networked markets ashish goel mukund sundararajan
TRANSCRIPT
Networked Markets
Ashish GoelMukund Sundararajan
Networked Markets
• Markets that could not have been possible without the Internet “at scale”
• Will briefly describe the concepts and business ideas underlying these markets
• Will largely focus on the mathematical and technical aspects that lie at the heart of how these markets use and are used by human beings
• Can not cover all these technical topics in great detail, but will give you a good understanding of the underlying technical ideas
Examples
Business: Groupon
Technical ideas: Group Action; Network Formation; Virality
Examples
Business: Netflix and Amazon
Technical ideas: The long tail of human interests; personalization; reviews and ratings
Examples
Business: eBay
Technical ideas: The long tail of production; new pricing models; reputation systems
Examples
Business: TripAdvisor, Yelp
Technical ideas: Reviews and ratings
Examples
Business: Mechanical Turk and oDesk
Technical ideas: crowdsourcing
Examples
Business: Search
Technical ideas: Crowd-sourced reputation systems; advertising networks
Examples
Business: Social networks, social media
Technical ideas: advertising networks; personalization
Examples
Business: Online retail and advertising
Technical ideas: targeting; personalization; different pricing models for advertisements
Examples
Business: Ashley Madison
Technical ideas: privacy
Examples
Business: Silk Road
Technical ideas: virtual currencies; anonymity
Examples
Business: Peer-to-Peer Content Sharing
Technical ideas: reputation systems; virtual currencies; anonymity
Examples
Business: Online education
Technical ideas: peer grading (similar to reviews and ratings)
Technical Topics Covered– Groupon: Group action; network formation; virality– Netflix and Amazon: The long tail of human interests; personalization; reviews
and ratings– eBay: The long tail of production; new pricing models; reputation systems– TripAdvisor, Yelp: Reviews and ratings– Mechanical Turk and oDesk: distributed work– Search: Crowd-sourced reputation systems; advertising networks– Social networks, social media: advertising networks; personalization– Online retail and advertising: targeting; personalization; different pricing models
for advertisements– Ashley Madison: privacy– Silk Road: virtual currencies; anonymity– Peer-to-Peer Content Sharing: reputation systems; virtual currencies; Anonymity– Online education: peer grading (similar to reviews and ratings)
Today: Long Tails
• Long tails are the signature of human activity.– Christos Papadimitriou
• Popularized by an article in “Wired” by Chris Anderson– Details in a book called “The Long Tail” by Chris Anderson
• Another interpretation: Long tails are a balance between human individuality and social behavior
• Technical details in notes
Implications of Long Tails
• Challenges for search: search keywords are diverse; can not just process the top few keywords offline; need to assemble search results in real-time!
• Opportunities for advertising: search keywords express intent; advertisers can use the long tail of search to sell a long tail of products
Implications of Long Tails
• Example: Blinds.com . Started 1996, and soon grew to be a large business.
• Blinds are a product in the long tail: they are bought infrequently, but the total demand is still quite large when aggregated over a country (Blinds.com has > 50M in revenues)
• The Internet allowed them to have a virtual storefront
• Search advertising allowed them to reach the long tail of customers very efficiently
Implications of Long Tails
• Example: Netflix’s DVD business• DVD licensing fees per rental are negotiated
between Netflix and the studios; depend on popularity of the movie
• If Netflix rents out a less popular DVD, its costs are lower
• Enter the recommendation system: by offering you personalized recommendations, Netflix makes good use of the long tail in people’s tastes