Thank you to…
-Yahoo! (especially Jeanne Condit & Havi Hoffman) for hosting the event and providing the munchies
-Beach for organizing the event
-David Weinrot (Shopzilla), Chris Saito (Yahoo! Shopping), Siva Kumar (TheFind), and Neil Mayle (BrowseGoods) for participating
-Everyone for attending!
-SD Forum & Jeff Clavier (the Chair of the Search SIG)
-and the Academy…
Here’s some coverage of the event…
-Freshelectrons (Havi helped launch Yahoo! Shopping back in ’98!)
Someone’s anonymous notes/thoughts…
SDForum ‘Search’ SIG meeting, at Yahoo Sunnyvale
Tuesday Feb 13, 2007
Brian Smith (www.comparisonengines.com)
Chris Saito (Yahoo! Shopping)
David Weinrot (Shopzilla)
Siva Kumar (TheFind)
Neil Mayle (Dotted Pair)
• Yahoo Sunnyvale is huge. Meeting held in Classroom C. Big enough to hold several hundred people. Maybe 150-200 there. The Shopzilla VP on the panel joked that his entire company could fit into it.
• My impression was that >50% of the people worked at Yahoo. They seemed older than the Googlers in general, and less cocky (not too hard to understand that, I guess).
• Session was led by Brian Smith of Comparison Engines. He was amusing and self-deprecatory – mentioned his background as a former stand-up comedian, said it was easy to be ‘the leading analyst in this space’ when you’re the only one.
• Some pretty funny technical problems, given where we were. Terrible feedback from the audio in the room, persisted for some time. Then the internet connection used for the demos was slow. When this happened during the Yahoo Shopping demo one wag suggested “Maybe it would run faster if we moved closer to Yahoo”.
• The Yahoo Shopping guy was trying to demo a results page, but no sooner had be started when he was interrupted by three huge pairs of disembodied legs dancing all over the page (a Circuit City ad). This unfortunately says a lot about Yahoo (and why I never use it).
• Shopzilla and Yahoo Shopping are conventional shopping comparison search engines. The Find is a start-up focusing on soft goods, has bigger pictures than the thumbnails shopping engines usually show. Dotted Pair has a uniquely visual approach – you look down on pictures of the products as if they were laid out on the floor of a gigantic aircraft hangar, and then zoom in on categories and eventually products in a Google Maps-like fashion.
• There was a discussion on crawling the web for price and product data versus feeds. Some of this, either for or by people unfamiliar with shopping search, was about the desirability of the one approach versus the other, as if they were exclusive or one was superior (not in my view a helpful or insightful perspective). Some more interesting observations though (phrases in quotes may be paraphrased):
- “You get more data from a crawl”. There’s often more information on the product web page than the merchant gets around to putting in his feed. You can get some or all of it if you take the time to direct your crawler accordingly.
- “Some sites are java-heavy or completely in Flash”. Makes you more reliant on
crawling data feeds for those products.
- “You have to be careful what you mean by ‘crawling’ these days, it’s not always the same thing”. I think what TheFind’s CEO meant was that crawlers are increasingly sophisticated robots, don’t necessarily break with every change on their target sites.
- “Crawling throws up a frequency issue”. Some supplier sites have more unique product offerings than you can crawl every day, so you have to adopt a selective targeting and scheduling strategy.
- “Some offers [you acquire by crawling] may be ones the retailer doesn’t want to appear on your price comparison site”
- “More feeds are becoming available all the time, but you still have to make decisions about people who don’t have feeds”.
- Shopzilla’s VP made a point about how concentrated the demand can be (and presumably therefore how important it is to focus on getting a tiny subset of the offers right) – currently it seems that a single Canon digital camera accounts for over 4% of Shopzilla’s entire volume in the digital camera category.
- “Shopzilla is now all feeds [no crawling at all]“.
- I think myself you have to recognize that people start off using crawled data because that’s what they’ve got. Feeds come later on, when you’ve attracted enough users that the suppliers will pay attention to you, and become more important subsequently when you can begin to influence what data comes to you in the feed.
Read the rest of this entry »