Data-Driven Development

Data, Data, Everywhere

A self-driving car, one of the most anticipated developments of the next decade, is expected to generate some 4,000 gigabytes of data for each hour of driving, according to chip maker Intel. To put it another way, just 3 million autonomous vehicles would generate, or consume, more data than the combined human population of more than 7 billion. Vehicles provide just one example of how data generation and use are growing explosively. Other machines generating an overload of data include satellites, environmental sensors, security cameras, and, of course, the ubiquitous mobile phone.

We are undoubtedly experiencing a data revolution in which our ability to generate, process, and utilize information has been magnified many times over by the machines that we increasingly rely upon. By 2016, according to IBM, some 90 percent of data that exists had been created within the previous 12 months, a rate of 2.5 quintillion bytes per day. Firms are increasingly finding hidden value in some of that data. Some 7 of the top 10 companies worldwide, by market capitalization, are data driven in that they create value primarily from the data they collect from or sell to their customers. The remaining 3 firms in the top 10 ­– in the more traditional financial services, energy, and health care sectors ­– also increasingly build data into their products and services or use it to improve them (table 1.1).

Table 1.1 Data hogs: Top 10 private companies globally, by  market capitalization, May 2017

With some justification, therefore, data has been called the new gold, the new oil, or the world's most valuable resource. Like oil, unprocessed data has relatively little value and needs to be mined, refined, stored, and sold on to create value ­– albeit in data centers rather than in oil rigs. But unlike oil, the quantity of data is ever increasing, not diminishing. Even though data is a nonrivalrous good, in the sense that my consumption of it does not affect yours, it is also excludable, which means it can be sold for profit, many times over. This makes it what economists sometimes call a club good, like privately owned safari parks or pay-per-view television. But because of the ever-increasing quantity of data, extracting value from it requires ever-greater computer power. Thus, the spoils from data-driven markets typically go to the largest players; those with the deepest pockets, the most users, the largest data centers, and most wide-ranging ability to collect and analyze data. Consequently, it is possible for market capitalization in companies like Facebook, Tencent, or Alibaba to exceed their annual revenue by 15 times or more and for the market capitalization of Apple and Amazon to touch the US$1 trillion mark in mid-2018, because investors view them as well positioned to take advantage of future data trends.