shopping search (plus some odd rules)

Don’t know why they are doing this, but… they are suddenly spouting rules) rule #4 pick the right […]

Don’t know why they are doing this, but… they are suddenly spouting rules)
rule #4 pick the right porridge (good data
#4 mind the metadata allows for structured search, helps defensible position.

moderator gary stein, of jupiter– this convention center is between landfill and roller coasters… great analogy for the Internet.

Michael yang, become.com, founder of mysimon.
saw opportunities for structured and unstructured data to make a better search
graham jones, pricegrabber. recently added travel and autos. will be working with MSN to enrich their offer.
Chris Saito, Yahoo shopping.
use combination structured data via feeds and scraped to get breadth and comprehensiveness. just added auctions back in finally also added free listings. now have new and used. integrating yshopping across network.

Q: consumer benefit of breadth.
price is important, breadth is important. people don’t’ shop just on price. brand of merchant is incredibly important, credibility. community to rate merchant is critical… being able to rate, review merchants. If a merchant knows they have a good brand, they welcome this.
graham: a big myth is many merchants aren’t’ participating in comparison shopping because they can’t compete in price, but the reality is their brand carries them through, and they spend money on tv to acquire, but they could be doing comparison shopping and reaching out. They have trouble with SEO too– are buying “bed bath and beyond” but they aren’t buying “bed” or “bath”. Consumers want to comparison shop, they have to be there.
Michael– people today use general purpose search engines to buy, and they provide junk. It’s not trustworthy, it’s spammed, it’s not helpful. It’s more than price, it’s quality of product, it’s what others think of the product– so they are doing a algo based on multiple types of information.
Q: is there a shift yet form general search engines to verticals?
graham– we provide black/whitelabel to websites to reach out, such as cnet or pcworld. people are using the product, learning to value it, and looking for more. syndicating functionality. google adsense is not a feature, it’s money but we provide a real feature and real value.
Chris– integrate product search throughout the whole network. many people start with a search engine like yahoo, and we integrate with a tab, with actually results called out, integrate in the rest of the network, providing purchasing on say, yahoo movie reviews, so you can buy the dvd then. Also provide rich tools like saving lists of things you are researching, ad community is mixed into that.
Michael– search has no reason to be unless its good. We have our own ranking called AIR, affinity something ranking. as you type in the query, we show you different suggestions. So if you can’t articulate what you want the tool helps. it can suggest things you haven’t thought of.
Q: what about narrowing?
Michael: what’s nice is you have context… general purpose search does everything but not well, but vertical search engines can do certain things better so that people will migrate over. swiss army knife versus power drill. (hah! marissa loves to call google a swiss army knife)
Q: issues around feeds and crawling.
Graham. you can crawl as many sites are you want, but that’s just half of it. You need to categorize it.. bbq grill, car grill. it’s hard to categorize when you crawl, automated. but it’s easier when you have a product numbers, part numbers. but you don’t’ always get that and if someone clicks through on what hey think is a bbq grill but it’s a car grill, you aren’t sending a qualified link and your clients are bummed. feeds help solve this.
Michael: we think there is a lot of qualified info on blogs, buying guides, etc, but merchants don’t’ give you that. So we think there is a need for crawled data and merchant data feeds. value in both.

Q: talk about crawl& feeds. also, merchants complain about nonstandard feeds.
Chris: we embrace both. feeds give you high quality information, crawl helps fill the tail. better to provide results to the user, even if the data is not as clean. About comprehensiveness. regarding standardization. we are in favor, like to work with others to do so.
graham notices that feeds are starting to converge.

Q:have you seen improvement in comparison engine consumer quality?
Michael: we’ve seen coverage of 46%, but on become.com it’s 96%, so shopping customers are much more focuses, targetable. much higher click through as well.
raham: i never discourage people to use google or overture, because they are good campaigns. but you don’t’ get the same kind of targeting. a search engine search for grill, it might be a person looking for cooking tips.. but not on pricegrabber?

Q: how reliant are you on the big three for ad revenue
chris: integration is key. want users to stay on yahoo.
a search will tell you yahoo buys keywords
graham: we buy keywords, and it’s worth it.
Michael: we aren’t reliant on search engine marketing yet, but we plan to. still it’ll be a small part… we have our own algo engine.

Q: do you know how much of your traffic are doing online versus offline purchasing.
Chris: we know the majority of people buy offline. Local search helps address that problem.. still hard to measure. people print things out, take to store. clipping service and you can send to telephone so you call store helps bridge.
Q: what are people researching?
Chris: definatly product, but also consumer reports, trust in merchant and often they want a map. everything.
Michael: working to get inventory info in so if you want to buy local, you can see on their site if a local merchant has it. we don’t’ care where you buy it. it’s a research site.
graham: we’d love to get feeds form merchants on their inventory, it’s hard. often out of date, unavailable. story of online and offline site.. the online site had great conversation, the offline one horrid, and it was that people liked going to the physical store.

Q: how do you sort through the products and categorize them efficiently.
graham: feeds, plus a standard taxonomy they send out. incentive is better conversion. also, part of company does categorization. bad categorization is so bad, they won’t even list it.

Q: how do you balance revenue and unbiased relevency.
chris: we have special programs for advertisers, but relevancy stays untouched. merchants can do special guides, bold names, etc. but if you game results, users won’t use the site.

Q: would standardized feeds make it easier for startups to cannabalize?
graham: I’m not worried about it… they can do it now. we don’t’ treat the taxonomy as company jewels. we feel if we can provide a quality product we’ll do fine. it takes time to build trusted relationships with merchants.

Q: what about outgoing feeds, rss, api, persistent searches.
chris: we’re using standards to open up our data. we have rss feeds for top and new products.
graham: we have 300+ cobrands,a nd provide rss on top products. you can see most popular, updates 6x a day. allows partners to keep fresh.
Michael: we plan to support rss feed. we plan to become a shopping news engine so you can get recall notices, or see a new product coming out.