Sunday, March 06, 2005

Personalized search: who owns the information?

In an interesting post forecasting the rise of personalized searching, Greg Yardley raises an important concern about who actually "owns" the personal information/history upon which search companies will base "personalization."
Pretend it’s December, 2007. You’ve been using Google (or Yahoo, or MSN, or A9) as your primary search engine for the past two years. When these companies first introduced effective personalized search in late 2005, you might have been a little hesitant at first, but the benefits quickly spoke for themselves and you (like everyone else) became a regular user of the service. Two years later, personalized search is just taken for granted.

Now pretend your RSS-feed aggregator delivers an announcement to your personal internet appliance during your morning commute. A start-up has spent the last year indexing the internet and promises to deliver results to you using a revolutionary new algorithm, far better than anything currently available. Curious, you go to try the new engine out. But when you get there, you realize that this search engine doesn’t know you yet. When you search for ‘python,’ you get results on snakes, and ‘ruby’ brings up jewelry stores. It can’t anticipate your behavior; it makes no useful suggestions. The base search results are better - but without the personalized search history you’ve built up over the past two years, it simply can’t compete with Google (or Yahoo, or MSN, or A9). If only you could feed this new competitor your personal search history - but that’s owned by Google (or Yahoo, or MSN, or A9), and they’re not about to hand it over. Great for them - not so great for anyone itching to compete with them. Not so great for you, either.

Personalized search engines that make intelligent use of your search and browsing history are coming, and they will make search technology today obsolete. You will want to use one. Yet this will make you effectively beholden to the engine you use and give data to. I don’t want to have to depend on anybody in this way. That’s why I think we need standards for recording and storing personal search history now. Just like I can transfer my RSS feeds from one newsreader to another using OPML, I should be able to transfer my search history from one engine to another. Maybe I can’t persuade a site like A9 to actually delete my information, but at least I could keep a local copy and take my business to another engine.

Technorati tag: