Monitoring Google via Patent Documents, Definitely Fun

May 8, 2010

As soon as I returned from San Francisco, it was telephone day. Call after call. One of the callers was a testosterone charged developer in a far off land. The caller had read my three Google studies and wanted to know why my comments and analyses were at variance with what Googlers said. The caller had examples from Google executives in mobile, enterprise apps, advertising, and general management. His point was that Google says many things and none of the company’s comments reference any of the technologies I describe.

I get calls like this every couple of months. Let me provide a summary of the points I try to make when I am told that I describe one beastie and the beastie is really a unicorn, a goose, or an eagle.

First, Google is compartmentalized, based on short info streams shot between experts with sometimes quite narrow technical interests. I describe Google as a math club, which has its good points. Breadth of view and broad thinking about other subjects may not be a prerequisite to join. As a result, a Googler working in an area like rich media may not know much or even care about the challenges of scaling a data center, tracking down SEO banditry, or learning about the latest methods in ad injection for YouTube advertisers. This means that a comment by a Google expert is often accurate and shaped for that Googler’s area. Big thinking about corporate tactics may or may not be included.

Second, Google management—the top 25 or 30 executives—are pretty bright and cagey folks. Their comments are often crafted to position the company, reassure those in the audience, or instruct the listener. I have found that these individuals provide rifle shot information. On rare occassions, Google will inform people about what they should do; for example, “embrace technology” or “stand up for what’s right”. On the surface these comments are quotable but they don’t do much to pin down the specific “potential energy” that Google has to move with agility into a new market. I read these comments, but I don’t depend on them for my information. In fact, verbal interactions with Googlers are often like a fraternity rush meeting, not a discussion of issues, probably for the reasons I mentioned in point one above.

Third, Google’s voluminous publicly available information is tough to put into a framework. I hear from my one, maybe two clients, that Google is fragmented, disorganized, chaotic, and tough to engage in discussion. No kidding. The public comments and the huge volume of information scattered across thousands of Google Web pages requires a special purpose indexing operation to make manageable. I provide a free service, in concert with Exalead, so you can search Google’s blog posts. You can see a sample of this service at I have a system to track certain types of Google content and from that avalanche of stuff, I narrow my focus to content that is less subject to PR spin; namely, patent documents and papers published in journals. I check out some Google conference presentations, but these are usually delivered through one of Google’s many graduate interns or junior wizards. When a big manager talks, the presentation is subject to PR spin. Check out comments about Google Books or the decision to play hardball with China for examples.

My work, therefore, is designed to illuminate one aspect of Google that most Googlers and a most Google pundits don’t pay much attention to. Nothing is quite so thrilling as reading Google patent applications, checking the references in these applications, figuring out what the disclosed system and method does, and relating the technical puzzle piece to the overall mosaic of “total Google”.

You don’t have to know much about my monographs to understand that I am describing public documents that focus on systems and methods that may or may not be part of the Google anyone can use today. In fact, patent documents may never become a product. What a patent application provides includes:

  1. Names of Google inventors. Example: Anna Patterson, now running I don’t beat up on because Dr. Patterson is one sharp person and I think her work is important because she is following the research path explained in her Google patent documents, some of which have now become patents. In my experience, knowing who is “inventing” some interesting methods for Google is the equivalent of turning on a light in a dark room.
  2. The disclosed methods. Example: There’s a lot of chatter about how lousy Wave was and is. The reality I inhabit is that Wave makes use of a number of interesting Google methods. Reading the patent applications and checking out Wave makes it possible to calibrate where in a roll out a particular method is. For that reason, I am fascinated by Google “janitors” and other disclosures in these publicly available and allegedly legal documents.
  3. The disclosures through time. I pay attention to dates on which certain patent documents and technical papers appear. I plot these and then organize the inventions by type and function. Over the last eight years I have built a framework of Google capabilities that makes it possible to offer observations based on this particular body of open source information.

When you look at these three points and my monographs, I think it is pretty easy to see why my writings seem to describe a Google that is different from the popular line. To sum up, I focus on a specific domain and present information about Google’s technology that is described in the source documents. I offer my views of the systems and methods. I describe implications of these systems and methods.

I enjoy the emails and the phone calls, but I am greatly entertained by my source documents. My fourth Google monograph, Google Beyond Text, will be available in a month or so. Like my previous three studies, there are some interesting discoveries and hints that Google has reached a pivot point.

Stephen E Arnold, May 8, 2010

Sponsored post. I paid myself to write this article. Such a deal.


Comments are closed.

  • Archives

  • Recent Posts

  • Meta