“In those days Caesar Augustus issued a decree that a census should be taken of the entire Roman world. This was the first census that took place while Quirinius was governor of Syria. And everyone went to their own town to register.”
It’s hard to fathom the incredible costs this first census imposed on the administration, and on the population. By stark contrast, collecting data today is so cheap that we do it without noticing.
Our precise geo-locations, driving behavior, facial expressions, our heart frequency, sleep quality, and typing speed – all these “features” are continuously sucked from our mobile phones, web browsers, and wearables. Thanks to 5G, we can look forward to a hundred times greater speeds for such data transfers.
Tech firms all over the world comb through this data for indicators of our willingness to pay for all conceivable goods and services. Despite what you read about in the newspapers, this has much less to do with robots suddenly acquiring human-like thinking skills (they haven’t) than with computers being very very good at rather banal but impressively large-scale data analysis.
This is the “Business of Big Data” – thus is the title of my Oxford lecture series and the eponymous book (which is co-authored with Uri Bram, as are all funny parts of this blog post). This business becomes more lucrative as the costs for data collection, storage, and computation fall.
Because these costs fall for everybody on the planet simultaneously, machine learning and AI will transform all industries within the next few years – whether the decision makers in those firms already know it or not. Nobody can hide from this wave of technological progress. Either you ride on that wave or you drown under it.
Many traditional firms don’t seem to have understood that new reality yet. In Germany, for example, the public discussion on how carmakers can catch up with Tesla focuses almost exclusively on when they should switch from combustion to electric engines. The first order of business, however, is not the drivetrain, but data, which Tesla drivers generate each time they touch the breaks, or the accelerator. Germans should be able to understand that Tesla isn’t worth ten Daimlers because it produces more and better manufactured cars; it’s worth ten Daimlers because it produces more and better data.
Europe’s more stringent privacy laws only present a small cost of doing business to the large tech firms. Even the recently-imposed record-level GDPR fines — to the tune of more than 100 million Euros — are but one thousandth of the market value of the violating firms. Why would they change their modus operandi in response?
Europe’s competition authorities — the German Bundeskartellamt in particular — once used to be pioneers in defining and enforcing limits to tech firms’ business models. Today looks very different. Whereas China writes aggressive new rules for their tech firms, strong enough to send those firms’ shares tumbling, and the U.S. sues Facebook and Google for various violations of competition rules, the European Commission’s idea to give itself new competition tools is lagging far behind. At the same time as Australia blocked Alphabet’s acquisition of Fitbit for the time being, Europe cleared the deal, with one main restriction: that Fitbit’s health data can’t be used to target Google ads. However, Fitbit’s health data can be used to inform your employer about the health-related risks of signing you on. (Your Fitbit should definitely notice an increase in your heart-rate as you read this.)
You don’t have to use a Fitbit yourself to be affected by these new practices; you only have to look sufficiently similar along other characteristics to your colleague who does use a Fitbit to be pooled in the same risk category. As such, your colleague imposes a “data externality” on you. She benefits from potentially attractive deals she receives in exchange for her data. You don’t enjoy the same deals, but you suffer the same consequences in terms of resulting discrimination in insurance and labor markets. That is to say: not only firms, but also individuals can’t hide from the wave that’s coming.
The good news is that entry costs are low, even for interested readers with non-technical backgrounds. After half a day of practice, my MBA students can write code that could replace human loan officers. Therefore, whoever hasn’t yet re-oriented her career towards complementing data-driven business models might want to give that idea some thought over the holidays; it might be the most important time you spend this year. We’re all very busy of course, but we do have some time that wasn’t available to Joseph and Mary: after all, we don’t have to make the two-week trek to Bethlehem for the census.