Video available: http://www.designforcontext.com/insights/designing-next-generation-search-user-experience
Search applications aren’t "just like Google" anymore – even Google is no longer the simple application it once was. Design is coming to the forefront of effective search applications, to help make sense of mobile search, data search, semantic search, enterprise search, federated search, and embedded search within websites and applications.
So what do we need to know about designing for search? We need to understand our users’ mental models for how they perceive seeking within an information environment. We need to understand how to provide powerful user control over results and yet keep it extremely simple. We need to know how to test for effective comprehension as well as task execution. And we need to know how to get the most out of the new available technologies and data. This UXPA 2015 talk is a deep dive into the essentials for a new generation of search designs.
2. @design4context
NextGen Search #UXPA2015
2
USER
EXPERIENCE
&
USABILITY
Understand user needs and
tasks. Design applicaHons
that work the way people
expect them to.
Photo
of
addiConal
presenter
here
(in
grayscale)
Lisa
Ba.le
INFORMATION
ARCHITECTURE
/
DESIGN
Understand the models that
make informaHon Hck, and
help people get things done.
Duane
Degler
3. @design4context
NextGen Search #UXPA2015
3
The
ubiquitous
search
box
4. @design4context
NextGen Search #UXPA2015
4
What
can
you
do
to
influence
the
user’s
experience
with
search?
5. @design4context
NextGen Search #UXPA2015
5
Public
Search
Site
Search
Enterprise
Search
What
is
a
search
experience?
6. @design4context
NextGen Search #UXPA2015
Stages
of
seeking
Catalyst
Info Task
Achieve in the
search space
Goal
Achieve in the
world
Seek
Type in the box
Evaluate
Assess results
Review
Assess content
Formulate
Strategy
Tactics
AcCvity
Use
Apply what you
gained
Keep
Persistent
resource
Outcome
REFINE
ITERATE
7. @design4context
NextGen Search #UXPA2015
Seek
Type in the box
Evaluate
Assess results
Review
Assess content
Formulate
Strategy
Tactics
REFINE
User’s
perspecCve
“Mechanics”
of
search
Trends
for
the
future
Stages
of
seeking
8. @design4context
NextGen Search #UXPA2015
8
http://en.wikipedia.org/wiki/Incandescent_light_bulb#mediaviewer/
File:Edison_incandescent_lights.jpg
The
user’s
mental
models,
moCvaCons
and
level
of
understanding
affect
their
approach
to
search.
TesHng search has some nuances.
USERS’
PERSPECTIVE
9. @design4context
NextGen Search #UXPA2015
9
SEARCH STYLE:
Survey
“
What
are
the
symptoms
of
arthri0s?
”
» SaCsficing,
"good
enough"
» Usually
single
session
» Look
for
correlaCons
between
results
to
bolster
confidence
» Focused
on
authoritaCve
sources
» Preview
to
support
decision-‐making
REFINE
Formulate Seek Evaluate Review
10. @design4context
NextGen Search #UXPA2015
10
SEARCH STYLE:
Targeted
"What
is
the
maximum
dosage
for
aspirin
per
day?"
» Typically
start
with
a
known
domain
» Clear,
known
expectaCon
of
outcome
or
nature
of
outcome
» High
recogniCon
when
evaluaCng
result
set
REFINE
Formulate Seek Evaluate Review
KEEP
11. @design4context
NextGen Search #UXPA2015
11
“
I
need
to
write
a
research
paper
on
effec0veness
of
physical
therapy
vs.
drugs
for
arthri0s
treatment”
» Longitudinal
» IteraCve,
layered,
evolving
» Search
strategy
is
a
criCcal
component
» Inputs
to
search
formulaCon
must
be
acConable
» Outcomes
must
be
acConable:
capture,
organize,
save
REFINE
Formulate Seek Evaluate Review
KEEP
SEARCH STYLE:
Exploratory
12. @design4context
NextGen Search #UXPA2015
12
SEARCH STYLE:
Archival
”In
what
situa0ons
has
methotrexate
caused
or
contributed
to
pa0ent
adverse
events?"
» Search
has
been
comprehensive
and
complete
» Important
to
know
you
have
found
all
that
there
is
(within
limit)
» Haven't
missed
criCcal
findings
» Comprehensive
search
is
itself
a
product;
if
I
can't
find
it:
• it
likely
doesn't
exist,
or
• It’s
not
relevant
to
the
context
that
I
have
applied
REFINE
Formulate Seek Evaluate Review
KEEP
13. @design4context
NextGen Search #UXPA2015
13
SEARCH STYLE:
RouCne
"Keep
me
up-‐to-‐date
on
hematological
cancer
involving
Velcade
as
an
agent"
» Take
a
successful
search
strategy
and
give
it
an
idenCty
» Focus
on
new
and
unique
results
» Search
parameters
and
filters
are
not
necessarily
related
to
a
specific
result
set
REFINE
Formulate Seek Evaluate Review
KEEP
14. @design4context
NextGen Search #UXPA2015
14
» Agent
for
others
» Sharing
» Refinement
might
be
mulC-‐session,
iteraCve,
not
real-‐Cme
» FormulaCon
task
requires:
• NegoCaCon
to
arCculate
goals
and
scope
• TranslaCon
REFINE
Formulate Seek Evaluate Review
KEEP
SEARCH STYLE:
CollaboraCve
"What
is
the
maximum
dosage
for
aspirin
per
day?"
”Is
this
for
an
adult
or
a
child?"
”A
child."
”
How
many
milligrams
is
your
aspirin
pill?"
16. @design4context
NextGen Search #UXPA2015
● Set
the
scene
● Let
the
user
define
scenarios
whenever
possible
● Be
clear,
together,
about
the
context
that
you
are
tesCng
for,
so
the
person
can
get
in
that
mental
space
● Test
in
their
environment
whenever
possible
• Comfort
• Secondary
resources
● IdenCfy
expected
outcomes
● Dig
to
learn
what
the
person
wants
to
do
with
informaCon
● Collect
examples
of
work
products
/
outputs
they
create
now
ConsideraCons
for
tesCng
search
designs
16
17. @design4context
NextGen Search #UXPA2015
● It’s
a
thinking
process
● Test
with
real
content
–
as
much
as
you
can
get
● Use
content
that
is
meaningful
/
familiar
● Encourage
conversaCon
–
formaCve
protocols
encouraged
● Iterate
rouCnely
and
rapidly
● Mix
your
parCcipants
–
some
regular,
some
new
each
Cme
● Work
with
real
code
in
the
back-‐end,
lightweight
interacCve
prototypes
on
the
front-‐end
• The
experience
is
tacCle
and
cogniCve
● Include
a
longitudinal
study
● Search
tools
for
large
content
environments
and
enterprise
take
Cme
for
a
person
to
incorporate
into
a
rouCne
● ReflecCon
is
an
important
aspect
of
user
feedback
ConsideraCons
for
tesCng
search
designs
17
19. @design4context
NextGen Search #UXPA2015
The
mechanics
of
search
Index
Query
Parser
Search
Engine
Document
Unstructured
Content
Models
Metadata,
Terms,
Thesaurus
Results
&
Facets
List
of
items
List
of
terms
and
counts
Results
Page
Text,
Image
Voice,
20. @design4context
NextGen Search #UXPA2015
Unstructured
Content
Image
The
mechanics
of
search
.
.
.
enhanced
Index
Query
Parser
Search
Engine
Document
Voice,
Models
Metadata,
Taxonomy,
Thesaurus,
Ontology
Results
&
Facets
List
of
items
List
of
terms
and
counts
Results
Page
FUTURE
Text,
ApplicaCon
23. @design4context
NextGen Search #UXPA2015
Deciding
what
to
search
for,
and
how
What
do
I
need
to
find?
Where
should
I
look?
How
is
this
different
than
what
I
searched
for
before?
What
words
or
phrases
should
I
use
to
describe
the
subject?
24. @design4context
NextGen Search #UXPA2015
Formula/ng
aligns
goals
with
concepts
about
informaCon
The Design of Browsing and Berrypicking Techniques for the Online Search Interface
Marcia Bates, 1989, http://pages.gseis.ucla.edu/faculty/bates/berrypicking.html
Berry
picking
ê
Evolving
understanding
How
do
we
help
users
align
their
goals
with
the
available
informaCon
space?
