Here is a bunch of links about Google that I have had lying around for a while – trying to think about the first one and to what extent Hal Varian is right about Google not having a network externality competitive advantage. I think he is wrong, but why is hard to articulate.
So, here goes (note that Google, rather nicely, includes a list of links to each blog post, which is fodder for further discussion):
- Hal Varian: Our secret sauce, arguing that Google’s competitive advantage is due to experience and innovation, not network externalities.
- Tom Evslin: Sitemaps and how the rich get richer: Essentially, Google has an advantage because they are the biggest and people adjust their web sites to the Google engine and its various algorithmic quirks.
- Hal Varian: Why data matters. Brief overview of search and PageRank.
- Hal Varian: How auctions set ad prices. Brief explanation of Google’s auction system for ads. One interesting effect, not mentioned here, is that the more precisely the user can describe the targeted population, the lower the ad price – thus, Google has both an incentive to make targeting imprecise (to have enough actors competing for a particular keyword/target) and an incentive to make it precise (to increase click rates).
- Marissa Mayer: A peek into our search factory. Various presentations, with notes, about the infrastructure underlying Google’s various offerings.
- Udi Manber: Introduction to Google search quality. Overview of what Google does to fight spam, increase precision, and other things. (Reads like a transcript of a talk.)
Here are two articles that everyone trying to understand Google should read (come to think of it, this blog post is starting to resemble the layout for a class):
- Brin, S. and L. Page (1998). The Anatomy of a Large-Scale Hypertextual Web Search Engine. Seventh International WWW Conference, Brisbane, Australia. (The classic on PageRank.)
- Ghemawat, S., H. Gobioff, et al. (2003). The Google File System. ACM Symposium on Operating Systems Principles, ACM. Description of the architecture of Google’s index, a file system geared for few writes and very many reads, redundancy, and low response time. PDF here.