KEMBAR78
8 Information Architecture Better Practices | KEY
8 better practices
from information architecture


Lou Rosenfeld •  Rosenfeld Media •  rosenfeldmedia.com
Hello, my name is Lou




 www.louisrosenfeld.com | www.rosenfeldmedia.com
The state of
contemporary
  findability
Why can’t we get
findability right?
• Semantic illiteracy
• Siloed organizations
• Ill-equipped decision-makers prone to
  short-term thinking
• We don’t know how to diagnose
• We don’t know how to measure
Information architecture:
8 better practices for findability
 1. Diagnosing the important problems
 2. Balancing our evidence
 3. Designing for the long term
 4. Measuring engagement
 5. Supporting contextual navigation
 6. Improving search across silos
 7. Combining design approaches effectively
 8. Tuning our designs over time
#1
Diagnosing the
important problems
A handful of queries/tasks/ways to navigate/features/
 A little goes a long way
documents meet the needs of your most important audiences
A handful of queries/tasks/ways to navigate/features/
 A little goes a long way
documents meet the needs of your most important audiences
                                                  Not all queries
                                                   are distributed
                                                      equally
A handful of queries/tasks/ways to navigate/features/
 A little goes a long way
documents meet the needs of your most important audiences
A handful of queries/tasks/ways to navigate/features/
 A little goes a long way
documents meet the needs of your most important audiences
                             Nor do they
                        diminish gradually
A handful of queries/tasks/ways to navigate/features/
 A little goes a long way
documents meet the needs of your most important audiences
A handful of queries/tasks/ways to navigate/features/
 A little goes a long way
documents meet the needs of your most important audiences




                 80/20 rule isn’t
                 quite accurate
(and the tail is quite long)
(and the tail is quite long)
(and the tail is quite long)
(and the tail is quite long)
(and the tail is quite long)
The Zipf Curve, textually
It’s Zipf’s World;
we just live in it
 A little...
  •   queries

  •   tasks

  •   ways to navigate

  •   features

  •   documents

 ...goes a long way
unverified rumor alert
unverified rumor alert

90% of Microsoft.com content
unverified rumor alert

90% of Microsoft.com content
  has never been accessed...
unverified rumor alert

90% of Microsoft.com content
  has never been accessed...
        not even once
unverified rumor alert

90% of Microsoft.com content
  has never been accessed...
        not even once
#2
Balancing our evidence
from Christian Rohrer: http://is.gd/95HSQ2
Balanced research
                                             leads to true insight,
                                              new opportunities




from Christian Rohrer: http://is.gd/95HSQ2
Lou’s TABLE OF
OVERGENERALIZED            Web Analytics              User Experience
  DICHOTOMIES

                                                   Users' intentions and
    What they         Users' behaviors (what's
                                                   motives (why those things
     analyze          happening)
                                                   happen)

                                                   Qualitative methods for
 What methods         Quantitative methods to
                                                   explaining why things
  they employ         determine what's happening
                                                   happen

                                                   Helps users achieve goals
  What they're    Helps the organization meet
                                                   (expressed as tasks or
trying to achieve goals (expressed as KPI)         topics of interest)

                                                   Uncover patterns and
  How they use        Measure performance (goal-
                                                   surprises (emergent
     data             driven analysis)
                                                   analysis)

                  Statistical data ("real" data    Descriptive data (in small
What kind of data
                  in large volumes, full of        volumes, generated in lab
   they use       errors)                          environment, full of errors)
#3
Designing for the long-term
Stewart Brand’s Pace Layering
model              Typical design
                       focus




                     Stuff that gets ignored:
                     mission, vision, charter,
                      goals, KPI, objectives
#4
Measuring engagement
Measuring
conversions?
No problem...
..measuring
anything else?
 Good luck!
The missing metrics
of engagement
• Orientation (“What can I do here?”)
• Authority (“I trust this”)
• Social (“Who else likes this?”)
• Connection/cross-promotion (“What goes
  with this?”)
• and many more...
Conversation architecture
uncovers levels of engagement
Level 0: I visit site (unauthenticated)
Level 1: I ask site a question (e.g., a search)
Level 2: Site asks me a question (“can we save those
settings?”)
Level 3: Site suggests something
to me (“you might also like this”)
Level 4: Site acts on my behalf
(“I’ve added this to your favorites
 list in case you’d like to reorder”)
                            trust and value grow progressively
#5
Supporting
contextual navigation
Contextual navigation:
your site’s desire lines
1.Choose a content
type (e.g., events)
                                      



2.Ask: “Where
should users go
from here?”