25. @design4context
NextGen Search #UXPA2015
RecogniCon,
not
recall:
Search
suggesCons
CURRENT
Consider
how
you
retrieve
and
structure
the
terms
you
present.
26. @design4context
NextGen Search #UXPA2015
Capture
thinking
process
when
it
happens
FUTURE
The increase in OA-related musculoskeletal pain especially in
the growing elderly population has created heightened interest in
effective treatments with better safety profiles. The Glucosamine/
chondroitin Arthritis Intervention Trial (GAIT) was designed to
examine the effects of the dietary supplements glucosamine and
chondroitin sulfate given alone or in combination as compared to
celecoxib or placebo on pain associated with osteoarthritis of the
knee. The primary symptomatic outcome assessment of the
GAIT study was after 24 weeks of randomized, double-blinded,
placebo-controlled therapy.[6] An ancillary structural study
describing the effects of the agents on radiographic joint space
width loss for up to two years has also recently been reported.[7]
To date, very few data have been reported for the long-term
treatment of OA with any agent including the supplements and
celecoxib studied here. Patients in the structural study continued
to have safety and clinical efficacy assessments at all scheduled
visits. This paper details the clinical efficacy and safety
experience with these agents alone and in combination along
with celecoxib as compared to placebo in patients from this
subset of GAIT over 24 months offollow-up.
chondroitin sulfate
celcoxib
radiographic joint space
Add note, action
Search on this now
Tag for!Tag for search!
Prior
search
formulaCons
and
strategies
are
products
Search Phrase
Search Builder Assist:
Operator Toolbar
Special Symbols
Equations and Formulas
arthritis +4
and methotrexate
arthritis – synonyms and instances
" Ankylosing spondylitis
þ Gonococcal arthritis
" Gout
" Juvenile rheumatoid arthritis
þ Psoriatic arthritis
þ Reactive arthritis (Reiter syndrome)
þ Rheumatoid arthritis (in adults)
" Scleroderma
" Systemic lupus erythematosus (SLE)
+ Add another
27. @design4context
NextGen Search #UXPA2015
RouCne,
automated
search
CURRENT
How
will
these
searches
evolve
as
a
person’s
interests
and
needs
change?
Think
longitudinally.
28. @design4context
NextGen Search #UXPA2015
Under
the
covers,
using
AI,
this
is
about
pa.ern-‐matching.
Seqng
personalizaCon
30. @design4context
NextGen Search #UXPA2015
Refine
personalizaCon
via
richer
models
FUTURE
Showing
aspects
of
structure
helps
someone
focus
on
their
specific
interests.
31. @design4context
NextGen Search #UXPA2015
Unstructured
Content
Image
The
mechanics
of
search
Index
Query
Parser
Search
Engine
Document
Voice,
Models
Metadata,
Taxonomy,
Thesaurus,
Ontology
Results
&
Facets
List
of
items
List
of
terms
and
counts
Results
Page
FUTURE
Profile
Text,
SemanCc
Processing
Save
User
Model
Index/
Graph
db
Formulate
32. @design4context
NextGen Search #UXPA2015
32
● Help
a
person
make
needs
and
interests
more
explicit
● Capture
what
they
are
thinking
before
they
start
searching,
in
context
of
other
tasks
and
applicaCons
● Find
streamlined
ways
to
support
refining
and
changing
interests
over
Cme
● RouCne
searches
decay
–
refinement
extends
engagement
● Surface
underlying
models
● ParCcularly
for
knowledge-‐intensive
and
longitudinal,
mulC-‐session
tasks
● Be
humble
–
and
create
a
humble
experience
Design
consideraCons
Formulate
34. @design4context
NextGen Search #UXPA2015
CURRENT
Keywords,
type
ahead
and
scoping
35. @design4context
NextGen Search #UXPA2015
CURRENT
Mobile
search
relies
heavily
on
type-‐ahead
36. @design4context
NextGen Search #UXPA2015
CURRENT
DisambiguaCon
Present
informaCon
that
is
recognizable
and
supports
user
decision-‐making.
