After our discussion of LinkedIn v. hiQ earlier this week, we received a lot of questions from clients and readers about the implications of the case. We thought it might be easier to compile some of these questions and our responses to continue the conversation about this important case.
Q. It seems unfair that hiQ is just taking LinkedIn’s data. It may be the case that LinkedIn didn’t create the profile data, but it certainly assembled and presented it. How do they not have the right to protect that?
This is a reasonable point, and was effectively what LinkedIn argued in its affirmative case. To use a metaphor, if the data that LinkedIn compiled was raw material, it is not appropriate for a company like hiQ to come in and take the finished products that LinkedIn presents. Yes, LinkedIn wasn’t the creator of the data it assembled, but that’s irrelevant to the economic process of creating a social networking site, creating a framework for users to input data, and hosting that data. A carpenter doesn’t create the wood she uses, but you can’t simply walk in and steal a chair from her showroom.
Yet this idea is misplaced. Data — say it with me now — is not the raw material of old, useable once in material form and forever transformed. Not to be overly philosophical (who, me?!) but this is a very Lockean approach to data: when a company mixes its labor with data it is transformed into a new species of property. Except not, or at least not in all cases and not for all uses. Personal data is absolutely not the same as property, given that it is a digital manifestation of personhood, protected as a fundamental right in Europe and elsewhere. And even in Lockean theory, you can’t transform someone’s personhood into property.
So there are limits to what data comprises a protectable property right, and there are even more limits to when the underlying data isn’t itself created by, or for, the company that hosts it. That’s why we always cringe when we hear “data is the new oil” or “data is the new gold.” Data not only isn’t a finite resource, it’s not even the same in all contexts and times. The value of data changes not necessarily because an outside metric says it is more valuable (although that can happen) but because of the circumstances in which we use, combine, and analyze it. An employee’s LinkedIn profile lists four languages she speaks. When she’s happy at her job in Tulsa, that information is worth something very different than when she’s unhappy and wants to move to Brussels. In other words, claiming a right in data at all times and for all purposes is, at best, an unmanageable task and, more likely, a contradiction in terms.
Q. What does it means for there to be a “circuit split?”
We mentioned the possibility of a circuit split in the article without really explaining it, so here’s a quick primer on federal courts. District courts are where trials take place, civil and criminal, and where the courts collect and assess evidence. When one side loses at the district court, they can appeal to a higher court, called a U.S. Court of Appeals. There are thirteen of them: eleven based on geography, one dedicated to federal claims like patents or copyrights, and one in D.C. which hears mostly cases related to the federal government. The geographic divisions can be huge (the Ninth Circuit is basically the western fifth of the U.S.) or relatively small (the Third Circuit is just Delaware, New Jersey, and Pennsylvania.
People often assume that federal law is a uniform thing, but that’s far from true. Unless and until the Supreme Court issues a definitive ruling on a subject, it’s open to interpretation. And, because the Supreme Court only hears 90 or so cases per year (and many of those cases are about civil and criminal procedure, rather than substantive law), questions about a particular law are unlikely to be heard at the Supreme Court, ever. So it is the intermediate appellate courts, the Circuits, where most analyses of laws reach their final adjudication. For some, that seems a little unsatisfying, because not reaching the Supreme Court seems like ending a season in the playoffs. But the reality is that the judges at the Circuit Courts are of such a high caliber and their rulings so respected that a ruling by a Circuit is much more like the Superbowl than the playoffs.
For that very reason, the Circuits will often disagree among themselves and, if the disagreement is fundamental enough, it creates what is known as a “circuit split.” If the Ninth Circuit says that the Computer Fraud and Abuse Act doesn’t apply to non-forcible commercial access to data, that’s a reasonable interpretation of the statute. And if the Eleventh Circuit says that the CFAA definitely applies to non-forcible commercial access to data, that’s a reasonable interpretation too. Either one of the Circuits could be right, but there can’t be two (or more) conflicting interpretations of a law in the country at once — which is why we said that scraping can’t be legal in Fresno but illegal in Fort Lauderdale. But that’s exactly the situation we have now. The only way to resolve the dispute is for the Supreme Court to hear and decide the matter once and for all. As Justice Jackson once said of the highest Court, “we are not final because we are infallible, but we are infallible because we are final.”
Q. Would it have been possible for hiQ and LinkedIn to come to some kind of agreement or signed a contract for the same data?
Yes, absolutely. This case is interesting (among other reasons) because it shows how things can go wrong in the absence of a well-defined data partnership. It’s clear that hiQ and LinkedIn worked closely in the past — the opinion relates that, at one point, LinkedIn happily collaborated with hiQ, took in data analytics from them, and participated in conferences together. A few short years later, they were enmeshed in litigation costing millions of dollars and occupying thousands of hours of work time, all to come back to the status quo ante.
Worth it? Hardly. This was a relationship gone sour, a sourness driven by a lack of guidelines or working rules. LinkedIn lacked the right to control hiQ’s data access under its existing framework, and so could have either completely changed how people access profile data (which would have been a costly disaster) or sought out a way to collaborate with hiQ. It’s simply a question of framing the issue. Could LinkedIn have shut hiQ out altogether? No, and so hiQ could have gone on scraping data, infuriating LinkedIn’s leadership. But, given their past relationship, it is far more likely that, had LinkedIn offered a partnership to hiQ — anything from a specialized license for some data sets, to a promise not to modify hiQ’s access rights, to a full-blown analytics sharing protocol — hiQ would have agreed. Instead of using its position (tenuous though it has now proven to be) to its advantage, LinkedIn went straight to litigation, and now has no leverage with hiQ at all.
In other words, back in 2015 before this litigation began, LinkedIn may not have been able to dictate the terms of hiQ’s access to data exactly as it wanted. But it definitely would have been able to negotiate an arrangement that offered it value. Now, with a federal appellate ruling that blesses hiQ’s business model, LinkedIn has nothing to show for its efforts but massive legal fees. A well-crafted data partnership doesn’t only protect your data or your business model — as the LinkedIn v. hiQ drama shows, sometimes a good data partnership protects you from yourself, too.