Skrentablog

In large parts of Europe Google has a market share of 90% or more, like for instance in the Netherlands.

I think that a service which fulfills the search-need of people better than Google will be a competitor. That could also be social media services.

The second important thing is the link with search. When you think of search, you think of Google. This is a very strong association which other search services need to accomplish.

Speaking unofficially as a long-time member of the search quality team at Google, in my opinion the assertion that "Anna built the current version of Google" is not right.

That said, I really want lots of competition for Google, both to keep Google on its toes and to deliver better search results to users, so I hope that search start-ups such as Cuill and Blekko do well.

This isn't the first time someone has anticipated some company getting 90% market share. But it never happens. Skrenta is right, someone will come along with a better mousetrap as log as the gov't stays out of the way. Then Chelsea Clinton (when running for president in 2016) can preach about the dangers of free trade and capitalism, and can promise some new gov't program to help the poor afflicted Googlers.

Rich, how does Google have a network effect in search? Your argument is that Google *doesn't* have a network effect, isn't it?

They have a network effect in their ad system. It's like ebay. They have 500k+ sellers, far more than anyone else. Any publisher prefers to run Google's ads because they have more inventory, so publishers make more money. Advertisers prefer google's ads, because they have more viewers. Media buyers would rather buy from one place than many, especially if the price is "fair" -- set by an auction.

So they may have a network effect with their base of advertisers. You could develop a great new search property, but that wouldn't necessarily help you any with the advertiser problem. You'd be pressured to take Google's ads, as Ask does and folks have suggested for Yahoo.

There is another other factor you don't mention: Due to its high market share, Google has access to better, more complete data. Google knows more about what keywords specific types of users are using in their searches and which results they click on.

Adsense for Publishers, Google Analytics and the Google maps api, among other ancillary offerings, also provide Google with access to valuable data.

All of this provides Google an important competitive advantage, particularly in comparison with the smallest firms and startups.

That said, I don't think the barriers to entry are high enough to indefinitely maintain an 80-90% market share.

The problem with Yahoo and MSN, as well as many of the smaller startups is that they are competing head on -- they aren't offering anything fundamentally different, they aren't focusing on a narrower niche, and they haven't offered a product that is sufficiently innovative, or truly superior to Google's industry-standard offering.

Trying to beat Google at its own game is probably not a wise idea, given it's mindshare and data advantages.

To be successful, a startup will probably need to focus on a narrower niche, and do a much better job within that niche. Ideally, it would also innovate in ways that are hard for Google to duplicate, and which are closely matched to the specific niche that is being targeted.

Question: so now that Cuil is widely considered an industry joke, is this off the table as a credible competitor to Google?

I have some personal experience with this; even their spider (twiceler) is really quite dumb. It doesn't even make gzip compressed requests, which I consider a cardinal newbie sin for spiders.

You'd think spidering the web would be a known science by now, but man, I deal with so many utterly retarded spiders on a daily basis that it makes my head hurt. Sadly, Cuil is not one of the non-retarded ones.