Sunday, November 2, 2008

Google knows everything

One of the major fears about Google’s march towards total world domination is that it increasingly knows everything. As SJSU professor (and Google chronicler) Randy Stross wrote on page 10 of his recent book,

In 2006, an unknown person at Google prepared a PowerPoint presentation that included an offhand remark in the slide’s comments that Google at that point had collected only 5 percent of the information it seeks. Another slide’s comments emphatically added: “We plan to … get all the world’s information, not just some.”
(Although the slides were quickly removed, Stross’s footnotes helpfully note where the slides were archived by blogger Paul Kedrosky.)

Or, as Stross put it a month ago, “ this is a company that will know more about us than any entity — public or private — in the world.”

Normally we think about this information gathering in a personal sense, i.e., that of consumer privacy. So when I edit this (or another blog), I have to log on with my Gmail ID, and Google keeps me logged in — matching my searches to my personal ID — long after I’ve finished editing. (In fact, after a while I’m logged in for search but must re-login for blogging).

There’s another way that Google’s all-knowing is scary: its unlimited ability to gather market research in a way that no firm ever has — research where the data is all free, and also is entirely proprietary to Google. Data is not only addictive for business researchers, but academic ones, which is why economist Hal Varian left Berkeley for Google to become their chief economist and data miner.

On Friday, my friend David Wood of Symbian blogged about a presentation by Google manager Sumit Agarwal last week at a mobile phone conference in San Diego. A Google search suggests that Wood has the only coverage of Agarwal’s talk on the web, by either a journalist or a blogger.

The upshot? From its search and maps and other applications, Google knows what mobile users want — even better than the carriers. This is consistent with Google’s empirical, data-driven approach recounted in Stross’s book, the Varian interview, and many public speeches by Google managers and researchers.

For mobile phones, consumer wants and desires are not yet clear — handsets and mobile web browsers are just becoming good enough to provide (unlike in the WAP days) a good quality web experience. Thus, information about where mobile phone products and services are going is very scarce and valuable.

Google has all the data, and has concluded (according Agarwal via Wood) that mobile users seem to want pretty much the same information (at least in search) as static users.

However, through reverse IP lookup, Google knows how the wants of consumers differ by carriers, allowing it to observe the relative effects of the competing carriers strategies in a way that the carriers never could. One example was that when one operator instituted free data weekends, Google saw a huge spike in mobile search engine usage (that gradually increased weekday usage).

Wood summarizes other aspects of the information that Google shared on mobile phone users, and its implications for the industry. This is obviously only the tip of the iceberg, since Google has information that only it has — and will largely keep private for competitive reasons. Perhaps Google is willing to institute a commercial market research division to sell some of this information to the rest of us peons.

For someone who worries about Total World Domination, it seems like there are two bright spots. First, Yahoo got to mobile before Google and it still should have enough information to draw similar conclusions. Because that information is valuable, it could develop the same information — even if it only has 30% as many mobile users. What it would need is a smart, research-driven partner (not worrying about layoffs) of which only one company comes to mind.

The second encouraging tidbit was that Agarwal did his demo on an iPhone and a BlackBerry in addition to an HTC G1, even though he’s the product manager for Google’s mobile division. Even though the gPhone is shipping, Google remains client platform agnostic. (Wood obviously would have liked to have seen a Nokia N95 running Symbian OS 9.1, the preferred smartphone for much of Europe.

If Google will continue to emphasize the use of its services by all web-capable clients, that means that it is less interested in “tying” across its value chain than previous IT monopolists like Microsoft and IBM.

No comments: