recommender systems - ppt
DESCRIPTION
RecommenderTRANSCRIPT
Data Sources of Amazon
Purchase dataItems Abandoned in cartsPrice elasticity experimentationReferral SitesObtain ratings from social networksDwell time: Amount of time spent on productMetrics Shipping address -> demographic information
Generic searchSort search resultsFilter search resultsOne click buy featureBig Data comes into picture
Collection of data over the years (address, payment info, items bought, items viewed, etc.)
Level 1 – Productivity Improvements
Level 2 – Improving Business Process
IS: Recommender Systems - Algorithm that uses customer’s historical behavior and product attributes to recommend other products that are most likely to be preferred by the user.
Anticipatory Shipping Patent - A patent to ship us goods before we have even made a decision to buy it, purely based on their predictive big data analytics. Though traditional retailers stock right items based on predictive analytics, here the new thing is Amazon is trying to take it to a personal level i.e., what you or I might buy. (based on past preferences - purchase data or other data sources)
Level 3 – Creation of a New business Model
The 3rd Party sellers get valuable information about the preferences of their customers through purchase data. These preferences could be used by the manufacturers to identify the evolving needs of the customers and used in new product development.
Create new products (hardware), services (software) which use these authentic user preference data at their core.
Eg - Kindle,Echo