37. @design4context
NextGen Search #UXPA2015
Advanced
search
CURRENT
The
common
trend
is
to
move
detailed
parameter
refinement
to
Results
page,
as
facet
filters,
to
avoid
“0
results”…
38. @design4context
NextGen Search #UXPA2015
Parameter-‐driven
“advanced”
searches
CURRENT
Parameters
are
the
same
(to
the
user,
in
the
code)
for
the
search
and
for
facet
management
on
the
results
page.
The
OpCcal
Society
(OSA),
h.ps://osapublishing.org
39. @design4context
NextGen Search #UXPA2015
EnCty
disambiguaCon
from
a
model
FUTURE
40. @design4context
NextGen Search #UXPA2015
Parsing:
Exposing
thesaurus
for
user
control
For
very
knowledgeable
researchers,
it
builds
trust
when
you
are
transparent
and
offer
control.
42. @design4context
NextGen Search #UXPA2015
What
states
surround
Missouri?
InterpretaCon
of
search
phrases
42
43. @design4context
NextGen Search #UXPA2015
What
is
the
populaCon
in
the
states
surrounding
Missouri?
InterpretaCon
of
search
phrases
43
Geographic
bordering
relaConships
US
State
as
focus
State
(geographic)
Number
of
residents
44. @design4context
NextGen Search #UXPA2015
Calculate
difference
Specific
Cme
period
for
data
Geographic
bordering
relaConships
US
State
as
focus
State
(geographic)
Number
of
residents
What
is
the
change
in
populaCon
in
the
states
surrounding
Missouri,
between
2012-‐2014?
InterpretaCon
of
search
phrases
44
FUTURE
And
this
is
sCll
a
simple
quesCon,
when
compared
with
what
researchers
and
analysts
ask
of
their
data
today.
45. @design4context
NextGen Search #UXPA2015
Smarter
search
boxes
(clinical decision support) (device or console or computer or apparatus)AND
Q-22 «
OR OR OR SAME Q-17 »
“clinical decision”
“decision support”
“clinical decision support”
Phrase
structures
support
supported
supporting
supportive
Stemming
Synonyms
(thesauri)
Operator combinations Dynamic controls
FUTURE
46. @design4context
NextGen Search #UXPA2015
Image
search
46
FUTURE
http://culturecam.eu
AlternaCve
entry
methods
for
large
amounts
of
rich
media
should
be
explored.
47. @design4context
NextGen Search #UXPA2015
Voice-‐acCvated
search…
a
conversaCon
FUTURE
Try
building
in
comprehension
check-‐points.
48. @design4context
NextGen Search #UXPA2015
ProacCve
search:
Solving
a
real-‐world
content
problem
48
User’sawarenessofneed
Nature of content related to task
Prohibited Not available
Specific awareness of
content and ability to
articulate problem
Required Available
Awareness of problem
but little ability to
articulate
No awareness of need
for information
“I don’t know
what I need to know”
“I know what I’m
looking for--
get me to it!”
Identify
via analyzing
content and
term use
Identify
via direct user
feedback
Identify
via task
performance
assessment
“I didn’t know
I needed that”
Degler/Ba.le
(2003).
Can
Topic
Maps
Describe
Enterprise-‐Wide
Context?
h.p://www.designforcontext.com/publicaCons/dd-‐lb_topic-‐maps_context_extreme2003.pdf
“Just tell me
what I
absolutely
must know”
49. @design4context
NextGen Search #UXPA2015
Unstructured
Content
Image
The
mechanics
of
search
Index
Query
Parser
Search
Engine
Document
Voice,
Models
Metadata,
Taxonomy,
Thesaurus,
Ontology
Results
&
Facets
List
of
items
List
of
terms
and
counts
Results
Page
Speech
Language
Images
FUTURE
Profile
Context
Parameters
Text,
SemanCc
Processing
Terms,
Tags
Interpreted
meaning
NLP,
Enrich,
Auto-‐Tag
Unstructured
Content
Semi-‐structured
Index/
Graph
db
Seek
50. @design4context
NextGen Search #UXPA2015
50
● Simple
is
best
.
.
.
● But
no
simpler
that
what
supports
a
person
to
express
what
they
need
as
precisely
as
they
are
able
● Allow
someone
to
use
examples
wherever
possible
● The
trend
is
to
move
complexity
to
the
results,
because
decisions
are
evaluated
against
real
informaCon
● Think
iteraCvely
.
.
.
And
longitudinally
● In
many
cases,
search
is
a
mulC-‐step
and
mulC-‐session
task
(exploratory,
archival,
collaboraCve
in
parCcular)
● Be
humble
–
and
create
a
humble
experience
Design
consideraCons
Seek
52. @design4context
NextGen Search #UXPA2015
Results
page
Refinements
User
choices
Ac/ons
View
controls
Results
Meta-‐results
CURRENT
(Ranked
Instances)
(Facets)
53. @design4context
NextGen Search #UXPA2015
Facets
53
CURRENT
HierarchyMulti-select Incremental DisclosureSimple list
Scoping
54. @design4context
NextGen Search #UXPA2015
Managing
the
evaluaCon
process…
CURRENT
These
habits
will
remain
–
consider
how
to
work
with
them.
55. @design4context
NextGen Search #UXPA2015
Mobile
results
–
answering
direct
quesCons
Shazam
CURRENT
Google
Maps
56. @design4context
NextGen Search #UXPA2015
Mobile
results
–
could
a
conversaCon
help?
iTriage
CURRENT
WebMD
How
quickly
do
you
go
to
individual
results
items?
57. @design4context
NextGen Search #UXPA2015
Answers
instead
of
“results”
(“Things,
not
Strings”
-‐
2012)
FUTURE
59. @design4context
NextGen Search #UXPA2015
Exploratory:
Blend
results
&
topics
(Intent
Streams)
Salvatore Andolina, Khalil Klouche, Jaakko Peltonen, Mohammad Hoque, Tuukka Ruotsalo, Diogo Cabral, Arto Klami, Dorota Głowacka, Patrik Floréen, and Giulio Jacucci. 2015.
IntentStreams: Smart Parallel Search Streams for Branching Exploratory Search. In Proceedings of the 20th International Conference on Intelligent User Interfaces (IUI '15). ACM,
New York, NY, USA, 300-305. DOI=10.1145/2678025.2701401 http://doi.acm.org/10.1145/2678025.2701401
FUTURE
60. @design4context
NextGen Search #UXPA2015
VisualizaCons:
User
control
is
key
FUTURE
Antti Kangasrääsiö, Dorota Glowacka, and Samuel Kaski. 2015. Improving Controllability and Predictability of Interactive Recommendation Interfaces for
Exploratory Search. In Proceedings of the 20th International Conference on Intelligent User Interfaces (IUI '15). ACM, New York, NY, USA, 247-251.
DOI=10.1145/2678025.2701371 http://doi.acm.org/10.1145/2678025.2701371
61. @design4context
NextGen Search #UXPA2015
Snippets
Snippets
displayed and
highlighted
upon
mouseover
Upon click,
display snippet
expansion
Transition display to single
document view/layer,
anchoring to the specific
snippet in context
FUTURE
62. @design4context
NextGen Search #UXPA2015
Expanded
snippets
Title
Number
Date
Class
Inventor
Type
Preview
Assisted Clinical Decision Support 20040091847 11/06/2003 434/353 Pandya, et al. Patent
Claim Text:
assisted clinical decision support, the method comprising: identifying, with a processor, a
plurality of associated terms from a medical ontology, the associated terms including
associated drugs; generating...
assisted clinical decision support, the system comprising: a memory operable to store a mining
engine with a domain-knowledge base from associated terms in a medical ontology, the mining
engine configured...
assisted clinical decision support, the storage media comprising instructions for: creating a
knowledge base from a medical ontology; and building a mining engine operable to search
unstructured medical data...
assisted clinical decision support, the method comprising: identifying a plurality of associated
terms from a medical ontology; and training, with a processor, a model of relationships of the
associated terms...
assisted clinical decision support, the system comprising: a memory operable to store a
probabilistic model having machine learned probabilities for relationships from a medical
ontology, the machine learned...
assisted clinical decision support, the method comprising: identifying, with a processor, a
plurality of associated terms from a medical ontology; generating, with the processor, a domain-
knowledge base from...
