The business of search.

In what I’m sure will be oft quoted, Joe Kraus, one of the founders of Excite, has started a blog about entrepreneurship. The reason I feel the need to quote him is related to a comment about a very highly placed person at Microsoft vehemently indicating that search was not a business.

That was back in 1995, and I was working for a search company at the time. We were making money on ads through web site, but we also sold enterprise search technology. Search was a business then. Yahoo was new, as was Excite, but Alta Vista didn’t exist yet. Companies were building portals – web starting points – that organized the information for the users in the ways they were expected to want to use it.

Then search and portals fell out of favor as a business, probably because it wasn’t driving lots of revenue and companies couldn’t succeed based on ad revenue alone. Then of course the internet bubble burst, and lots of consolidation happened.

But about that time Google appeared with spectacular search technology and the ability to find the answers. Google even became a verb, and provided many users with the definitive way to find things on the web. So search became a business again, as did search engine optimization, as marketers tried to find ways to make search engines treat them better. Even Microsoft now feels the need to do substantial search research.

So search, the business that wasn’t now drives a lot of revenue, and Google, the company that brought it back, has really changed the way people interact with the web. In fact, it is probably easier to find a site I liked by using Google, rather that setting a bookmark. Search, for many, is the first point of entry when they are looking for information.

Yet Eric Brill, a senior Microsoft researcher said that the search business may fall apart as search gets better at finding things, forcing advertising revenue to decrease. He also suggested that it will get cheaper and cheaper to build a search engine, to the point where the technology is commoditized. It is surprising though that, given the resources Microsoft is throwing at search, they have yet to get it right.

Comments like this assume that one type of search, keyword for example, is suitable. However there may be many more types of search, such as context or latent semantic analysis, that prove to be useful in some combination with keywords. Other suggestions have included combining search with social networks to enhance the searches based on what my friends are searching for.

Search is still in its infancy, and there will certainly be numerous ways in the future for us to find what we need.