Ask.com: Housecleaning, New Paint, Same Damp Basement
October 6, 2008
My newsreader is stuffed like a six-year old’s teddy bear with views, opinions, and analyses of the “new” Ask.com. I remember when humans worked like beavers to hand craft “answers” for the AskJeeves.com “natural language search engine.” That is in the mid 1990s. About a year ago, a semi-wizard from a third-tier consulting firm regaled me at dinner with insights into the innovations that AskJeeves.com (updated as Ask.com) was unleashing to hoards of Web searchers. Then I learned that some of the Rutgers’ wizards had returned to academe. I suffered through a technical paper that explained how Ask.com could scale. As skeptical as I am of Amazon’s approach, the Ask.com presentation was even less convincing.
Today’s stories are capped with the snow on the journalistic Mount Everest. Miguel Helft’s “Ask.com Revamps Search Engine” is the summit from which this Ask.com bobsled launches. You can find his write up here. Please, read this story. See if you agree that one of the most interesting comments in the piece is this statement:
The new Ask.com also includes an index of various question-and-answer sites from around the Web, including Yahoo Answers and WikiAnswers, that proves effective at returning results for some queries posed as questions.
I interpret this as a metasearch play. Now metasearch technology has made great strides since the early days of Dogpile.com. Among the services I use are Ixquick.com, the remarkably helpful Devilfinder.com, and the relative newcomer, Clusty.com. I also keep a copy of Copernic on one of my machines because it’s built in collection narrowing function is helpful for some of my queries.
I tested Ask.com with this query, “What’s the capital of Tasmania?” The system’s first result was “Hobart.” The second result was correct as well and pointed me to Wikipedia, the go-to site for general information. Powerset leveraged a demo featuring Wikipedia content into $100 Microsoft dollars not long ago. My second query was a different kettle of fish or “answers”. The query was, “What is the architecture of the Google File System?” The first result pointed to a useful article from the HighScalability.com site here. My other test queries returned useful results, and I concluded that the “new” version of Ask.com was pretty good.
The Search Engine Land write up focused on the notion of “structured search”. With the shift to XML and the crossover from flat ASCII to XML having taken place sometime last year (according to a Google document I located using the Google.com search engine), this is a good point to make. In fact, the Search Engine Land story here made this point:
…we expected this to come. After seeing it, I personally still do not consider Ask.com to be a core search engine and thus do not consider them to be in the race with Google, Yahoo or Microsoft. In fact, I find it interesting that Ask.com is bring back the Jeeves approach, which failed back then – but they hope will work now.
I did not see this coming because I don’t pay much attention to the machinations of Barry Diller’s Web and online empire. I don’t think of Ask.com as a resource suitable for my research needs. I probably won’t change my information retrieval habits to much either. Now look at the usage data for the top search engines. I have used the comScore data from this source.
Here’s a Google Trends’ report on the incidence of queries for the top four search engines: Ask.com, Google.com, Live.com, and Yahoo.com. You can update this query here.
Yahoo.com is leading the pack in queries on Google. My hunch is that Yahoo’s financial and business challenges make it popular. What do these data tell us? A Web site can get traffic by becoming a major business story. A Web site with a five percent share of Web query traffic with an improved search engine has a lot of work to do to get traffic.
Ask.com, therefore, has to have a better search engine, and it has to set the media on fire with its search engine. I will check out my newsreader 24 hours from now to see if the Ask.com “news” has staying power. My hunch is that this upgrade won’t have enough horsepower to pull up the Ask.com market share. More is needed. I want to be candid. I am not sure what Ask.com can do to build buzz. Crazy advertisements, staff churn, and a different interface catch my attention, but these are not sufficient to change my research behavior. Google will probably get a bump because people will navigate to Google and type the query “ask” in order to get a direct link to the Ask.com service. Agree? Disagree? I really want to hear from the Traces and Whitneys working at consulting firms to set me straight on my perception that Ask.com is stuck in the damp basement of consumer online Web traffic.
Stephen Arnold, October 6, 2008