89 pages | Cover page | 12 drawings
#$
forward-backward citations
X
assisted clinical decision support, the system comprising: a memory
operable to store a mining engine with a domain-knowledge base from
associated terms in a medical ontology, the mining engine configured
Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do
eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad
minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip
Duis aute irure dolor in reprehenderit in voluptate velit esse cillum
dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non
proident, sunt in culpa qui officia deserunt mollit anim id est laborum.
#$
434/353: Specifications Pg XX
FUTURE
Help
evaluaCon
at
different
informaCon
levels.
63. @design4context
NextGen Search #UXPA2015
Unstructured
Content
Image
The
mechanics
of
search
Index
Query
Parser
Search
Engine
Document
Voice,
Models
Metadata,
Taxonomy,
Thesaurus,
Ontology
Results
&
Facets
List
of
items
List
of
terms
and
counts
Results
Page
FUTURE
Structured
Data
Panels
Graph
Model
Images
&
MulCmedia
Profile
Context
Parameters
Text,
SemanCc
Processing
User
Model
Terms,
Tags
Interpreted
meaning
NLP,
Enrich,
Auto-‐Tag
Unstructured
Content
Semi-‐structured
Index/
Graph
db
Evaluate
64. @design4context
NextGen Search #UXPA2015
64
● The
results
page
is
a
query
interface
● Provide
sufficient
signposCng
and
user
control
● Allow
for
quick
skimming
and,
for
knowledge-‐intensive
tasks,
incrementally
deeper
dives
● Consider
capabiliCes
for
exploratory
behaviors
● Design
to
suit
the
task
and
mental
model
● Consider
“how
much
is
sufficient”
● Establish
trust
in
the
results
–
algorithms
need
transparency
● Be
humble
–
and
create
a
humble
experience
Design
consideraCons
Evaluate
66. @design4context
NextGen Search #UXPA2015
Snippets
Snippets
displayed and
highlighted
upon
mouseover
Upon click,
display snippet
expansion
Transition display to single
document view/layer,
anchoring to the specific
snippet in context
FUTURE
67. @design4context
NextGen Search #UXPA2015
People
organize
their
thinking
in
different
ways
at
different
Cmes.
Sensemaking
and
organizing
FUTURE
Marti A. Hearst and Duane Degler. 2013. Sewing the Seams of Sensemaking: A Practical Interface for Tagging and Organizing Saved Search Results.
In Proceedings of the Symposium on Human-Computer Interaction and Information Retrieval (HCIR '13). ACM, New York, NY, USA, , Article 4 , 10 pages.
DOI=10.1145/2528394.2528398 http://doi.acm.org/10.1145/2528394.2528398
69. @design4context
NextGen Search #UXPA2015
The
World
Bank
invests
about
one-‐quarter
of
its
budget
for
country
services
in
knowledge
products…
About
49
percent
of
the
World
Bank’s
policy
reports,
which
are
published
Economic
and
Sector
Work
or
Technical
Assistance
reports,
have
the
stated
objecCve
of
informing
the
public
debate
or
influencing
the
development
community…
About
13
percent
of
policy
reports
were
downloaded
at
least
250
Cmes
while
more
than
31
percent
of
policy
reports
are
never
downloaded.
Almost
87
percent
of
policy
reports
were
never
cited.
