ENTERPRISE
CROWDSOURCING
USE
CASES
ANALYSIS
ENTERPRISE
CROWDSOURCING
USE
CASES
ANALYSIS
Research
Report
November,
2011
Copyright
©
2011,
Crowdsourcing,
LLC
P a g e
|
0
ENTERPRISE
CROWDSOURCING
USE
CASES
ANALYSIS
SCOPE
This
report
addresses:
• • massolution’s
proprietary
methodology
for
the
categorization
of
the
Crowdsourcing
use
cases
Definition
of
existing
Crowdsourcing
use
cases
within
the
Enterprise
domain
and
examples
of
the
projects
and
tasks
delivered
through
the
use
of
Crowdsourcing
METHODOLOGY
Rapid
growth
of
the
Crowdsourcing
market
leads
to
continuous
emergence
of
new
use
cases.
In
order
to
create
the
most
complete
inventory
of
Crowdsourcing
use
cases,
the
massolution
team
conducted
an
extensive
review
of
the
publicly
available
information
on
Crowdsourcing
transactions,
announcements
and
industry
news.
In
addition,
we
interviewed
a
number
of
leading
Crowdsourcing
providers
and
analyzed
the
information
available
on
our
industry
website
crowdsourcing.org,
reviewing
more
than
2,500
content
items
(articles,
blogs,
documents)
published
over
the
last
two
years.
The
illustration
below
highlights
some
of
the
Crowdsourcing
providers,
interviewed
during
the
study.
On
crowdsourcing.org
alone
we
reviewed
over
2,500
references
to
the
use
of
crowdsourcing
(individual
content
items)
and
identified
of
these,
435
provided
relevant
examples
of
enterprise
crowdsourcing,
which
we
then
classified
using
massolution’s
proprietary
Use
Case
categorization
framework.
The
subsequent
analysis
of
these
content
items
produced
information
on
the
crowdsourcing
efforts
of
83
enterprises
from
13
industries
in
addition
to
multiple
government
clients.
In
additional
to
specific
references
to
enterprises,
a
large
number
of
items
also
described,
in
general
terms,
Copyright
©
2011,
Crowdsourcing,
LLC
P a g e
|
1
ENTERPRISE
CROWDSOURCING
USE
CASES
ANALYSIS
other
offerings
of
Crowdsourcing
service
providers,
targeted
at
the
enterprise
market,
that
were
not
specifically
attributed
to
named
enterprises.
The
chart
following
shows
how
specific
enterprise
examples
were
broken
down
by
industry.
Copyright
©
2011,
Crowdsourcing,
LLC
P a g e
|
2
ENTERPRISE
CROWDSOURCING
USE
CASES
ANALYSIS
ENTERPRISE
USE
CASES
In
total,
we
identified
65
specific
Use
Cases
(i.e.,
discreet
applications)
of
Crowdsourcing
being
performed
today
within
enterprises,
which
we
then
organized
under
four
categories:
1. Knowledge-‐based
tasks
are
completed
by
online
workers
that
possess
knowledge
based
on
a
familiarity
with
someone
or
something
that
can
include
philosophy,
facts,
descriptions,
information,
and/or
skills
acquired
through
experience
or
education.
It
can
refer
to
both
the
theoretical
or
practical
understanding
of
a
subject.
It
can
be
implicit
(as
with
practical
skill
or
expertise)
or
explicit
(as
with
the
theoretical
understanding
of
a
subject);
and
it
can
be
more
or
less
formal
or
systematic.
2. Expertise-‐based
tasks
are
completed
by
online
workers
that
are
widely
recognized
as
a
reliable
source
of
technique
or
skill,
and
whose
faculty
for
judging
or
deciding
rightly,
justly,
or
wisely
is
accorded
authority
and
status
by
their
peers
or
the
public
in
a
specific
well-‐distinguished
domain.
3. Micro-‐tasks
are
short
duration
tasks
completed
by
online
workers.
Micro-‐tasks
require
no
specialized
knowledge
or
expertise
other
than
a
basic
level
of
comprehension
and
cognitive
judgment.
