Transcript Web2.0-Master series

The future of Media
Personalization
Agenda
• Customization and personalization aspects:
▫ Place (home, mobile)
▫ Time
▫ Content
• Personalization/Recommendation engines
▫ Intro to media recommendation methods
▫ Media personalization in IPTV, Mobile and Web
• Personalization in advertisement - Targeting
▫ Personalized video Ads on mobile & Web
▫ Seambi example
Web2.0-Master series
Media Customization
Place adaptation
• Place
• Time
• Content type
Mobile
PC
• This lecture concentrate on
how to adapt/personalize
the content.
UGC
Content
Type adaptation
Web2.0-Master series
Broadcast
TV
VOD
PVR
Personalized
content
Owned
content
Time adaptation
Content personalization
• There is a surge in the amount of available
content:
▫ 1 channel -> Thousands of channels
▫ 10 VoD Titles->Thousands of VoD titles
▫ Tens of millions of UGC clips
• How to select the right content?
▫ Let the system select the content for you
Media Recommendation Engines
Web2.0-Master series
Why Recommendation Engine?
Know
• Get to Know Each customer as an Individual
Recommend
• Make Personal Recommendations
User Experience
• Enhance the User Experience
Web2.0-Master series
Increase
sales
Recommendation Engine Methods
Source:
TrustedOpinion
Recommendation
Systems
Collaborative
(Peer based)
Technology is mostly:
natural language processing
Correlation matrix
Content Based
Hybrid
(use both
methods)
Web2.0-Master series
Music Recommendation Services
Web2.0-Master series
MeeMix
• Analyze the user taste in order to provide the
best personalized music channel
Web2.0-Master series
Video Recommendation samples
• Commercial version of
MovieLens with better
features & GUI
Collaborative /Peer based
Content based
Web2.0-Master series
Desktop/web application sample
• Recommendation/Rating:
▫ Based on the known 1-5 star system
▫ Use of peer/group recommendation icon
Web2.0-Master series
Mobile sample
• Rating is kept simple:
▫ Like it/hate it/no opinion
• Recommend ions are:
▫ You’ll love it
▫ You might love it
Web2.0-Master series
Movie selection ->personal channels
• Once we are sure of users needs we can assist
him in creating his own “Personal TV Channel”
• Personal channel is most common in:
▫ Music Recommendation Engine
▫ Mobile recommendation engine
• Starting to catch up on Web TV channels
• Channel Creation Platforms
Web2.0-Master series
Trends
Process
Past
Present
Future
Distribution
Broadcast
Pre defined
channels
UGC / Pre defined
channels
My Customized
channel
Consumption
Lean Back – open
loop
Lean Forward
Text search of
clips
Lean back
Content type
Premium Content
UGC content +
Stolen Premium
content
UGC + Legal
Premium
content+Meshups
Web2.0-Master series
Trends
Studio
Creates
Content
Distributer
Broadcast
• Recommendation
Systems closes the
loop between content
creator/distributers
and the users
Distributer
Web2.0-Master series
User
Content
Creator
Recomm.
System
User
Recommendation engines in
Video Advertisement
Advertisement Personalization-Targeting
• In broadcast TV ->broadcast Ads
• In IPTV and Web we can use flash->video
advertisement without transcoding.
• Use targeting and peronalization engine
• Advertisers use viewers information for ad
targeting including:
▫
▫
▫
▫
Location
Demographic
User profile
User content
Web2.0-Master series
Personalized Video Ads
• Adobe Flash and Microsoft Silverlight Enables:
▫ Using one video version
▫ Changing the video advertisement per user without
transcoding the video
Video &
Audio
Images,
Scripts,
special
Effects
Web2.0-Master series
Clip +
Personalized
Video
Advertisement
Video Advertisement - Standard
• Pre and Post Roll
• Overly
Doomsday Movie advertisement
as Overlay on a pre-roll of
Doomsday video
Web2.0-Master series
Samples of Personalized AD
Web2.0-Master series
Thank You !