The next big controversy: Who owns the data?
By Michael Kanellos
A year ago an acquaintance was excitedly telling me about his company's new business plan at a mixer at the IoT World Conference. The company-a worldwide contract manufacturer-was going to become a data broker. The company would retain rights to the data generated by equipment it produced. It would then use this data to analyze and improve products or anonymize it and sell it to third parties. Right about then, another acquaintance from a brand-name manufacturer sidled up. "Like hell you will," the brand-name manufacturer bellowed. "You work for me. It's my data."
Expect to hear some variant of this conversation every few days for the next several years. Silicon Valley is currently bonkers about data brokerages, and for good reason. The company that coins the magically delicious way to harness data will achieve vast wealth and fame. Data is also a low capital business. OK, Uber is still losing money on every ride, but in industry there are already sterling results. Companies like Syncrude and Duke Energy have saved millions through service. Caterpillar has launched CAT Connect, a value-added service. Aurelia Metals achieved payback on analytics in 12 days by boosting gold production by 1 percent or about $750,000.
Unfortunately, the one thing many tech execs were counting on-that no one else would figure out the value of data before they locked it up-is clearly not going to happen. Manufacturers, end users, and others already know their information is valuable, and they are not going to relinquish it freely. No more IBM giving Microsoft free rights to resell DOS, or Intel getting the blueprints to the microprocessor back as an afterthought.
"Why should we be paying for the data? Why shouldn't OEMs [original equipment manufacturers] pay us, the operators, for the data?" said Gavin Hall of Petronas, the Malaysian oil company at a recent event. Petronas was showing off PROTEAN, a homegrown analytics tool. "Perhaps we need to change the business model."
Lockheed Martin's Don Kinard at this year's IoT World agreed. Lockheed might own the information and data regarding the shape of a wing, but the government typically owns the engine performance data. "They own it, mostly. They paid for it," he said. In some countries, sensitive data ownership will be mandated by law.
So does that mean digital transformation is dead? No, what it means is that the business model for sharing and using data will likely be different from what we have seen in the past. "Industry clouds" are an example of how the data industry might evolve. In industry clouds, companies share their (anonymized) data with an independent third party. Because they have access to data from multiple vendors, the third party can more quickly fine-tune its algorithms and achieve relevant results quicker.
The catch? Often, these industry clouds are owned by manufacturers and others generating the data. Clarient, a company that helps seek out financial fraud, was formed and is owned by six major banks. Sage Insights, incubated by John Deere, provides data services to farmers. Likewise, you will see large companies like IBM buy companies with large data stores to help animate their algorithms: Big Blue's Weather Channel purchase derived as part of its push into agriculture.
You will also likely see concepts come from real estate law. One of the first things you learn in real estate law is that ownership is rarely exclusive. Cities and counties own easements for water and sewer lines over a property. Leases often deliver more rights than a contract. A piece of property is a bundle of rights rather than a plot of dirt.
In the same way, manufacturers will provide access to their data in exchange for services, or for proportional interests in second-generation data produced through analyzing their original core information. Utilities will provide raw information to service providers for free but might be able to charge advisory services based on insights from that information for a fee. Consumer data is owned by consumers, says Bruce Edelston at Southern Company, but utilities get access to some of it, so they can perform their operations, while third parties will need consent. Metadata will be segmented from raw data and machine data from design.
The rule book will be massive, but in the end, it will make both OEMs and their customers happy.