3.Analyze the                             


frequent queries
from this content
type

                      from aiga.org
                                              






                             





                                                          

                                 





                                     
                    




Analyze frequent queries generated from each content sample
Content models emerge (example: BBC)     concert calendar




  album pages     artist descriptions
                                              TV listings




album reviews    discography            artist bios
User studies are another great
way to get at content models
#6
Improving search across silos
Reconsidering the search UI...
...convert “advanced” search
into refinement
...convert “advanced” search
into refinement
 search session patterns
 1. solar energy
 2. how solar energy works
...convert “advanced” search
into refinement
  search session patterns
  1. solar energy
  2. how solar energy works




 search session patterns
 1. solar energy
 2. energy
...convert “advanced” search
into refinement                search session patterns
  search session patterns     1. solar energy
  1. solar energy             2. solar energy charts
  2. how solar energy works




 search session patterns
 1. solar energy
 2. energy
...convert “advanced” search
into refinement                    search session patterns
  search session patterns         1. solar energy
  1. solar energy                 2. solar energy charts
  2. how solar energy works




                           search session patterns
 search session patterns   1. solar energy
 1. solar energy           2. explain solar energy
 2. energy
...convert “advanced” search
into refinement                        search session patterns
  search session patterns             1. solar energy
  1. solar energy                     2. solar energy charts
  2. how solar energy works




                             search session patterns
 search session patterns     1. solar energy
 1. solar energy             2. explain solar energy
 2. energy


                           search session patterns
                           1. solar energy
                           2. solar energy news
Design for
specialized queries
(e.g., proper nouns,
dates, unique ID#s)
...and design specialized search results
...and design specialized search results
...and design specialized search results
Content objects
 from product
content model

       ...and design specialized search results
Poor search
                                   results returned
                                   by search engine




Content objects
 from product
content model

       ...and design specialized search results
#7
Combining design approaches
effectively
Yes, manual effort is still as
important as tools
Yes, manual effort is still as
important as tools

                   Narrow, deep
                   content access
Vanguard’s Tax Center is a
simple, low-tech, editorial
Vanguard’s Tax Center is a
simple, low-tech, editorial


      ...to editorially
        rich content
Manually
selected results
Manually
selected results




   ...complement
      raw results
Treat your content
                                                                      Each layer is
                                                                       cumulative;
                                                                    most important

        like an onion                                                 content is at
                                                                           the core



            information
layer                                  usability           content strategy
            architecture
          indexed by search
 0             engine
                                     leave it alone           leave it alone

                                 squeaky wheel issues
 1          tagged by users
                                      addressed
                                                             refresh annually

        tagged by experts (non-   test with a service
 2            topical tags)     (e.g., UserTesting.com)
                                                             refresh monthly

          tagged by experts      “traditional” lab-based    titled according to
 3           (topical tags)           user testing               guidelines
          content models for                               structured according
 4       contextual navigation
                                      A/B testing
                                                                to schema
#8
Tuning designs over time
Your site is a moving target
built on moving targets
Impact of change on design
(queries)
Impact of change on design
(queries)


 Interest in the
  football team:
     going...
Impact of change on design
(queries)


 Interest in the
  football team:
     going...

                   ...going...
Impact of change on design
(queries)


 Interest in the
  football team:
     going...

                   ...going...



                                 gone
Impact of change on design
(queries)


                                        Time to
 Interest in the
                                         study!
  football team:
     going...

                   ...going...



