Strands develops personalization technologies to help users discover new things by capturing and analyzing their preferences from their online activities. The company aims to understand consumer preferences at different levels - in real-world activities, commerce transactions, and with individual products - to provide highly targeted recommendations and placements. As the web has evolved from static content to user-generated content, user profiles have become more complex and multi-dimensional. Strands sees a need for modeling user tastes, a portable format for exchanging profiles online, and potentially ontologies to represent preferences semantically.
1 of 10
More Related Content
UMAP 2009
1. Personalization on the Web
UMAP, Trento, Italy. June 25th, 2009
Marc Torrens
Chief Innovation Of鍖cer
Copyright 2003-2008, Strands, Inc
2. Understanding Peoples Preferences
mission Help People Discover New Things
vision Lead the Personalized Digital World
C A P
strategy Aggregate
Capture Personalized
preferences &
preferences services
extract patterns
Strands develops recommendation and discovery
technologies to help people discover new things
Copyright 2003-2008, Strands, Inc
3. The Strands Umbrella
Understanding Understanding consumer Understanding
consumer preferences habits at the commerce consumer behavior
in real-world activities (transaction) level at the product level
Consumer Space Business Space
Highly-targeted Product Placement
Copyright 2003-2008, Strands, Inc
4. Internet Evolution
from Individual Experts to Collective Intelligence
Web 1.0 Web 2.0
the mostly read-only Web the wildly read-write Web
INTERNET INTERNET
250k sites + 80M sites
Collective
Intelligence
Published
Content
Published
Content
Generated
Content
User
1996, 45M users 2007, +1.3B users
Inspired by a Social Computing Magazine article.
Copyright 2003-2008, Strands, Inc
5. Our Web 2.0 Online Life is Active
We express our tastes as we...
manage our money trade communicate with friends
are informed search for jobs
learn
get organized
buy clothes play games blog our passions
listen to music share videos
rate movies exchange goods bet
plan holidays
post pictures read news buy books
share presentations
...thus users implicitly create meaningfull and powerfull pro鍖les!
Copyright 2003-2008, Strands, Inc
6. Complexity in User Tastes
Our Tastes are Multi-Dimensional and Dynamic
friends
Me (today)!
music
newspapers clothes
places
investments
websites movies
books
games blogs
1. We need a way to model complex user tastes
Copyright 2003-2008, Strands, Inc
7. Our tastes should be ours to share (or not)
2. We need a portable format to exchange taste online
Copyright 2003-2008, Strands, Inc
8. APML within DataPortability initiative
Attention Pro鍖ling Mark-up Language
Attention Data Interest Cloud XML Attention Pro鍖le
3. Is it enough? do we need ontologies?
Copyright 2003-2008, Strands, Inc
9. some observations...
犢 Pro鍖les are (the) key for Personalization (more than algorithms?)
犢 Distribution: Pro鍖les are distrituted over the web
犢 Ownership and Privacy: Pro鍖les should be owned by users
... thus, we need more on:
1. MODELING: complex pro鍖les (multi-dimensional and dynamic)
2. PORTABILITY: portable format to exchange tastes online
3. SEMANTICS: do we need it? ontologies?
Copyright 2003-2008, Strands, Inc
10. Contact
Dr. Marc Torrens
Chief Innovation Of鍖cer
torrens@strands.com
Copyright 2003-2008, Strands, Inc