March 28, 2005

Email Q&A With Michael Yang of Become.com

Become.com, the new shopping search engine that released its public beta in February, has been a recurring subject at the Organized Shopping Blog. I'll admit to being impressed with their initial search results, but I'm most interested in what they plan to do next... if they will find new synergies between the structured shopping databases and the unstructured Web.

I recently emailed some questions to Michael Yang, the Founder, President & CEO of Become.com. We had exchanged emails before, but had not found time to talk. Not knowing when our schedules would cooperate, I went the email route to get the ball rolling, and Michael was kind enough to answer the following questions about their direction. We are planning to talk later this week, and I will try to add detail on these and other topics. If you would like to pass along any ideas before then, feel free to contact me.

+ + +

Do you see much room for innovation on the "comparison" side, or will the bulk of the innovation come from the "crawler" side... or the  integration of the two sides?

Right now many of the comparison shopping engines have very similar offerings, so we definitely see the opportunity to innovate on the comparison side. Likewise, there hasn’t been a real quantum shift in web search technology since Google introduced PageRank, and that was in the late ‘90s. Become is first and foremost a technology company, and we are looking to revolutionize both these areas, as well as being in a unique position to combine them.

Besides the reliance on outgoing links (in addition to incoming links), can you provide any more detail on what is different about the Affinity Index Ranking algorithm?

AIR ranks web pages based on their level of interconnection with web pages on the same topic. Both inlinks and outlinks are evaluated to understand the level of interconnection among the sites. Links from valuable web pages on the same topic work to increase the score of a web page; links from off-topic sites do not help. Linking to spam pages or off-topic pages can reduce the score of the page generating the link. Advanced mathematics and concepts from Applied Physics and Engineering Dynamics are used to calculate specific scores.

Do you think standalone link-based algorithms will eventually be supplemented by direct visitor feedback? Is your five-star rating system a hint in this direction?

Link-based connectivity is an important aspect of evaluating web pages, and we believe that our methods of evaluating connectivity are significantly more accurate than the methods used by major search engines today.

Direct visitor feedback is another indicator of the value of a web page. While direct visitor feedback may be used to supplement connectivity information, we do not believe that it will replace link-based algorithms in the near term.

How many pages do you estimate Become.com would index if you did not filter non-shopping pages?

We have specifically focused our crawling efforts on US-based content which is oriented toward shoppers. We do not have an accurate estimate of the number of pages indexed if we conducted a worldwide crawl of all types of content.

Do you plan to extend the upcoming shopping-comparison service into other countries or languages?

Yes. We plan to expand internationally after establishing our service in the US market.

+ + +

see also:
Become.com | Overview of Services | Press Center | Jobs | Contact

Email Q&A With Michael Yang of Become.com
Posted by Sean O'Rourke on March 28, 2005 at 6:46 AM
Archived at Become.com | Interviews



Subscribe
RSS
Twitter

Favorites
Product Review Mini-Series: Features
Product Review Mini-Series: Numbers
Attribute-Based Shopping Search Shootout
More Context for Feature Finders
What if Popular Products had 1000+ Reviews?
Product Portals - How Big Are They?
Why I Don't Use Shopping Comparison Sites
How Many Shopping Engines Do We Need?
Thoughts on Aggregation (No Respect)
Feedback Engines (Why The Symphony Sucks)
Why the Lack of Attributes at Amazon.com?
Study: Searchers Operating w/ Blunt Instruments

Profiles
LinkedIn
Hunch

Copyright © 2005-2009
Organized Shopping, LLC

Contact Sean