You are where you live. That’s the message that Prizm, a market research tool from the company Claritas says. In the United States, groups of similar people tend to live near each other, and the size of these groups roughly corresponds to ZIP code boundaries. Claritas has identified 66 lifestyle types, and they geographically aggregated demographic information to companies looking to refine targeted marketing. This demographic information has proven to be very valuable. Try your zip code and I think you’ll agree they’ve done their homework.

The web magnifies this effect. On the web, companies know more about you as an individual as opposed to a member of a group. They track your individual activity in a way that physical companies couldn’t dream of. Search engines have managed to turn one activity, answering simple questions (search queries), into one of the most profitable businesses in the history of the world. The combination of wide appeal, high effectiveness, ease of use, low transaction cost, and an effective method of monetization (paid ads alongside normal search results) has produced hundreds of billions of dollars of wealth and many billions in revenue and profit over the last 10 years, Google being the leader.

Google has over 50 products (not counting Labs products), including dozens that aren’t variations on search. Google is the undisputed king of search, but none of its non-search products are #1 in their category (except YouTube, but that was an acquisition). Orkut is way, way behind MySpace and Facebook in popularity. Finance is behind Yahoo. Maps, despite redefining how online maps should work a few years ago, still isn’t as popular as others (it boggles my mind that people still use MapQuest). Talk is way behind AIM and MSN. Gmail is a fraction the size of Hotmail or Yahoo Mail. I’ve never heard anyone say anything positive about Checkout. The list goes on. How can Google, which completely owns search, be so dysfunctional in every other space?

Because it’s not trying to win those categories!

Google made $10.6 billion in 2006, a 73% increase from the previous year. That’s almost $1B/month (and at the rate they’re growing, it probably is by now). Nearly all of that comes from two products: AdWords (paid search ads) and AdSense (contextual ads on other websites). Simply put, for any of their other products to even be a blip in their revenue, it would have to be one of the most successful software projects in history! Google is smart enough to know this, and so what they’re really trying to do with these other products is to gather usage data and refine their picture of customers.

There are really only two ways Google can bump its revenue: increase Internet usage in general, and improve monetization of search. Google is indeed trying to increase internet usage (through things like the citywide wifi initiatives, bidding on the 700 MHz spectrum, and the Android smart phone platform) because they’re the biggest provider of one of the most common and most profitable activities on the web. But their ability to speed this up relative to general societal trends is small. On the other hand, if they can improve their search monetization even slightly, they make swimming pools worth of cash. If data mining users’ search queries against email messages against blog entries and comments could improve monetization by just 1%, that’s $100,000,000 a year. A couple percent improvement and pretty soon you’re talking about real money!

How could usage data from other products improve search? Here are a couple ideas off the top of my head. Right now (as far as I know), Google returns the same search results to everyone, and just counts on covering all interpretations of a query in the first few terms. If they ever do (do they already?) offer personalized search results, they could use other data to get a better idea of how to answer a query. For instance, geographical ambiguities would disappear (Aurora, IL vs Aurora, CO, Portland, ME vs Portland, ME, Springfield wherever), and terms with multiple meanings could be handled better (if you’ve never asked or talked about programming, then a Lisp is a speech impediment). Heck, location specific results could be tailored to information in your calendar that identifies where you are (ditto with mobile queries). Searches for products could return results from Base. I don’t know how advanced their Natural Language Processing is, but if they could understand your email and documents, then they could get to know you pretty well. They could go beyond what you ask to who you are and what you feel. Even if they don’t give personalized results, if they could create hundreds (or thousands or more) of “lifestyle types” (as opposed to Claritas’ 66), they could target general search results more effectively. I’m going to stop this now because it’s starting to scare me.

Plus, by being second or third (or eighth or whatever), Google can still get a useful amount of data without worrying about running into antitrust regulations. They (especially Eric Schmidt) have seen what happened to Microsoft when they bought or crushed competitors in different markets, and how much of Microsoft’s decision making in the late 1990’s was cramped by fear of the Justice Department. So it doesn’t matter if Knol is smaller than Squidoo, if Base is less popular than Craigslist, or if Docs never kills Office. If the data these apps collect on users helps improve their searches even the tiniest smidge, Google makes much more money than the company they’re losing to (well, maybe not with Docs). Interestingly, for the area where they’re most likely to run into antitrust issues (online advertising), they find it worth the risk to make a valuable acquisition (DoubleClick).

What does this mean for a company that Google copies? People used to be terrified that Microsoft would move into their space, because they would make a similar product and then just crush you, and all you accomplished was “doing market research for Microsoft”. With the internet, though, so much of the value is in the quality and size of the audience and the data they contribute. Many popular sites have technology that wouldn’t be that hard to duplicate, but their community is irreplaceable (take Google Video vs. YouTube for example). This means that if Google wants what you have, they have to buy you. Second, if they copy you, all it means is that you’ve found a useful online activity that they can mine behavioral data from. They will build something to get a representative data sample, but no matter how valuable your business is to you, it’s less than a drop in the bucket for them. Just consider your idea validated and keep your nose to the grindstone.