Designing Search For Humans


Published on

Published in: Technology
  • Best one
    Hope you are in good health. My name is AMANDA . I am a single girl, Am looking for reliable and honest person. please have a little time for me. Please reach me back so that i can explain all about myself .
    Best regards AMANDA.
    Are you sure you want to  Yes  No
    Your message goes here
  • Great presentation - I especially liked the incorporation of flow.
    Are you sure you want to  Yes  No
    Your message goes here
  • Can you explain why you say this? I only removed a few images that were there just for decoration.
    Are you sure you want to  Yes  No
    Your message goes here
  • Too bad half of the presentation is missing.
    Are you sure you want to  Yes  No
    Your message goes here
No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Designing Search For Humans

  1. Designing Search for Humans Dr. Marti Hearst UC Berkeley Enterprise Search Summit Keynote Speech May 11 2010
  2. Consider the Human Feelings Language, Memory, and Planning Sociability
  3. Feelings Aesthetics Emotional Stages Flow
  4. Feelings: The Importance of Aesthetics <ul><li>With an aesthetically pleasing design: </li></ul><ul><ul><li>People will enjoy working with it more </li></ul></ul><ul><ul><li>People will persist searching longer </li></ul></ul><ul><ul><li>People will choose it even if it is less efficient </li></ul></ul>Nakarada-Kordic & Lobb, 2005, Ben-Basset et al. 2006, Parush et al. 1998, van der Heijden 2003
  6. Feelings: The Importance of Aesthetics <ul><li>Small details matter </li></ul><ul><ul><li>A left hand side line vs. a box for ads </li></ul></ul><ul><ul><ul><li>The line integrates the results into the page </li></ul></ul></ul><ul><ul><li>Balancing white space with content </li></ul></ul><ul><ul><li>Balancing font color, shape, and weight </li></ul></ul>Hotchkiss 2007
  7. Feelings <ul><li>Kuhlthau on informational AND emotional stages in search </li></ul>(Assuming novice researchers engaged in challenging tasks) Uncertainty and apprehension Optimism (after deciding) Confusion, uncertainty, doubt, frustration Confidence dawning * Confidence growing Relief and satisfaction (or disappointment) Initiation Selection Exploration Formulation Collection Presentation
  8. Feelings: The Importance of Flow
  9. Feelings: The Importance of Flow <ul><li>From Csikszentmihalyi, M. (1991). Flow: The Psychology of Optimal Experience. HarperCollins </li></ul><ul><li>via Bederson, Interfaces for staying in the flow, ACM Ubiquity 5(7), 2004 </li></ul>
  10. Properties of Interfaces with Flow <ul><li>Inviting </li></ul>Support interrupt-free engagement in the task No blockages Easy reversal of actions <ul><li>Next steps seem to suggest themselves </li></ul>
  11. Language, Memory, & Planning Address Anchoring and Vocabulary Problems Provide Memory Aids Suggest Helpful Next Steps
  12. Language
  13. Language: The Vocabulary Problem <ul><li>There are many ways to say the same thing. </li></ul><ul><li>People remember the gist but not the actual words used. </li></ul>
  14. Language: The Vocabulary Problem <ul><li>With no other context, people generate different words for the same concepts. </li></ul><ul><ul><li>The probability that two typists would suggest the same word for a given function: .11 </li></ul></ul><ul><ul><li>The probability that two college students would name an object using the same word: .12. </li></ul></ul>Furnas et al., 1987
  15. Language: The Problem of Anchoring <ul><li>Try this experiment: </li></ul><ul><ul><li>Tell people to think of the last 2 digits of their SSN </li></ul></ul><ul><ul><li>Then have them bid on something in auction </li></ul></ul><ul><ul><li>The SSN numbers they thought of influences their bids. </li></ul></ul>Ariely, Predictably Irrational, 2008, Harper
  16. The Problem of Anchoring <ul><li>Anchoring in search </li></ul><ul><li>A user starts with a set of words, then anchors on them </li></ul><ul><ul><ul><li>Harry Potter and the Half-Blood Prince sales </li></ul></ul></ul><ul><ul><ul><li>Harry Potter and the Half-Blood Prince amount sales </li></ul></ul></ul><ul><ul><ul><li>Harry Potter and the Half-Blood Prince quantity sales </li></ul></ul></ul><ul><ul><ul><li>Harry Potter and the Half-Blood Prince actual quantity sales </li></ul></ul></ul><ul><ul><ul><li>Harry Potter and the Half-Blood Prince sales actual quantity </li></ul></ul></ul><ul><ul><ul><li>Harry Potter and the Half-Blood Prince all sales actual quantity </li></ul></ul></ul><ul><ul><ul><li>all sales Harry Potter and the Half-Blood Prince </li></ul></ul></ul><ul><ul><ul><li>worldwide sales Harry Potter and the Half-Blood Prince </li></ul></ul></ul><ul><li>Contrast with the Vocabulary Problem! </li></ul>Russell, 2006
  17. Provide Memory Aids Support “Recognition Over Recall”
