Tag Archives: Google

Things I Learned this Week

Among the things I learned this week:

* Despite the Beastie Boys’ suggestion, Jamaica is not a significant producer of mangoes. (Courtesy: FAOSTAT, nicely displayed on Mongabay.com, and University of Florida’s Institute of Food and Agricultural Sciences)

* Physiocracy. (Courtesy: Car Talk reference; Wikipedia)

* “Package store” is Northern slang for liquor store. (Courtesy: HD)

* Bitrates of various streaming services:
– Spotify: Vorbis 160kbps with some at 320kbps if a Premium subscriber
– Amazon Cloud Player: MP3 at the rate the file was saved
– Pandora: AAC 64kbps and up to 192kbps, depending on subscription
– Google Music: play back at the quality of the file, except for FLAC, which is played back as mp3 320kpbs

* Lord of the Rings is “fundamentally religious and Catholic work”, according to Tolkien. (Courtesy: Wikipedia and follow up readings).

Search Index as the Web (Alternative Conceptualizations of the Internet)

The Register has a great piece on Cuil‘s launch, its impact on Google, and what the Web really is these days. While I don’t completely agree with the article’s point, thinking of the Web not as the culmination of linked documents but as The Index (i.e., search engine handling of the Web) is interesting and useful. Here are some of the key points from the article (“Spammers, Cuil, and the rescue from planet Google”):

With a little thought, Cuil not being as good as Google at finding what we want online is the least surprising piece of news since people familiar with the situation said JPII was partial to fish on a Friday. In 2008, Mountain View’s all-seeing algorithms in many ways are the web.

It’s easy to identify what happened. When it first surfaced in 1998, Google made sense of the web a bit better than anyone else. It was a useful improvement on existing services. Ten years later, the web does its best to make sense of Google.

The sorry upshot is that barring some unimaginable technological leap no search engine’s results will ever be better than Google’s, at least in the West. And the switch leaves the likes of Microsoft and Cuil (and a dozen other doomed start-ups) effectively attempting to reverse-engineer Google, not understand the information on the web.

…

The people at the vanguard of reverse-engineering Google are not its jealous search rivals. They’re the spammers and SEO consultants. They have driven an ever-closer relationship between the quirks and whims of Google’s algorithms and policies, and the structure and content of the web. It’s a feedback loop that was unavoidable once Google’s early rivals proved unable to respond to its better search results and presentation.

The Google (and GIM and GMail)

Given the attention Google receives, including from techies, investors, and media-types, it is surprising to see how little attention and analysis there has been regarding Google’s integration of Google Talk (or GIM) with GMail.

In general, Google uses sophisticated models and data-mining techniques to deliver more relevant advertising to users. Increasingly, however, the company is microscoping those techniques down to a user-specific level. The bait to tie advertising to a specific user (e.g., me), rather than a general type of user (e.g., people who perform a search for “glass + composer + ‘tour dates'”), was/is GMail. To use the service, people log in and have a specific account. In its initial form, e-mails were scanned and relevant advertising provided.

But that crude method, which was and to some degree remains the standard method, only provides a one-dimensional (i.e., e-mail) information/revenue stream. Steadily, Google has increased the number of services/revenue streams that improve/adjust when a user is logged into what was his or her GMail account and is now more generally referred to as a Google account. These includes News.Google, Base.Google, Images.Google, Groups.Google, and (Search.)Google.

While the Google (nee GMail) account was, and remains, the bait for attracting users to create trackable/information-creating accounts, GIM’s incorporation into GMail–which is done adequately, although not impressively, well–is the hook that keeps users logged in, trackable, and able to be advertised to in a personal (rather than customizable or typed) manner. That’s because GIM requires you to be logged in constantly and because there are people who only access their e-mail account when checking for new messages (as opposed to running it continuously and being notified when there is e-mail). The combination of the two is a significant barrier to Google’s efforts to collect user-specific data, but a barrier that is overcome because of GIM’s always-logged-in character. Important to remember is that once a user remains logged in the other personalizable Google services are activated, thereby generating not just the GIM and GMail information/revenue streams, but also a number of others (e.g., News.Google). That one GIM/GMail door opens up many worlds of new information gathering and advertising potential.

In fact, GIM’s integration with GMail may be one of the most important new products/developments the company has rolled out, and yet it met with little discussion and fanfare from the tech and investment community.