May 17, 2011
"New ways to exploit raw data may bring surge of innovation". This type of title does not ordinarily brings good news for the defenders of eprivacy, the right to our own personal data. But Steve Lohr's article (*) pleasantly surprised me. While progress promises to "exploit Internet-scale data sets to discover new businesses and predict consumer behavior and market shifts", "one hurdle is a talent and skills gap". What a relief!
What prompted Steve Lohr's call to rally "math majors" is a new report from the McKinsey Global Institute (**). Bain & Company, a competitor of theirs, has recently advocated to look at "personal data [as] a new asset class". When two leading consulting firms address the same subject, it is not a coincidence . Beyond mere words, their very actions prove them to be under market pressure to provide their clients with relevant guidance.
I took issue with Bain's reluctance to deal with the need to clarify our property rights as personal data providers but, together with the World Economic Forum, Bain put forward the welcome and revolutionary notion companies "will need to start paying for end user data". From this perspective, the McKinsey report is a step backward. Steve Lohr confirms it "does not deal with privacy issues in detail".
This is not to say that McKinsey ignores the subject. They do list "a set of policies issues that will become important, including, but not limited to, privacy, security, intellectual property and liability" and the word "privacy" is used almost fifty times within one hundred and thirty six pages.
Yet the very fact that its report is professionally researched makes McKinsey all the more to blame for reinforcing three myths against eprivacy.
One reason behind the popularity of these myths is their canny capacity to shift most costs on those too weak today to defend access to their data and on the future generations, conveniently tasked with the promised provision of safe sustainability. This should not be a surprise, knowing how prone society is found to generate asset bubbles and whom it burdens with the pollution of its environment.
- Big Data is a good in itself. Enable its "effective use" and it will "transform economies" and deliver "productivity growth and consumer surplus"
- Big Data is based on inescapable "trade-offs between privacy and utility", between the "risks of sharing data and the benefits of sharing data"
- educate consumers, "helping [them] to understand" these trade-offs, and any "third-party [which] might have not considered sharing [its data]"
Behind the lure of taxing one's fellow citizens, though, lies a genuine difficulty. These three myths are linked into a bundle to be broken as a unit, a tall order according to La Fontaine (1). McKinsey devotes a whole chapter to "personal location data", let us use it in our proof.
No doubt this is Big Data. "Navigation devices are a major source of data volume because they update their locations so frequently". But this does not make it bad in itself. In fact with the right attachment, a smartphone could readily double as a justice auxiliary. Recall Kara Scannell reports that, pending sentencing, "the Manhattan federal judge overseeing the [Rajaratnam] case ordered him to wear an elecronic monitoring bracelet" (***).
So if Big Data is not a good in itself, it is only because its proclaimed utility is wanting, because it is not always necessary to generate Big Data and trade off its sharing against privacy for society to achieve the same benefits.
This means the whole project of convincing consumers of "understanding", i.e. accepting the associated trade-offs, is a confidence trick whose goal is to hide the harm done until it cannot be reversed.
But how can we reach such sweeping conclusions without building an alternate reality? Without sharing, could we for instance implement "location based-marketing", a major application of "personal location data"? Consumers normally know where they are, of course, and have always been tempted by telling signs or attractive displays to step into the shops they happen to pass by. Noting that such time honored behavior collected no personal data and breached no privacy is not to glorify the past as idyllic but to take it as an inspiration for the future.
As the density of local shops increases, the visual advertising of the past becomes an ugly jungle of competing signals with decreasing efficiency. Since cellphones carried in consumers' pockets also know where they are, society will gain if local marketers use them to attract nearby prospects.
But to track all consumers, resell this data live to marketers and filter their ads to beam the most relevant ones to approaching consumers is only one way to go about it. The alternate way drops consumer tracking as unnecessary as it is dangerous. Why not instead label each local ad according to merchant location and targeted prospect profile and package and broadcast them all to the cellphones of all consumers in a greater marketing area?
If a smartphone is to live up to its billing, can't it filter all these ads itself against its current position and its owner's profile and, as before, present the most relevant ones? Rather than a monitoring bracelet, can't the smartphone enhance the consumer's known ability to quickly sift through a mass of competitive claims on his or her attention? As digital storage boasts go, can't a cellphone hold a dynamic Yellow Book for its current area code?
This alternative reality sketch is not to be confused with an operations manual (2). But it is enough to open a debate which would expose the three myths propagated by the McKinsey report for what they are. Instead
I share with Bain the conviction that personal data as a new asset class offers opportunities to entrepreneurs. I strongly disagree with McKinsey when it assumes it requires us to share our own personal data and implicitly endorses Mark Zuckerberg's faith that sharing with data aggregators expresses social solidarity. This is not how to best balance the two active forces of progress, spirit of enterprise and sense of solidarity.
- as harmful noise makes most of Big Data, increasing data collection may end up decreasing the signal to noise ratio (3).
- the only inescapable trade-off is in deciding who controls private data as it supplies utility, individual owners or collective aggregators
- educate consumers on the mounting costs of retrofitting privacy-conscious policies once an economy is shaped around data aggregators
Solidarity is rather born by decentralization, which dilutes the pernicious economic rents levied by data aggregators by virtue of their central power.
- (*) ..... New Ways to Exploit Raw Data May Bring Surge of Innovation, a Study Says, by Steve Lohr (New York Times) - May 13, 2011
- (**) ... Big Data: the next frontier for innovation, competition and productivity,
........... by James Manyika, Michael Chui, Brad Brown, Jacques Bughin, Richard Dobbs, Angela Hung Byers (MKinsey Global Institute) - May 2011
- (***) . Rajaratnam found guilty, by Kara Scannell (Financial Times) - May 12, 2011
- (1) for those of my readers who read French, the story is in Le Vieillard et ses Enfants
- (2) for more details, see US Patents Number 6,092,197 and 7,945,954 and US Patent Application 2009/0076914. For an implementation, please contact ePrio Inc.
- (3) "math majors" know that of course but their employers have selective hearing, as when big banks hired them before 2008 to create credit derivatives