  18. Provide Memory Aids <ul><li>Suggest the Search Action in or near the Query Form </li></ul>,
  19. Memory Aids <ul><li>Provide Access to Recent Actions </li></ul>PubMed Dumais et al., Stuff I’ve Seen, SIGIR 2003
  20. Memory Aids; Anchoring Aids <ul><li>Dynamic Query Suggestions </li></ul>
  21. Memory Aids; Anchoring Aids <ul><li>Augment suggestions with images or faceted classes. </li></ul>
  22. Suggest Next Steps: Query suggestions <ul><li>Show suggestions after the query has been issued. </li></ul>
  23. Suggest Next Steps: Query suggestions PubMed
  24. Suggest Next Steps: Query Destinations <ul><li>Recorded search sessions for 100,000’s of users </li></ul><ul><li>For a given query, where did the user end up? </li></ul><ul><ul><li>Users generally browsed far from the search results page (~5 steps) </li></ul></ul><ul><ul><li>On average, users visited 2 unique domains during the course of a query trail, and just over 4 domains during a session trail </li></ul></ul><ul><li>Show the query trail endpoint information at query reformulation time </li></ul><ul><ul><li>Query trail suggestions were used more often (35.2% of the time) than query term suggestions. </li></ul></ul>White et al., SIGIR 2007
  25. Suggest Next Steps: Related Documents <ul><li>In some circumstances, related items work well </li></ul>PubMed
  26. Putting It All Together: Faceted Navigation <ul><li>Suggests next steps </li></ul><ul><li>Helps with Vocabulary Problem and Anchoring Problem </li></ul><ul><li>Promotes Flow </li></ul><ul><ul><li>Show users structure as a starting point, rather than requiring them to generate queries </li></ul></ul><ul><ul><li>Organize results into a recognizable structure </li></ul></ul><ul><ul><li>Eliminates empty results sets </li></ul></ul>
  27. A New Development: Faceted Breadcrumbs <ul><li>Nudelman, </li></ul>
  28. Sociability People are Social; Computers are Lonely. Don’t Personalize Search, Socialize it!
  29. Social Search Implicit: Suggestions generated as a side-effect of search activity. Asking: Communicating directly with others. Collaboration: Working with other people on a search task. Explicit: knowledge accumulates via the actions of many.
  30. The DARPA Network (Red Balloon) Challenge The ultimate in social question answering
  31. Social Search: Asking <ul><li>What do people ask of their social networks? </li></ul>Morris et al., CHI 2010 Type % Example Recommendation 29% Building a new playlist – any ideas for good running songs? Opinion 22% I am wondering if I should buy the Kitchen-Aid ice cream maker? Factual 17% Anyone know a way to put Excel charts into LaTeX? Rhetorical 14% Why are men so stupid? Invitation 9% Who wants to go to Navya Lounge this evening? Favor 4% Need a babysitter in a big way tonight… anyone?? Social connection 3% I am hiring in my team. Do you know anyone who would be interested? Offer 1% Could any of my friends use boys size 4 jeans?
  32. Social Search: Implicit Suggestions <ul><li>Human-generated suggestions still beat purely machine-generated ones </li></ul><ul><ul><li>Spelling suggestions </li></ul></ul><ul><ul><li>Query term suggestions </li></ul></ul><ul><ul><li>Recommendations of book, movies, etc </li></ul></ul><ul><ul><li>Ranking (clickthrough statistics) </li></ul></ul>
  33. Social Search: Explicit Help Question-Answering Sites <ul><li>Content produced in a manner amenable to searching for answers to questions. </li></ul><ul><li>Search tends to work well on these sites and on the internet leading to these sites </li></ul><ul><ul><li>This suggests that for the intranet, content is best generated and written this way. </li></ul></ul><ul><ul><li>Like an FAQ but with many authors and with the questions that the audience really wants the answers to. </li></ul></ul>
  35. Explicit Suggestions: Building Knowledge <ul><li>Social knowledge management tools seem promising </li></ul><ul><li>Utilize the best of social networks, tagging, blogging, web page creation, wikis, and search. </li></ul>Millen et al., CHI 2006
  36. Collaborative Search Pickens et al., SIGIR 2008
  37. Summary: Consider the Human <ul><li>Feelings </li></ul><ul><ul><li>Emotional responses to information seeking </li></ul></ul><ul><ul><li>Aesthetics </li></ul></ul><ul><ul><li>Flow </li></ul></ul><ul><li>Language / Memory / Planning </li></ul><ul><ul><li>Scaffold memory by suggesting next steps, providing context and feedback </li></ul></ul><ul><ul><li>Tools to aid with the anchoring and the vocabulary problems </li></ul></ul><ul><li>Sociability </li></ul><ul><ul><li>Search as a social experience </li></ul></ul><ul><ul><li>Turning to others for certain types of task </li></ul></ul><ul><ul><li>Sharing information for next-generation knowledge management </li></ul></ul>
  38. Thank you! Full text freely available at: