. This was the biggest risk we ever took as a family—and it probably helped me understand the value of taking risks based on a vision of how things can be, which is really what I do as a venture capitalist (VC).