Moving Google Beyond Search
May 19th, 2006
Bill Tracer, over at Hitwise, shares with us this lovely bit of information:
“Leveraging the custom category capability of Hitwise, I’ve created a category of the top 20 Google domains in order to understand the popularity of Google’s varied services. The table below details the percentage market share that each property accounts for in relation to all visits to the top 20 Google Domains. “
Not unlike what Josh noticed when studying UIE.com’s logs, Google displays a Long Tail distribution, something which occurs again and again when we study the frequency of web content. The first property, Google Search, far outruns the second property, Google Image Search, by a factor of eight. Coming in at half of Image Search is Google Mail. And the fourth property, Google News, is half of Mail, with Google Maps following right behind it. (When you combine the two “search” properties — Google Search and Google Image Search — they add up to more than 89% of all use.)
Unlike what Josh found, the sum of the tail for these properties does not exceed the head. Google’s Search is far more popular than all the other Google properties combined.
There are a couple implications we can read from this:
- There is no single formula that dictates how users demand content across a site or brand. You need to collect your own data and determine your own users’ behaviors.
- If you frequently use the less popular Google properties, such as Google Finance, Google Local Search, and even Google News, you are not normal. Yet more evidence that you shouldn’t extrapolate your behaviors and preferences into those of your users.
- Google, with all of its market capitalization and clout, doesn’t automatically win instant popularity when it introduces a new service, such as Google Calendar. The designers at Google have to work as hard as everyone else to provide a great experience to make a service something people want to use.
May 19th, 2006 at 10:49 am
I wonder what *data patterns* are emerging under the *google hood* to allow the smart folks at Google to create innovative mashups that people really need.
For example, think of a service like *Google News* that is completely machine generated and serves up answers to key questions people have in a specific domain. How does Google find those key questions & domains?
Apart from the wealth of data and user behaviours they have on hand …
My guess is field research and understanding user needs BEFORE technology and instead of serving up services that are already a dime a dozen on other sites.
May 19th, 2006 at 11:21 am
[...] Jared Spool — “Google displays a Long Tail distribution” (Related: Google 2.0) [...]
May 19th, 2006 at 11:38 am
This data could be affected by users’ home page settings. I set Google as my home page; that means I visit it every time I launch my browser regardless of whether or not I’m doing a search. Perhaps half of my visits to Google do not result in me using their service.
Google is a likely target to be a popular home page. I wonder how that would skew the results.