TV Scout Lowering the entry barrier to personalized TV program recommendation Patrick Baudisch & Lars Brueckner AH 2002 June 1 th 2002
Contents Motivation TV Scout user interface –Retrieval part… –…leading to the filtering part Results of usage data analysis Conclusions
Motivation: Information overload Too many research papers, books, movies, web pages… even TV programs Germany: printed program guides list programs per two weeks Content of interest has not increased proportionally planning TV has become a challenge* Goal: Reduce the set of programs that users have to look at to find relevant programs Allow users to watch TV more selectively
The initial concept… We wanted to offer: personalized TV program listings “at a single mouse click” Resulting user interaction: “Sure, we’ll tell you what’s on tonight, but before we do that, please answer these 30 questions…” Guess how users liked that…
We did some field work… Users’ expectations are inspired by printed TV program guides –Step 1: Find the right listing –Step 2: Sift through the listing –Step 3: Remember or mark-up programs to watch –Step 4: Watch User interface design challenge: –Pick people up where they are (printed TV program guides) –… –…and guide them to personalized listings at a mouse click
Best match Step 1: Select a query Exact match
program description list program description table retention menus Step 2+3: Read & retain program descript.
program description list program description table retention menus Step 4: Print it out & watch TV video labels laundry list
Emulating a printed guide Printed program guide Step 1: Pick the right listing Step 2: Sift through listing Step 3: Mark-up programs Step 4: Watch TV Scout Step 1: Pick the right query Step 2: Sift through listing Step 3: Retain programs Step 3b: Print it out Step 4: Watch
But then: suggestions and bookmarks
Personalized schedules at a mouse click
Not that users have to, but…
Summary of usage system compiles one-click TV program S3 TT user updates system learns T3 U3 bookmarked queries user defines system suggests S2 U2 T2 queries S1 U1 T1 start system provides user writes
TV Scout usage data TV Scout user interface concept = delayed disclosure of the filtering functionality Does this actually reduce the entry barrier to personalized filtering? => Informal analysis of log file data from actual web usage
Procedure 18 months of log file data, extracted from the web server log files and the system’s database Gathered data –10,676 registered users –In total, users had executed 48,956 queries –53% of all queries (25,736 queries) were specific queries different from the default query. Bias: the suggestion feature became available later
Goals Goal 1: Repeated usage would indicate that users had taken the entry hurdle Goal 2: Learn more about the users’ demand for the offered filtering functionality: How many would use bookmarking and/or query profiles? Goal 3: How useful users would find the query profile. Query profile users, would they use or abandon it?
Results & conclusions Repeated log-ins: 9,190 of 10,676 users logged in repeatedly (= 86%) Very high percentage for a web-based system => Delayed disclosure of filtering functionality is a successful approach to keeping the entry barrier for first-time users low
Results & conclusions Bookmarks & Query profiles –1770 users had bookmarked 4383 queries (= 17%) –270 users executed query profile (= 15% of bookmark users) –They executed their query profiles 5851 times (21 times per user). –Once they used the profile they liked it Only 17% used filtering functionality, isn’t that low? –Survey: only 12% of the users of printed TV guides planned TV schedule for a week or longer –=> The 83% non-bookmark users may have found retrieval to be the appropriate support for their information seeking strategy Future work: An online survey as well as an experimental study should help to verify this interpretation.
Thanks to: Dieter Böcker, Joe Konstan, Marcus Frühwein, Michael Brückner, Gerrit Voss, Andreas Brügelmann, Claudia Perlich, Tom Stölting, Diane Kelly, and TV TODAY Further reading & demo: END
If time left –Explain system architecture –Demo paintable interfaces
TV Scout Architecture program descriptions Content provider Movie database Program description database Query subsystems Exact match filtering Date Time Profile ChannelProfile feedback QSA filtering QSA profile Retention tools Video labels Laundry list Time Dialog ChannelDialog Editors’ tips User tips Text search Genres Estim. Pop. ACF query hoc ad
P+. Exact match profiles channel profile editor viewing time profile editor
Slides to bring up during questions
TV Scout UI TV Scout interface with starting page viewing time profile editor channel profile editor query menus QSA menu text search program description list program description table suggest queries QSA profile editor QSA profile editor (experts) retention menus video labels laundry list
Structure of TV Scout user profiles user profile QSA profile q1q1 A qnqn … e.g. news, sports, Comedy shows e.g. news, sports, Comedy shows How does user like news compared to sports…? How does user like news compared to sports…?
Cooperation with German TV TODAY 17,000 registered users TV Scout: retrieval usage summary retention tools TV listing & table
Further reading P. Baudisch. Dynamic Information Filtering. Ph.D. Thesis. GMD Research Series 2001, No. 16. GMD Forschungszentrum Informationstechnik GmbH, Sankt Augustin. ISSN , ISBN P. Baudisch. Recommending TV Programs on the Web: how far can we get at zero user effort? In Recommender Systems, Papers from the 1998 Workshop, Technical Report WS-98-08, pages 16-18, Madison, WI. Menlo Park, CA: AAAI Press, P. Baudisch. The Profile Editor: designing a direct manipulative tool for assembling profiles. In Proceedings of Fifth DELOS Workshop on Filtering and Collaborative Filtering, pages 11-17, Budapest, November ERCIM Report ERCIM-98-W001. P. Baudisch. Using a painting metaphor to rate large numbers of objects. In Ergonomics and User Interfaces, Proceeding of the HCI '99 Conference, pages , Munich, Germany, August Mahwah: NJ: Erlbaum, 1999.