How data helps uproot businesses and transform markets.
An essay by Victor Mayer-Schönberger
Internet companies aim to build your next car. The world’s largest jet engine producers create more revenue repairing their jet engines before they break than selling them. The largest book-store in the world knows exactly when customers stop reading a book. A start-up making a mobile phone app knows better than the world’s best language trainers how a Spanish speaker should learn English. Almost every sector today seems to undergo disruptive change, often with established players blindsided and destroyed.
This isn’t just ordinary competition at work. As former Xerox PARC head John Seely Brown says, Fortune 100 companies have seen their average life expectancy decline in recent years. Young, smart and agile seems to offer a better route to success than large and established. At the core of this disruptive shift, however, isn’t a new invention, but a new way of comprehending the world. This drives improved decision-making, but even more importantly, it drives the generation of new economic value through novel insights. And what makes this all happen is data – lots of it. Humans have always made sense of the world by using data. When we observe what’s around us, we gather data. But throughout human history, gathering and analyzing data has been difficult, costly, and time-consuming. As a result, we used as little data as possible, and devised smart ways to squeeze the most insights out of the least amount of data.
But what happens if collecting and analyzing data becomes cheap and easy? When we can capture data comprehensively, a lot of limitations will disappear: We’ll be able to see things without the fuzziness of extrapolation, and we can drill deep into details without losing precision. In short, we’ll understand the world at an unprecedented scale, creating economic value. For instance, online retailer Amazon gathers and analyzes not just our past transactions, but every small interaction we had on its website: what we looked at, what products we compared, and with what sequence, when we abandoned a shopping impulse, and when we followed through. The result is a huge dataset that leads to an extremely powerful (and patented) method of recommending products to us. The goal, as one Amazon executive once put it, is not to recommend lots of products, but ideally just one: the product that we will buy next. Amazon is pretty good at it. It’s said to generate about a third of revenues with its recommendation engine.
But “big data”, as this data-driven approach at disruption is often called, is not just improving marketing. Its much more important feature is to drive innovation by creating completely new products and services. For instance, a startup in California is aiming to uproot household financing. By examining thousands of data points for each loan applicant, the company thinks it can calculate the risk of defaults much better than ordinary banks relying on crude credit scores.
Another startup is analyzing billions of price points gathered from websites every day to predict inflation rates in real time, and offer that insight to third parties for a fee. Wherever business models are based on crude simplifications of the complexity of reality, a big data startup can swoop in and disrupt: from mobility and transportation to manufacturing, from finance to health care.
At the core of this data-driven disruption is a shift in how to treat data. In the past, companies employed data only after they had a concrete question to answer, and threw data away after they had used it. But data’s value is far larger than what can be extracted by using it just once for a particular purpose. As the cost of handling data plummets, the economic incentives to reuse data for novel purposes skyrocket. For example, a U.S. startup offers a nifty app to route drivers around heavy traffic on highways. It knows when traffic jams are building up in real time because every one of its over one hundred million daily users is also a sensor, automatically sending valuable data about location and speed back to company headquarters. The result is a great service for drivers.
But by looking at the huge pile of data, engineers discovered that heavy traffic around shopping malls on weekends correlates with strong revenues of the shops at such malls. So the startup teamed up with an investment company to trade stocks of retailers based on the sales predictions derived from traffic data. It’s a novel reuse of data they already had, extracting new value from it. Existing companies can do the same. International payment specialist SWIFT, for instance, is reusing transaction data of its billions of cross-border money transfers to predict the health of local economies in real time, selling these insights to others at a premium.
A different understanding
As the latent value of data incentivizes reuse, data disruptors are not only keeping data as long as possible, but also collect data whenever they can, not just when they already have a concrete use for it in mind. It’s the total opposite of how data has been handled in most companies to date. And it requires a different understanding of data’s role. This is something that companies often have troubles with.
Take the car industry, where a recent startup, Tesla, is giving established car manufacturers a hard time. Many people see Tesla as a producer of electric cars. What they miss is data. Tesla’s cars are also highly sophisticated data gathering platforms with dozens of sensors capturing billions of data points. The gist of that data stream is sent back automatically to Tesla. This way, Tesla can improve its cars, develop and then push out new software, and thus offer new features literally overnight through software updates (such as automated parking). But most importantly, Tesla knows when and how much you used your car, for how long, and when and where you recharged.
So they can devise the optimal network for charging stations (rather than, as with traditional cars, wait years for the market to find the optimal number and location of gas stations). And they can use the data to discover road safety issues, or help insurance companies to adapt their risk models to actual driving behavior. Put bluntly, Tesla is a data company, whose data collection platform happens to have four wheels. It is this distinct and novel perspective on what matters that enables Tesla and its peers to disrupt well-entrenched industries.
More rather than less
Obviously, it is better to have more data rather than less. That is the reason large data disruptors, such as Google, buy up successful data platforms. In the spring of 2014, Google spent about three billion U.S. dollars to acquire NEST, which manufactures a smart thermostat that remembers what temperature people want to have and when. What Google, of course, was after wasn’t a thermostat producer, but a nifty platform that collects valuable data. Google isn’t alone. Facebook, Apple, Microsoft, Twitter and many other large players, too, are buying datasets.
But data is also remarkably fluid. With cloud computing services the cost of storing and analyzing data has come down drastically. Computer power, which used to require huge server farms and large initial capital investment, is now a commodity. This has dramatically reduced entry costs for data startups. Unlike in the industrial age, these new entrants no longer have to build expensive factories. And unlike the Internet age, they don’t need to finance huge server farms, either. The consequence is an incredibly diverse and large field of new data startups in the U.S. They are trying out new ideas, and realize value from insights gleaned from data. Small companies with a few dozen staff can now collect and crunch billions of data points every day at low cost, and serve their innovative insights to millions of customers.
With a powerful new approach at hand and affordable tools at their disposal, data disruptors are poised to shape business models and uproot established industries. It surely will be fun to watch.