4. Results
Term Users’ intent Facet/ structure
free I’m cheap. Pricing
brunch Full English vs. eggs benedict Mealtimes
jazz My music tastes are Music genres
sophisticated.
Les Miserables Is Russell Crowe as bad as Name
the reviews?
I want to see the theatre
musical.
Italian Pasta anyone? Cuisine
Shoreditch What’s local to me? Area
kids It’s half term, what can we Audience
do?
today I’m bored – what can I do? Date
clubbing Let’s dance! Subject
Cliff Richard Calendar anyone? Name
8. Modes of information seeking
Marcia Bates – Modes of Information Seeking
Http://gseis.ucla.edu/faculty/bates/articles/info_SeekSearch-i-030329.html
9. Intent in the search interface
OR
What Where When
Clapham All dates Search
Clapham
Clapham Common
Clapham Junction
10. Other search input models
Gig finder
1. Choose a music genre
All (1234)
Blues, roots & country(62) Jazz (67)
Cabaret & standards (9) Kids' music (25)
Classical & opera (96) Pop, rock & hip-hop (87)
Dance music (25) R & B (56)
Folk (87) World (76)
2. Choose a date
March
Today | Tomorrow | This weekend
W T F S S M T W T F S S M T
14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | >>
11. Other search input models
http://www.engadget.com/2011/10/15/siri-ported-to-an-iphone-4-old-phone-learns-a-new-trick/
13. Intent in the content
Does it match what users are looking for?
Some examples:
• Product numbers of panel boards and switches, searched
by contractors and engineers. The company didn’t
structure the content to allow this to happen.
• Mealtimes in Time Out – more popular than cuisines. Is
the content tagged or have keywords in the content?
14. Location in London
Levels of granularity
• Postcode (TW9 2SU)
• Street (Oxford St)
• Town/village (Hammersmith, Camden)
• Area (South-West London)
• City-wide (London)
• How do we allow people to search them all? Do we even
capture this information?
16. Intent in the results
• Search results – do you show the information in the
results people are asking? Enough to make a decision?
• Business goals – are you consciously showing the
content the business wants to promote? Does it enhance
the user experience?
• Content types – are we showing the information people
want? More events for a music search; more venues for
restaurant search.
19. Analytical, not technical
You need to:
• Do granular analysis to understand the patterns
• Perform benchmarking of results
• Prototype search results ordering by knowing exactly the
impact of each decision like boosting of tags and search
field indexing
• Understand your content, structuring for patterns and
search as well as browse
• Show the working (this is where it is like A Level maths!)
20. Example – boosting by tags in Solr
Data element or tag Additional score
Rating score
5 100
4 80
3 60
2 40
1 20
Bookable 20
21. Example – boosting tags in Solr
Search Current results With boosts
Burgers 1. London’s best burgers (0) 1. Honest Burgers (Brixton) - 5 star
2. The Burger Bar - 3 star (60) (100)
3. Honest Burgers (Soho) - 4 star 2. Honest Burgers (Soho) - 4 star
(80) (80)
4. Guerrilla Burgers - 2 star and 3. The Burger Bar - 3 star (60)
bookable (60) 4. Guerrilla Burgers - 2 star and
5. Honest Burgers (Brixton) - 5 star bookable (60)
(100) 5. London’s best burgers (0)
22. Logical, not Mystical
• Reject “It’s how it is set-up” from the technical team. It can
be tuned based on rules, someone needs to define them.
• Most of us don’t work for Google – don’t need a maths
PhD to do this. BUT some basic maths is useful; think
Venn diagrams
• Understand the content and business requirements. They
will suggest the algorithm.
• And if you are not this way inclined – find someone who
is.
24. Activities
• Review your search logs and interface analytics
• Analyse the interface and failing searches to see intent and
understand ways to change it
• Test it in usability sessions – talk aloud when typing in queries
and what the people expected the results to be
• Understand your content and how it matches to the intent of
the users
• Structure the content better and leverage in the interface or
query engine
25. Thank you
Mags Hanley
Twitter: @magshanley
LinkedIn:
uk.linkedin.com/pub/margaret-hanley/0/180/3a5/
26. People
• Peter Morville (http://semanticstudios.com)
• Louis Rosenfeld (http://louisrosenfeld.com/home/)
• Tony Russell Rose (http://isquared.wordpress.com)
• Joe Lamantia (http://www.joelamantia.com)
• Tyler Tate (http://tylertate.com)
• Greg Nudelman
(http://www.androiddesignbook.com/about-greg-nudelman/)
27. Books
Information Architecture for the World Wide Web
http://www.amazon.com/Information-Architecture-World-
Wide-Web/dp/0596527349/findability-20/
Designing the Search Experience
http://designingthesearchexperience.com
Search analytics for your site
http://rosenfeldmedia.com/books/searchanalytics/
Search patterns
http://searchpatterns.org
Designing Search: UX Strategies for eCommerce Success
http://eu.wiley.com/WileyCDA/WileyTitle/productCd-
0470942231.html
Notas del editor
Which one:Is easier to develop on?Can I apply my facets?Follows the users’ intent?What about advanced search?Domain searchesDirected browse
Combinations of search that meet the mental model for that type of content