                                 gone
IRS before Tax Day
Before
              Tax Day
IRS before Tax Day
IRS after Tax Day
After
               Tax Day
IRS after Tax Day
Summary:
8 IA better practices
1. Diagnosing the important problems
2. Balancing our evidence
3. Designing for the long term
4. Measuring engagement
5. Supporting contextual navigation
6. Improving search across silos
7. Combining design approaches effectively
8. Tuning our designs over time
Say hello


 Lou Rosenfeld
 lou@louisrosenfeld.com
 Rosenfeld Media 
 www.louisrosenfeld.com | @louisrosenfeld
 www.rosenfeldmedia.com | @rosenfeldmedia

8 Information Architecture Better Practices

  • 1.
    8 better practices frominformation architecture Lou Rosenfeld •  Rosenfeld Media •  rosenfeldmedia.com
  • 2.
    Hello, my nameis Lou www.louisrosenfeld.com | www.rosenfeldmedia.com
  • 4.
  • 5.
    Why can’t weget findability right? • Semantic illiteracy • Siloed organizations • Ill-equipped decision-makers prone to short-term thinking • We don’t know how to diagnose • We don’t know how to measure
  • 6.
    Information architecture: 8 betterpractices for findability 1. Diagnosing the important problems 2. Balancing our evidence 3. Designing for the long term 4. Measuring engagement 5. Supporting contextual navigation 6. Improving search across silos 7. Combining design approaches effectively 8. Tuning our designs over time
  • 7.
  • 8.
    A handful ofqueries/tasks/ways to navigate/features/ A little goes a long way documents meet the needs of your most important audiences
  • 9.
    A handful ofqueries/tasks/ways to navigate/features/ A little goes a long way documents meet the needs of your most important audiences Not all queries are distributed equally
  • 10.
    A handful ofqueries/tasks/ways to navigate/features/ A little goes a long way documents meet the needs of your most important audiences
  • 11.
    A handful ofqueries/tasks/ways to navigate/features/ A little goes a long way documents meet the needs of your most important audiences Nor do they diminish gradually
  • 12.
    A handful ofqueries/tasks/ways to navigate/features/ A little goes a long way documents meet the needs of your most important audiences
  • 13.
    A handful ofqueries/tasks/ways to navigate/features/ A little goes a long way documents meet the needs of your most important audiences 80/20 rule isn’t quite accurate
  • 14.
    (and the tailis quite long)
  • 15.
    (and the tailis quite long)
  • 16.
    (and the tailis quite long)
  • 17.
    (and the tailis quite long)
  • 18.
    (and the tailis quite long)
  • 19.
    The Zipf Curve,textually
  • 20.
    It’s Zipf’s World; wejust live in it A little... • queries • tasks • ways to navigate • features • documents ...goes a long way
  • 22.
  • 23.
    unverified rumor alert 90%of Microsoft.com content
  • 24.
    unverified rumor alert 90%of Microsoft.com content has never been accessed...
  • 25.
    unverified rumor alert 90%of Microsoft.com content has never been accessed... not even once
  • 26.
    unverified rumor alert 90%of Microsoft.com content has never been accessed... not even once
  • 27.
  • 28.
    from Christian Rohrer:http://is.gd/95HSQ2
  • 29.
    Balanced research leads to true insight, new opportunities from Christian Rohrer: http://is.gd/95HSQ2
  • 30.
    Lou’s TABLE OF OVERGENERALIZED Web Analytics User Experience DICHOTOMIES Users' intentions and What they Users' behaviors (what's motives (why those things analyze happening) happen) Qualitative methods for What methods Quantitative methods to explaining why things they employ determine what's happening happen Helps users achieve goals What they're Helps the organization meet (expressed as tasks or trying to achieve goals (expressed as KPI) topics of interest) Uncover patterns and How they use Measure performance (goal- surprises (emergent data driven analysis) analysis) Statistical data ("real" data Descriptive data (in small What kind of data in large volumes, full of volumes, generated in lab they use errors) environment, full of errors)
  • 31.
  • 32.
    Stewart Brand’s PaceLayering model Typical design focus Stuff that gets ignored: mission, vision, charter, goals, KPI, objectives
  • 33.
  • 36.
  • 39.
  • 40.
    The missing metrics ofengagement • Orientation (“What can I do here?”) • Authority (“I trust this”) • Social (“Who else likes this?”) • Connection/cross-promotion (“What goes with this?”) • and many more...
  • 41.
    Conversation architecture uncovers levelsof engagement Level 0: I visit site (unauthenticated) Level 1: I ask site a question (e.g., a search) Level 2: Site asks me a question (“can we save those settings?”) Level 3: Site suggests something to me (“you might also like this”) Level 4: Site acts on my behalf (“I’ve added this to your favorites list in case you’d like to reorder”) trust and value grow progressively
  • 42.
  • 43.
  • 44.
    1.Choose a content type(e.g., events) 
 2.Ask: “Where should users go from here?” 3.Analyze the 
 frequent queries from this content type from aiga.org 

  • 45.
    
