Thursday, December 4, 2008

The New Commodity

When you watch TV, you get to enjoy your show because they fill it with commercials. This monetizing model can be seen in radio and even some areas of the Internet too. Ads! But I submit to you that while money will still make the world go around, knowledge, structured knowledge, will make the money change hands. All the technology is heading that way, and if you aren't on board, you're not going to get into the money flood that's coming.

The demand is out there for knowledge - this broad "I don't know" type of idea, where people know they'll want to know 'things about stuff' but they're not sure to what end just yet. Google is poised very well for when the dam breaks. If you strip out the Google Ad Words revenue stream, you might think that Google has no way to pay the bills. But the truth is that they probably house more knowledge about people, demographics, chatting, email, images, video, and everything else, than any other library, corporation, or other KB group out there today. That is to say, that if knowledge is power, Google is will be the world's new king.

What is interesting, is how they do it. They give their software away. It's easy to use, there's no barrier to "buy" because you don't have to buy it, you don't even have to install it, and if you already have a Google account, you don't have to sign up for anything extra. Just point your browser at a URL and away you go. So now you're using their product giving them plenty of knowledge nuggets with each click and keystroke to be a data-miner's wet dream. And because of that, they can release new features or rearrange their UI to match what people want.

The really brilliant part is the platform and the price. Since it's all SaaS, they can push out an update whenever they want. And Google knows exactly what version you're running (hint: it's the same version as all of their other users). The price ($0) means that you're going to be less likely to complain if the server goes down, or the behavior is a little quirky. Sure, you might file a bug, but you're not going to be able to threaten to take you business else where, because you're just one person who has zero cash bargaining power. Yes, you have some value to Google, they want to know what you know, and they want to know what you don't know you know. But that knowledge you have is just a tiny slice of the millions of other users they have.

Other companies are starting to realize what's going on. JobScore is one such company that gives away the applicant tracking system they make, so that they can sell the knowledge that recruiters input into their system to other recruiters. They make the same old software that anyone else can make, but they sell the knowledge that only they had the foresight to see as a commodity. Now recruiters get better data thanks to the work that other recruiters were doing anyway. Here in the Bay Area, I see start ups all over the place with the same kinds of business models, though some don't have a market to sell their knowledge back to, they hope to be gobbled up by a knowledge giant like Google.

This idea that knowledge is a commodity is really what sets the big G apart from Microsoft and Yahoo who really seem to only fill a direct need with software, but neglect that they can get something extra from a user's experience. Ultimately, this leads us to write software that lets us solve hard problems. Things that fall into a category of 'dynamic pattern recognition' are great problems for humans to solve, but hard for machines to recognize. "Are these two faces the same?", "Does this MP3 contain profanity?", "Would a mother be upset if her child saw this picture?". One solution is to bring in all kinds of math-minded people to set up all kinds of algorithms that run in O(n*log n) time and gets it right only slightly more than if it just randomly guessed. Or, you can write software that asks a human "Hey, what do you think?" You have to work with humans though, humans who are sometimes lazy, who lie, who game the system for their own benefit.

Google has a happy set up for solving just such a problem. Actually they have millions of solutions - their users. By asking a lot of people the same question, they quickly can create some easy math to work with "What does the majority say, and who are the outliers?" You can ask your 15 year old son to compute the standard deviation of most any data set, but it is the kind of work that machines were built to handle.

Knowledge gathering is really quite simple. Step one: Write software that is easy to use and gathers knowledge. Step two: Get users to use it, thereby gathering knowledge. Step three: Profit! It's the bread and butter of Google, and that business scheme is rapidly being adopted in the start up world as well.

No comments: