Fork me on GitHub

Recommendation systems

cdixon:

It seems that users have spoken and that recommendation systems are a feature, not a product. On the web, this means that recommendations either need to be 1) embedded in a consumption experience, the way they are on Amazon or Netflix, or 2) attached to a discovery mechanism, which (usually) means SEO’d via Google a la TripAdvisor and Yelp.

I learned this the hard way at Hunch. Ultimately, it was obvious that the best use of our technology was to embed it on eBay. I’d be excited if a standalone recommendation system succeeded, but until it happens I’m skeptical.

Dang. I respect Chris’s directness.

See more
This post has 24 notes
Tagged with ideas,
Posted at 2:07 AM 22 August 2012
  1. viablebiz reblogged this from cdixon
  2. This was featured in #Tech
  3. gbattle said: +1. You’ve answered what I’ve been pondering for a long time now. I think you’re right on this one Chris.
  4. matandme reblogged this from cdixon and added:
    Hmm. Yes.
  5. nikos reblogged this from cdixon
  6. amitkumar01 reblogged this from cdixon and added:
    Chris Dixon on recommendation engines
  7. j2labs reblogged this from cdixon and added:
    respect Chris’s directness.
  8. cdixon posted this
Bookmark and Share