4. Ideation-‐process
tasks
are
completed
by
online
workers
engaged
in
the
creative
process
of
generating,
developing,
and
communicating
new
ideas,
where
an
idea
is
understood
as
a
basic
element
of
thought
that
can
be
visual,
concrete,
or
abstract.
Ideation
is
all
stages
of
a
thought
cycle,
from
innovation,
to
development,
to
actualization.
We
identified
a
large
number
of
Use
Cases
(e.g.,
content
generation,
image
validation)
that
are
process
agnostic
and
would
apply
to
multiple
vertical
and
horizontal
processes
within
the
enterprise.
A
smaller
number
of
use
cases
are
aligned
along
specific
enterprise
process
groups
(e.g.,
customer
support,
underwriting).
Copyright
©
2011,
Crowdsourcing,
LLC
P a g e
|
3
ENTERPRISE
CROWDSOURCING
USE
CASES
ANALYSIS
1. Expertise-‐Based
Tasks
We
identified
seven
types
of
enterprise
use
cases
that
involve
the
provisioning
of
specific
expertise.
The
table
below
provides
additional
details
on
these
use
cases.
Use
case
type
Translation
Use
cases
High
volume
translation.
Offline
translation
of
large
amounts
of
text
(e.g.,
product
documentation)
or
a
large
number
of
small
textual
elements
(e.g.,
product
descriptions).
Real
time
translation.
Automated
real
time
translation
or
improvement
of
machine
translation
of
online
content
such
as
Web
pages,
documents,
customer
support,
user
generated
content.
Example
Global
Internet
coupon
giant
Groupon,
crowdsourced
large-‐ scale,
global
translation
and
optimization
project
in
a
number
of
languages.
Translation
Cloud
allows
freelance
translators
the
ability
to
translate
ongoing
projects
in
real
time
using
their
Facebook
accounts.
Copyright
©
2011,
Crowdsourcing,
LLC
P a g e
|
4
ENTERPRISE
CROWDSOURCING
USE
CASES
ANALYSIS
Monitoring
Sentiment
analysis.
Review
user
generated
content
(UGC),
e.g.,
messages,
tweets,
blogs,
reviews
to
identify
and
judge
sentiment
(i.e.,
favorable/
neutral/unfavorable)
UGC
monitoring.
Regular
review
of
social
media
platforms,
customer
product
reviews,
etc.
in
search
of
specific
alerts
or
inappropriate
content.
Regulatory
monitoring.
Provide
regulatory
environment
oversight
in
multiple
localities,
using
local
experts.
Video
monitoring.
Provide
monitoring
of
real
time
video
content
(e.g.,
for
enforcement
purposes)
Information
policing.
Identify
and
report
specific
behavior
or
informational
signals.
Website
demographics.
Measure
the
demographics
of
the
audience
by
attribute
(e.g.,
likes/dislikes).
Pricing
analysis.
Use
a
distributed
expert-‐base
community
to
provide
advice
on
prices
and
complex
pricing
models.
Patent
search.
Use
qualified
distributed
patent
researchers
to
support
the
patent
process
for
new
applications
or
defending
existing
patents.
Business
consulting.
Use
a
global
expert
pool
to
conduct
distributed
consulting
engagements.
General
research.
Conduct
topical
analyses
in
a
wide
variety
of
areas
(e.g.,
artist/album
and
song
Mobile
operator
used
Crowdsourcing
to
understand
increase
in
churn
rate.
Their
automated
system
could
not
process
advanced
UGC,
such
as
sarcasms
or
complex
emotions.
US
software
firm
used
Crowdsourcing
to
monitor
regulatory
changes
in
multiple
countries
to
update
its
software
product.
Internet
Eyes
service
allows
volunteers
in
Britain
to
watch
live
security
camera
video
from
stores
to
catch
thieves
in
the
act.
Reform
Reading
Consumer
Agency
used
crowdsourcing
to
police
the
industry
for
deceptive
financial
practices.
