When there are companies that are able to search billions of web pages with sub-second response, it shouldn't be hard to search a product catalog an order of magnitude smaller than that. As part of my effort to learn about online retailing and the impact of search engine optimization, I realized that many people search using phrases such as "gifts below 100 dollars", "beautiful jewelry for under 50 dollars", "women's watches under fifty dollars" and so on. This may not be the most popular way of searching, but certainly a decent number of searches are done this way.
So, as part of taming the search's long tail I started identifying a set of key patterns applicable to GiftPickr and started tweaking the search query processing based on the phrases. So, when someone searches for any of the above mentioned phrases, they would get a decent set of results. Ofcourse, don't expect too much of natural language processing :). But even with my minimal effort, I think I have provided a much better user experience than the world's largest retailers!
Yes, as of 9/25/2008, I searched for "HTDVs below 1000 dollars" on Amazon, Walmart, Overstock and Newegg. The resulting webpages have been captured as images which you can view for Amazon, Walmart, Overstock, Newegg. When I initially set up on doing this on these major retail websites, my intention is to validate that I am doing it the right way. However, to my surprise, I found that none of these big guys have natural language search support at all!
Now, you can do the same search on GiftPickr for HDTVs below 1000 dollars and see what you get. May be the big players know more than me. May be they don't care about conversion optimization for the long tail. But I am happy to make GiftPickr support such simple phrases.