 
 
 
 
 Analyze frequent queries generated from each content sample
  • 46.
    Content models emerge(example: BBC) concert calendar album pages artist descriptions TV listings album reviews discography artist bios
  • 47.
    User studies areanother great way to get at content models
  • 48.
  • 49.
  • 50.
  • 51.
    ...convert “advanced” search intorefinement search session patterns 1. solar energy 2. how solar energy works
  • 52.
    ...convert “advanced” search intorefinement search session patterns 1. solar energy 2. how solar energy works search session patterns 1. solar energy 2. energy
  • 53.
    ...convert “advanced” search intorefinement search session patterns search session patterns 1. solar energy 1. solar energy 2. solar energy charts 2. how solar energy works search session patterns 1. solar energy 2. energy
  • 54.
    ...convert “advanced” search intorefinement search session patterns search session patterns 1. solar energy 1. solar energy 2. solar energy charts 2. how solar energy works search session patterns search session patterns 1. solar energy 1. solar energy 2. explain solar energy 2. energy
  • 55.
    ...convert “advanced” search intorefinement search session patterns search session patterns 1. solar energy 1. solar energy 2. solar energy charts 2. how solar energy works search session patterns search session patterns 1. solar energy 1. solar energy 2. explain solar energy 2. energy search session patterns 1. solar energy 2. solar energy news
  • 56.
    Design for specialized queries (e.g.,proper nouns, dates, unique ID#s)
  • 57.
  • 58.
  • 59.
  • 60.
    Content objects fromproduct content model ...and design specialized search results
  • 61.
    Poor search results returned by search engine Content objects from product content model ...and design specialized search results
  • 62.
  • 63.
    Yes, manual effortis still as important as tools
  • 64.
    Yes, manual effortis still as important as tools Narrow, deep content access
  • 65.
    Vanguard’s Tax Centeris a simple, low-tech, editorial
  • 66.
    Vanguard’s Tax Centeris a simple, low-tech, editorial ...to editorially rich content
  • 68.
  • 69.
    Manually selected results ...complement raw results
  • 70.
    Treat your content Each layer is cumulative; most important like an onion content is at the core information layer usability content strategy architecture indexed by search 0 engine leave it alone leave it alone squeaky wheel issues 1 tagged by users addressed refresh annually tagged by experts (non- test with a service 2 topical tags) (e.g., UserTesting.com) refresh monthly tagged by experts “traditional” lab-based titled according to 3 (topical tags) user testing guidelines content models for structured according 4 contextual navigation A/B testing to schema
  • 71.
  • 72.
    Your site isa moving target built on moving targets
  • 73.
    Impact of changeon design (queries)
  • 74.
    Impact of changeon design (queries) Interest in the football team: going...
  • 75.
    Impact of changeon design (queries) Interest in the football team: going... ...going...
  • 76.
    Impact of changeon design (queries) Interest in the football team: going... ...going... gone
  • 77.
    Impact of changeon design (queries) Time to Interest in the study! football team: going... ...going... gone
  • 78.
  • 79.
    Before Tax Day IRS before Tax Day
  • 80.
  • 81.
    After Tax Day IRS after Tax Day
  • 82.
    Summary: 8 IA betterpractices 1. Diagnosing the important problems 2. Balancing our evidence 3. Designing for the long term 4. Measuring engagement 5. Supporting contextual navigation 6. Improving search across silos 7. Combining design approaches effectively 8. Tuning our designs over time
  • 83.
    Say hello LouRosenfeld lou@louisrosenfeld.com Rosenfeld Media  www.louisrosenfeld.com | @louisrosenfeld www.rosenfeldmedia.com | @rosenfeldmedia

Editor's Notes