Business
analytics
PriceProphets
recruited
pricing
analysts,
pricing
architects
and
other
professionals
to
provide
recommendation
on
prices.
Microsoft
became
the
first
member
of
Article
One’s
Litigation
Avoidance
program,
which
uses
Crowdsourcing
to
find
evidence
of
“prior
art.”
Insurance
provider
used
10EQS
to
map
IT
organizational
structure
for
an
industry,
using
knowledge
of
insurance
experts.
Copyright
©
2011,
Crowdsourcing,
LLC
P a g e
|
5
ENTERPRISE
CROWDSOURCING
USE
CASES
ANALYSIS
Customer
support
research).
Social
CRM.
Enable
better
customer
voice
and
problem
understanding
through
community
engagement.
Peer
support.
Use
existing
customers
to
provide
customer
support
to
peers.
Support
knowledge
base
generation.
Use
company
employees
to
generate
support
content
for
internal
reuse.
Sales
and
marketing
Target
identification.
Create
validated
contact
lists
to
enable
marketing
campaigns.
Lead
generation.
Crowdsourced
lead
generation,
based
on
community
input.
Application
services
Channel
marketing.
Leverage
reach
and
reputation
of
a
large
number
of
qualified
experts
to
sell
products.
Customer
upsell.
Identify
brand
champions
and
reward
them
for
supporting
customer
acquisition.
Code
generation.
Produce
software
code
via
a
distributed
base
of
software
developers,
using
collaboration
or
competition
approach.
Code
testing.
Test
software
code,
based
on
pre-‐determined
use
cases,
with
the
help
of
a
distributed
base
of
testers.
Product
testing.
Test
company
products
by
engaging
an
external
community.
Oracle
integrated
an
application
that
allows
customer-‐generated
ideas
to
automatically
link
with
their
CRM
profiles
helping
sales
reps
and
marketing
managers
to
understand
each
user's
point-‐of-‐ view,
ideas
and
challenges.
Microsoft
created
Most
Valuable
Professional
program
to
crowdsource
technical
support
for
Microsoft
technology.
Government
agency
developed
an
internal
community
to
generate
knowledge
around
problem
solving
for
tier-‐II
support
requests.
IT
hardware
company
developed
list
of
IT
contacts
within
Colleges
and
Local
Government
Agencies
LeadTip.com
launched
a
“sales-‐ ready”
marketplace
for
sharing
tips
on
customers
ready
to
buy.
RFPdb
encourages
participants
to
upload
RFPs
that
they
discovered
in
their
search
in
exchange
for
credits.
Global
security
software
maker
AVG
used
CrossLoop's
Channel
Program
to
sell
services
through
experts.
Large
Italian
Telco
provides
discounted/free
services
as
a
commission
for
top
customers
helping
to
recruit
other
customers
NASA
and
Harvard
University
jointly
designed
a
program
to
facilitate
a
crowdsourced
competition
for
software
development.
AOL
Answers
leveraged
a
crowdsourced
testing
solution
from
uTest
for
functional
and
exploratory
testing.
Xerox
launched
a
website
allowing
customers
to
test
drive
its
technology
research
projects,
Copyright
©
2011,
Crowdsourcing,
LLC
P a g e
|
6
ENTERPRISE
CROWDSOURCING
USE
CASES
ANALYSIS
Industry
specific
including
PDF
converters
and
an
Arabic
language
analyzer.
Underwriting.
Follow
the
pre-‐ US
Financial
institution
used
US
defined
set
of
rules
to
generate
rural
workers
to
support
specific
required
decision
making
mortgage
processing
tasks
information.
Clinical
trials.
Use
of
patients
to
PatientsLikeMe
crowdsourced
assess
efficacy
of
a
drug
or
a
clinical
trials,
obtaining
data
on
treatment.
the
effectiveness
of
off-‐label
use
of
drugs.
Recruiting
referrals.
