April 19, 2011

There is no denying the reality behind the expression "Internet time".

Five years ago, my first weekly fillip sang of the epic battle between Microsoft and Google. Three years ago, it became obvious Google had displaced Microsoft as the dominant player, making search the control point of the consumer computer economy according to my diabolo effect.

Whereas Microsoft enjoyed its victory over IBM unchallenged for more than a decade, see how Google already feels embattled. No sooner had it won the crown, it has had to repeal the attacks of Apple, which tries to create a new diabolo around iTunes, its digital content distribution platform.

In effect the "search engine" is becoming but one way to be the preferred "recommendation dispenser". Steve Jobs however has imprinted on Apple that control over a recommendation system means assuming sole recommender dictatorial status. Were Apple to win, even the most corrupted pronaocracy could hardly countenance universal private censorship without accountability. Still, comparatively more open, Google should prevail.

Unfortunately for Google, Facebook appears to be an even more dangerous foe. Google itself says so, as it has paid this upstart the ultimate compliment, imitation. It did it with Buzz, to little avail, it does it again with its "plus one" button.

Beyond the tactical merit of this latest squirmish, Richard Waters has penned an incisive strategic analysis of the new battle (*).

On one hand he writes of "the uncomfortable feeling that the search group is running out of ideas". One could have said the same thing when Google followed the iPhone with a smart phone of its own. But there are imitations and imitations. For the dominant diabolo, one thing is to match a platform on which to run the software which defines its diabolo, another to replicate a radical departure in its very definition.

In the former case, enlisting external help does not lessen one's control. Google has indeed widely licensed its Android software among phone manufacturers. Also the paramount feature is to support of the original diabolo, leaving time for completing the full functionality of competing offers.

Selling fleets of personal phones for companies to equip their employees, RIM does not compete directly with Apple. Reviewing its PlayBook tablet, David Pogue acknowledges "[it] is supposed to appeal to the corporate network administrators [...] because they can deploy PlayBooks without having to worry about security breaches" (**). In a vignette printed with Chris Nuttall's review (***), the PlayBook fits in a "large jacket pocket". In tandem with a Blackberry phone and despite its current shortcomings, the PlayBook may be the better tablet for company business.

Apple's strategic challenge stems more from iTunes than the iPhone. But for a competitor as rich as Google, well versed in proto-store operations with Google Books, opening and running an "app store" is not that difficult, especially with looser rules on third party developers. Enabling social dialog and turning it into a source of recommendation, what social networks are all about, is something else. Ask Microsoft. Extending Windows with a me-too browser was enough to defeat Netscape but it has never caught up with Google in search as a business.

On the other hand, Richard Waters highlights that Google "undeniably has the staying power". It also "has some prime assets [...] in place for its social push", "its own, implicit networks of relationships to mine", starting with its total control of Gmail address books. Yet he acknowledges that "Google needs more users to set up profiles and add personal information about themselves".

While I like Richard Waters' analysis, add one layer to it for a deeper understanding.

If today's winning diabolo is built around a universal recommendation system, recommendations are expected to be personalized. Why not then apply the intuitive framework proposed by the World Economic Forum ("WEF") and Bain & Company to categorize personal data? It casts the difference between Google and Facebook in a stark contrast. Both companies have aggregated massive amounts of data on each of their users. But most of Facebook's data is, if not volunteered, at least disclosed by its users, whereas most of Google's data is observed, either when users search for information or exchange emails or when Google taps into consumer marketing databases, courtesy in part of past activities at DoubleClick

Google's staying power comes from monetizing the demand it observes as users search the web. Facebook's fame is to turn user disclosures into social pressure to wield as a marketing tool. But what advertising campaign manager wouldn't give for an integrated approach towards each user? Isn't indeed Richard Waters' implicit conclusion that the winner must offer both sources of personal information?

If so, the winning diabolo might not be based on the recommendation system itself but on its ability to collect and refine personal data. From this perspective competitors should be judged on three factors, the data refining process itself, its visibility and its legitimacy.

Besides user demand and social pressure, data refining should include the observation of purchase patterns, which feeds what most shortsightedly consider the only Internet based recommendations, the type illustrated by Amazon, Netflix and Pandora. The risk however is to overlook data refining is a generic process whose four steps should be better studied and optimized for their own sakes, i.e. raw data acquisition, profile aggregation, making recommendations and finally delivering them at the opportune time.

When I search for "car insurance", all four steps occur at once, a directness best mastered by Google. But in general data acquisition and recommendation delivery are best performed on a decentralized basis, focusing the diabolo on profile aggregation and recommendation making.

Recommendation making itself includes the third WEF and Bain category, inferred data. However powerful the algorithms behind the "if you like this, you will like that" approach, however persuasive is the social pressure "if all your friends like this, you should like it too", what advertisers really need to know is whether "this person wants my offer" or not. Any method whatsoever to reliably infer the answer to that simple question is valuable.

This is why the real pinching point in the diabolo may well be profile aggregation, the central role in a system open to all raw data contributors, whether observed or disclosed, all inference makers, whether through software or domain expertise, and all opportunity creators, whether their content entertains, inform or answer questions.

"The question is not whether Google "gets" social - it is whether it is as attuned to, and respectful of, its users' interests as it claims".

What Google claims is not necessarily the truth but Richard Waters' parting statement again is quite on the mark. A recommendation process which lacks transparency, like Google, or legitimacy, like several of Facebook's past attempts, is not sustainable over the long term.

Allow me to repeat myself. A platform which enables profile aggregation to occur under the sole control of each profile owner, which therefore combines visibility and legitimacy with true eprivacy is the best diabolo there is and the only believable challenger to Google's current dominance.

Nokia and Publicis take notice. Meanwhile Google should consider hiring Steve Ballmer. Has Larry Page any experience milking aging cash cows?.

Philippe Coueignoux

• (*) ..... Google bidding to go plus-one better than Facebook, by Richard Waters (Financial Times) - April 14, 2011
• (**) ... Much Missing In PlayBook, by David Pogue (New York Times) - April 14, 2011
• (***) . Enough to give Apple the Pip? , by Chris Nuttall (Financial Times) - April 15, 2011