Last night I participated in a panel discussion for the Web Apps & E-Commerce SIG of the SVForum in Palo Alto on Recommendation Engines and Brand Commerce . I was joined on the panel by Brian Johnson from eBay, Paul Twohey from Ness, and Vik Singh from Slice Data. Each of us had a chance to present our views on the subject and then answer some questions. I presented an excerpt of content from our recent webinar on the Psychology, Technology, and Science of Personalization.
Search is everything / everything is search. Paul made some interesting comments this area that I agree with at some level. When people tell you what they want in search, that is an awesome signal of intent that you can use to personalize experiences. Having gotten our start in social search and also having a term space and search-based algorithms as a critical part of our platform, we obviously understand his point.
There is confusion about the uses of Hadoop and Mahout for real-time applications. We spent some time on this. My comment was that while we use some of the Mahout libraries, we have built our own models and real-time infrastructure to meet the needs of our ecommerce customers.
There is a lot of interest in machine learning for applications outside of recommendations. We talked about weather, tracking the movements of vehicles and people and a number of other subjects. The fact that machine learning technology is being used in all of these areas just points to how integral it has become in our everyday lives, without many people even knowing it.
All in all, I am glad I could participate in the dialog and I encourage anyone who is in the Silicon Valley area to take a look at the SVForum and the good work that they are doing.