Engage
IBM
used
Crowdsourcing
to
reach
external
communities
to
millions
of
potentially
employable
generate
referrals
for
recruiting
individuals
in
rural
areas
and
in
purposes.
the
unorganized
sectors.
Other.
Completion
of
other
industry
specific
micro-‐tasks.
2.
Knowledge-‐Based
Tasks
Delivering
content
through
Crowdsourcing
is
the
largest
area
of
Crowdsourcing
work
to
date.
Use
cases
in
this
area
fall
into
one
of
five
types:
The
table
below
provides
additional
details
on
these
use
cases.
Use
case
type
Content
generation
Use
cases
Article
production.
Produce
authored
materials
on
specific
topics
(e.g.,
tailored
mini-‐articles).
Example
Mechanical
Engineering
magazine
announced
a
12-‐ month
initiative
to
generate
content
for
the
magazine
Copyright
©
2011,
Crowdsourcing,
LLC
P a g e
|
7
ENTERPRISE
CROWDSOURCING
USE
CASES
ANALYSIS
from
the
readers
themselves.
Ether
Books
(publisher
launched
a
Short
Story
Competition
of
the
Week
contest
Outdoor
shoes
and
apparel
company
Merrell
generated
photo
content
for
the
largest
panoramic
image
using
crowdsourced
photos.
Small
content
production.
Write
short
paragraphs
of
specific
content
(e.g.,
caption
for
a
photo,
Adword
campaign).
Audio/video/image
content
production.
Generate
audio,
video
or
image
content
on
specific
topics.
Content
summary
production.
Write
summary
descriptions
of
articles,
images
or
other
objects.
Content
editing.
Rewriting,
line
editing,
abstracting,
display-‐type
writing.
Content
discovery.
Identify
specific
relevant
content
items
(e.g.,
find
competitor
promotions).
Citizen
Journalism.
Use
a
community
to
assemble
news
from
multiple
locations.
Content
moderation.
Identify
or
filter
out
inappropriate,
offensive
or
illegal
content
(e.g.,
SPAM,
false
or
imposter
profiles).
Photo
moderation.
Verify
photographic
content
for
compliance
with
specific
rules.
Content
evaluation/verification.
Evaluate
the
quality
of
content
in
articles.
Content
enhancement.
Enhance
accuracy
of
the
existing
classifications
(e.g.,
verify
product
attributes),
including
annotations
and
descriptions.
Fact
check.
Verify
numbers,
sources
or
other
reference
data.
Location-‐based
content.
Create
or
adopt
location-‐centric
content
(e.g.,
use
local
idioms,
terms
and
definitions).
Location-‐based
information.
Deploy
local
resources
to
gather
or
Content
management
Google
offered
an
editing
tool
to
its
users
allowing
them
to
make
revisions
to
US
maps.
eBay
used
Crowdsourcing
to
conduct
large-‐scale
information
collection
and
annotation
tasks.
TapIn
launched
a
mobile
social
network
for
news
collection.
Facebook
partnered
with
Web
of
Trust
to
crowdsource
protection
against
dubious
web
links.
Google
crowdsourced
review
of
photos
to
enforce
family
friendly
content.
Guardian
of
London
enlisted
large
number
of
volunteers
to
evaluate
information
in
large
number
of
public
records.
Real
estate
company
Crowdsourced
addition
of
images
to
their
property
catalog
Travelati
travel
magazine,
announced
contest
for
the
best
local
travel
tips
and
unique
stories.
Real
Estate
company
crowdsourced
local
property
Location
services
Copyright
©
2011,
Crowdsourcing,
LLC
P a g e
|
8
ENTERPRISE
CROWDSOURCING
USE
CASES
ANALYSIS
verify
specific
information
(e.g.,
find
local
products,
verify
and
name
local
points
of
interest).
Consumer
surveys.
Use
a
broad
range
of
participants
to
collect
data,
source
opinions
or
conduct
measurements.
Opinion
polling.
Assemble
input
from
a
large
number
of
people
(e.g.,
customers)
to
inform
business
decision
making.
Internet
search
Search
relevancy.
