July 21, 2009

Steve Lohr has highlighted an interesting study which tracked the use of "frequently repeated short phrases" "during [...] the 2008 presidential campaign". (*). As co-author Jon Kleinberg refreshingly declares, "you can see this kind of research as further elevating the role of sound bites". I would not mind being credited for originating a popular sound bite or two. But my purpose today goes in the other direction. How to go from fleeting news to lasting information.

Take Sarah Lyall, who echoes the Mail on Sunday's splash made by Sir John Sawers' trunks (**). No doubt "the photographs, and a great deal more" that Facebook hosted on the newly appointed head of the UK Intelligence Service made for top drawer entertainment. Lest we get carried away, Sarah Lyall concludes her news article with a sober quote according to which "the ennemies of this country do not wholly rely on [...] Facebook for their information". Besides the site has pulled down the incriminating evidence.

Behind this bit of fun though, there lies some biting truth. Forget Sir John himself. To get to him, work out his and his loved ones' personal networks. Better, project yourself 30 years from now. Who can know who his successor will be? No ennemy is rich enough to anticipate. Yet, intent to share it with everyone, Facebook is likely to have a page today on this future target. How hard can it be to hack and mirror Facebook wholesale?

News reporters work under severe objectivity standards. Lasting information may be struck down by the editor as too subjective. So, to let the readers on, good journalists will pick their quotations with care. Follow how Stephanie Clifford keeps a straight face when writing about the latest "effort to fend off federal regulation" by the advertising industry through self-regulation, this great oxymoron (***).

Read what "a lawyer for the trade groups" thinks of giving consumers "access to the data being collected about them". "An interesting concept [showing up] as bunch of ones and zeros [which] to a consumer would mean nothing". A call to the Center for Digital Democracy sees through this evasion. "Of course they can give the profile information, [which] can be translated". The reader is left to uncover the truth. With self-regulation, the only translation you will ever get of the million bits aggregated in your profile is the targeted advertisements themselves. Traduttore, traditore.

Free from following sound bites, those lucky enough to have a column, or a blog, can openly deliver lasting information. This is the case with Gillian Tett's "insight" into the Markit investigation (****). "Eight years ago, the fact that any group was producing communal data on credit derivatives marked real progress [...] even if distribution of that data was controlled". At issue is whether "Markit - and the dealers that own it - [abused] such a powerful position in relation to data flows".

The insight is in stressing how aggregators' mastery of data flows is "structural", i.e. can deliver unique competitive advantages in our Information Age. Not only in the financial industry. Indeed read what Professor Jon Kleinberg and his associates had to say on social networks (1). "Relatively subtle structural features are crucial in distinguishing between groups likely to grow rapidly and those not likely to". In other words access to micro-structures gives social aggregators macro-structural predictive power.

The valuation of Facebook is now reckoned by David Gelles to be $6.5B at the price paid by their Russian friend for common stock (*****). The hard fact is that this is based on getting users to surrender ever more of their privacy, nothwithstanding Chris Kelly's grandstanding. "We believe that when tools are simple, people are more likely to use them". I wonder. What is simpler than a spoon? Is it used to drain swimming pools? Yet the promised "granular privacy settings" will be set to public by default, requiring users "to explicitly choose to make each piece of content private".

David Gelles notes (******), the chief privacy officer of Facebook is "also running for California Attorney General in next year's election". Government of the lobbyists, pronaocracy has a great future. Chris Kelly, why not give a call to Vladimir Putin too? You may already share a friend and I just wrote your sales pitch.

While columnists speak their mind, they are not the only ones to provide information. Well researched news about one industry becomes information when read from a broader perspective. In the wake of Air France Flight 447, Christine Negroni reports how the airlines could insure crucial flight data survive a crash at sea (*******) and finds out "transmitting all that data in real time [is] not that easy", one obstacle being, in the words of a representative for pilot associations, that "if the data is not adequately protected, transmission could be an invasion of privacy".

Eprivacy is all about money, I wrote. Let me amend myself. It is also a power play. In our pronaocracy, isolated individuals should expect no protection while members of professional guilds make their voice heard. Compared to other data aggregators, the structural weakness of Facebook is to enable users to form vocal coalitions. Chris Kelly, make sure to keep strictly private any deal you strike with the Russian government.

Nothing though beats reading presentations by experts. What I see as information, others may dismiss as opinions. But isn't scientific truth nothing but a bunch of yet to be falsified hypotheses? Simson Garfinkel happens to offer us such an instance on eprivacy (********). His material is too rich and will be reviewed in next week's fillip. Let just point out two of his points.

Speaking of free services by Internet giants, he states "few [users] realized they themselves were the product these companies were selling". Hard not to take at face value. "The real problem is not that your information is out there; it's that it's not protected from misuse". This of course is exactly the pilots' position and one of my major complaints. Since poor security only raises risks for consumers but cuts the costs of data aggregators, economics theory guarantees market failure in the absence of proper regulation. This one sorry state is stable.

Bad news go beyond the lack of eprivacy. Per the news cycle study reported by Steve Lohr, the best sites anticipate the peak of a political sound bite by less than 24 hours (2). This really measures on the internet how fast news loses its freshness and hence its economic value. According to David Carr's figures (*********), "a tough blogger" is worth more than $600k a year. Contrast this to Simson Garfinkel, who earns an independent living in academia.

Work for The Mail on Sunday, face students on weekdays. Or else better band together to build a business model on reputation and eprivacy.

Philippe Coueignoux

• (*) ................. Key Words Tell The Life Cycle Of a News Bit, by Steve Lohr (New York Times) - July 13, 2009
• (**) ............... On Facebook, A Spy Revealed (Pale Legs, Too), by Sarah Lyall (New York Times) - July 6, 2009
• (***) ............. Industry Tightens Its Standards For Tracking Web Surfers, by Stephanie Clifford (New York Times) - July 2, 2009
• (****) ........... How Markit turned from a camera into an engine, by Gillian Tett (Financial Times) - July 17, 2009
• (*****) ......... DST in Facebook common stock deal, by David Gelles (Financial Times) - July 14, 2009
• (******) ....... Facebook's privacy changes to mimic its rival Twitter, by David Gelles (Financial Times) - July 2, 2009
• (*******) ..... Building a Better Black Box, by Christine Negroni (New York Times) - July 14, 2009
• (********) ... Privacy Requires Security, Not Abstinence, by Simson Garfinkel (Technology Review) - July-August, 2009
• (*********) . A Hollywood Blogger Feared Not by Starlets, but Executives, by David Carr (New York Times) - July 17, 2009
• (1) Group Formation in Large Social Networks: Membership, Growth and Evolution, by L. Backstrom, D. Huttenlocher, J. Kleinberg, X. Lan
  ..... (ACM SIGKDD Intl. Conf.) - 2006, downloadable from memetracker.org
• (2) Time lag of different media sites on reporting a story, by J. Leskovec, L. Backstrom, J. Kleinberg - 2009