Granular
content
instead
of
“documents”
69
FUTURE
Doemeland, Doerte; Trevino, James. 2014. Which World Bank reports are widely read ? Policy Research
working paper ; no. WPS 6851. Washington, DC: World Bank Group. http://documents.worldbank.org/
curated/en/2014/05/19456376/world-bank-reports-widely-read-world-bank-reports-widely-read
70. @design4context
NextGen Search #UXPA2015
Craig S. Tashman and W. Keith Edwards. 2011. LiquidText: a
flexible, multitouch environment to support active reading. In
Proceedings of the SIGCHI Conference on Human Factors in
Computing Systems (CHI '11). ACM, New York, NY, USA,
3285-3294. DOI=10.1145/1978942.1979430 http://doi.acm.org/
10.1145/1978942.1979430
Summarize
&
compress
(LiquidText)
FUTURE
71. @design4context
NextGen Search #UXPA2015
FUTURE
ReflecCng
task,
user
and
content
relevance
71
73. @design4context
NextGen Search #UXPA2015
Unstructured
Content
Image
The
mechanics
of
search
Index
Query
Parser
Search
Engine
Document
Voice,
Models
Metadata,
Taxonomy,
Thesaurus,
Ontology
Results
&
Facets
List
of
items
List
of
terms
and
counts
Fragment
Results
Page
ApplicaCon
Embedded
Results
Panels
FUTURE
Context
Parameters
Text,
Graph
Model
Terms,
Tags
Interpreted
meaning
NLP,
Enrich,
Auto-‐Tag
Unstructured
Content
Semi-‐structured
Index/
Graph
db
Review
74. @design4context
NextGen Search #UXPA2015
74
● Review
is
a
learning
process
● It
is
iteraCve
and
longitudinal,
so
help
a
person
understand
what
has
previously
been
retrieved,
viewed,
saved,
used
● Documents
are
part
of
the
search
experience
● The
take-‐away
is
the
informaCon,
not
the
document
/
desCnaCon
● Design
should
facilitate
the
next
acCons,
e.g.
synthesis,
sharing,
ciCng,
learning,
doing
● Be
humble
–
and
create
a
humble
experience
Design
consideraCons
Review
76. @design4context
NextGen Search #UXPA2015
Image
The
mechanics
of
search
Index
Query
Parser
Search
Engine
Document
Voice,
Context
Parameters
Models
Metadata,
Taxonomy,
Thesaurus,
Ontology
Results
&
Facets
List
of
items
List
of
terms
and
counts
Fragment
Results
Page
Structured
Data
Panels
Speech
Language
Images
ApplicaCon
Embedded
Results
Panels
FUTURE
Graph
Model
Images
&
MulCmedia
Profile
Text,
SemanCc
Processing
Save
User
Model
Terms,
Tags
Interpreted
meaning
NLP,
Enrich,
Auto-‐Tag
Unstructured
Content
Semi-‐structured
Index/
Graph
db
77. @design4context
NextGen Search #UXPA2015
User
System
• Think
• Look
at
what
you
did
before
• Gather
terms
and
ideas
• Express
keywords
• Pick
from
lists
• Enter
parameters
• Speak
sentences
• Change
what
you
expressed
(refine)
• Review
lists
• Read
small
amounts
of
informaCon
• Apply
filter
parameters
(refine)
• Make
selecCons
(short-‐term
keep)
• Look
at
what
you
found
• Make
decisions
about
what
else
to
seek
(refine)
• Select
and
discard
• Take
follow-‐on
acCon
• Share
• Keep
• Capture
terms
in
notes
• Extract
terms
from
documents
• Provide
reviews
of
previous
materials
• Provide
examples
of
what
might
be
available
(proacCve)
• Ingest
and
store
content
in
a
structured
way
so
it
can
be
searched
• Incorporate
what
other
people
have
searched
for
previously.
• Provide
example
phrases
(type
ahead)
• Parse
language
to
try
and
understand
intent
• Gather
context
signals
• Apply
thesaurus/
addiConal
model
terms
• Send
requests
to
other
services
• IdenCfy
appropriate
items
to
return
• Rank
the
returned
items
• Present
the
structured
data
• Present
the
returned
items
• Present
the
model
of
vocabulary
for
filtering
• Present
summaries
and
abstracCons
to
profile
what
is
found
• Analyze
selected
informaCon
• Provide
a
place
to
store
things
• Provide
a
place
to
organize
things
• Provide
a
way
to
manage
what
you
have
• Provide
a
way
to
take
what
you
have
and
use
it
elsewhere
• Profile
the
user’s
process
when
they
did
that
Seek
Type in the box
Evaluate
Assess results
Review
Assess content
Formulate
Strategy
Tactics