Verify
search
relevancy
and
improve
accuracy
of
the
search
results.
Develop
click-‐ through
analysis
for
search
improvement
recommendations.
Search
Engine
Optimization.
Improve
the
use
of
keywords
by
finding
the
most
optimal
keyword
combinations
and
phrases.
listing
verification.
Bank
repo
company
Crowdsourced
local
data
analysis.
PBS
NewsHour’s
crowdsourced
keeping
records
of
the
total
time
taken
by
travelers
at
airport
security
counters.
Perry
Ellis
polled
its
fans
for
their
favorite
male
model
to
walk
in
the
Perry
Ellis
fashion
show
2011
via
their
Facebook
page.
Online
retailer
improved
an
accuracy
of
site's
search
algorithms,
making
products
easier
to
find.
Trada,
crowdsourced
paid
search
marketplace,
has
created
an
offering
to
make
it
easy
for
attorneys
to
run
paid
search
campaigns
on
Google,
Yahoo
and
Bing.
Online
portal
used
Crowdsourcing
to
discover
whether
people
were
gaming
the
moderation
system.
Survey
production
Algorithm
training.
Verify
automated
solutions
tо
discover
holes
and
exploits
for
computer
learning.
Copyright
©
2011,
Crowdsourcing,
LLC
P a g e
|
9
ENTERPRISE
CROWDSOURCING
USE
CASES
ANALYSIS
3. Micro-‐Tasks
Micro-‐tasking
is
the
fastest-‐growing
category
of
Crowdsourcing
work.
As
platforms
become
more
sophisticated,
some
of
the
Expertise-‐based
tasks
can
now
be
performed
as
Micro-‐tasks.
Use
cases
in
this
area
fall
into
one
of
four
types:
The
table
below
provides
additional
details
on
these
use
cases.
Use
case
type
Transcription
Use
cases
Example
Audio
transcription.
Transcription
The
Khan
Academy
of
audio
data,
following
pre-‐ introduced
community-‐ defined
rules
for
content
access.
provided
subtitles.
Video
transcription
and
subtitles.
Castingwords
offers
Transcription
of
video
materials
crowdsourced
video/audio
following
pre-‐defined
rules
for
transcription
leveraging
content
access.
MTurk’s
worker
pool.
Data
categorization.
Apply
tags
to
Services
company
processed
data
(e.g.,
articles)
to
create
internal
knowledge
repository
searchable
records
for
data
to
create
metadata
and
mining,
cataloging
or
knowledge
enable
improve
search.
management.
Image
categorization.
Tagging
of
Google
developed
the
Image
the
images
(e.g.,
digital
receipts,
Labeler
game,
designed
to
invoices)
for
subsequent
improve
its
image
search
processing
and
cataloging.
results
through
tagging.
Audio
and
video
categorization.
Security
company
Categorization
Copyright
©
2011,
Crowdsourcing,
LLC
P a g e
|
10
ENTERPRISE
CROWDSOURCING
USE
CASES
ANALYSIS
Tagging
audio
or
video
files
for
archiving
purposes.
Metadata.
Add
metadata
to
objects
or
cleanup
(e.g.,
remove
duplicates)
existing
metadata.
crowdsourced
review
of
security
camera
surveillance
for
archiving
purposes.
The
Getty
Research
Institute
crowdsourced
reliable
metadata
for
the
Built
Works
Registry
(architecture)
project.
National
Library
of
Finland
launched
a
project
to
crowdsource
volunteer
data
entry
to
index
its
complex
archives
(e.g.,
handwritten).
Data
management
OCR
Validation
Data
entry.
Enter
data
into
systems
and
databases
from
various
sources
(e.g.,
scanned
image,
audio
file,
unstructured
text).
Data
normalization
(or
standardization).
Edit
large
number
of
records
to
ensure
data
consistency
or
uniform
formats
(e.g.,
common
phone
number
format).
Data
de-‐duplication.
Remove
repetitive
records
from
a
database
(e.g.,
duplicate
products).
Data
verification.
Validate
and
refresh
existing
data
(e.g.,
business,
organization,
or
consumer
listings)
using
online
sources.
Data
mining.
Extract
data
from
various
databases
and
other
sources
to
gather,
aggregate
or
append
information.
OCR
validation.
Validate
output
of
OCR
systems
for
hard-‐to-‐process
materials
(e.g.,
handwriting,
charts,
diagrams).
Technology
company
used
Crowdsourcing
to
verify
stale
business
listings
bought
from
third
party
data
providers.
Facebook
helped
scientists
quickly
identify
nearly
5,000
fish
specimens
collected
in
Guyana.
National
Library
of
Australia
conducted
Newspapers
Digitization
Program,
using
Crowdsourcing
to
correct
OCR
errors
Copyright
©
2011,
Crowdsourcing,
LLC
P a g e
|
11
ENTERPRISE
CROWDSOURCING
USE
CASES
ANALYSIS
4.
Ideation-‐process
Tasks
Ideation
is
longest
established
area
of
Crowdsourcing.
Emergence
of
online
platforms
that
helped
solicit
ideas
from
large
groups
of
people
can
be
tracked
over
the
last
15
years.
In
the
last
couple
of
years,
however,
this
area
has
flourished
and
now
has
expanded
to
include
multiple
elements
of
R&D.
The
diagram
below
shows
use
case
types
most
common
in
the
Ideation-‐process
tasks
category.
The
table
below
provides
additional
details
on
these
use
cases.
Use
case
type
Idea
generation
Use
cases
Brainstorming.
Generate
ideas
through
community
engagement
(e.g.,
employees,
suppliers).
Idea
evaluation.
Create
a
community
of
experts
to
evaluate
ideas.
Marketing
design.
Develop
marketing
materials
and
organize
programs
through
community
engagement.
Example
GE
announced
the
Ecoimagination
Challenge
to
generate
ideas
on
new
smart
power
solutions.
HP
developed
an
internal
hub
for
employee
idea
evaluation.
Unilever
crowdsourced
short
commercial
films
to
design
advertisement
for
13
brands.
PepsiCo
sponsored
a
crowdsourcing
competition
to
produce
a
commercial
for
Super
Bowl.
Glamour
magazine
announced
international
logo
design
contest
for
PAD
magazine.
Huffington
Post
held
a
design
contest
for
a
new
HuffPost
Design
and
optimization
Creative
design.
Design
logos,
brands,
concepts
and
other
elements
of
creative
design.
Copyright
©
2011,
Crowdsourcing,
LLC
P a g e
|
12
ENTERPRISE
CROWDSOURCING
USE
CASES
ANALYSIS
Product
design.
Assemble
community
of
clients
(or
potential
clients)
to
develop
key
characteristics
of
new
products.
Politics
logo
Simple
Shoes
launched
a
community-‐driven
design
competition
to
develop
new
product
(shoes).
The
US
Army
crowdsourced
a
combat
support
vehicle
design.
Derek
Lam
partnered
with
eBay
to
decide
which
collection
items
should
go
into
production
and
sales
next
spring.
Research
and
Development
Product
optimization.
Optimize
(or
rationalize)
an
existing
product
line
using
Crowdsourced
information
for
decision-‐making.
Scientific
research.
Use
contests
Canadian
hi-‐tech
company
or
online
collaboration
models
offered
$45,000
for
the
best
to
solve
scientific
problems.
algorithm
for
identification
of
substances
from
electromagnetic
signatures.
Business
research.
Use
contests
Netflix
announced
$1M
contest
or
online
collaboration
models
for
the
best
algorithm
to
to
solve
business
problems.
recommend
movies
to
users.
Citizen
science.
Use
McGill
University
used
game,
communities
or
games
to
Phylo,
to
take
advantage
of
engage
people
in
solving
humans’
prowess
at
pattern
scientific
problems.
recognition
to
solve
a
DNA
analysis
challenge.
Copyright
©
2011,
Crowdsourcing,
LLC
P a g e
|
13