2. Crowdsourcing:
The
act
of
taking
a
job
tradi2onally
performed
by
a
designated
agent
(usually
an
employee)
and
outsourcing
it
to
an
undefined,
generally
large
group
of
people,
in
the
form
of
an
open
call
Jeff
Howe,
Crowdsourcing
3. The
logic:
Given
the
right
set
of
condi2ons,
the
crowd
has
the
poten2al
to
outperform
any
number
of
employees
4. More
and
more
companies
are
becoming
aware
of
this
logic,
and
are
aEemp2ng
to
exploit
it
by
tapping
into
the
excess
capacity
and
collec2ve
brainpower
of
the
crowd
–
usually,
for
liEle
or
no
compensa2on
5. Making
crowdsourcing
possible
• Emergence
of
open
source
soIware
movement
• Tools
of
produc2on:
widely
available,
faster,
cheaper,
easier
to
use
• Rise
of
online
communi2es
• Renaissance
of
amateurism/DIY
6. Why
does
the
crowd
do
it?
• Intrinsic
mo2va2ons
– A
belief
in
a
project
– Obliga2on
to
community
– Enjoyment
– Fulfillment
(crea2ve
or
other)
– Altruism
– Showing
off
(prove
how
smart/crea2ve
you
are)
– Reputa2on
enhancement
– OCD
• Extrinsic
mo2va2ons:
– Financial
reward
Non-‐comprehensive
list,
of
course.
Usually
several
mo2va2ons
will
coexist.
15. Eric
von
Hippel:
DemocraBzing
InnovaBon
Users
leading
companies
to
the
cubng-‐edge
• Scien2fic
instruments
• Computer
chips
• Sports:
windsurfing,
snowboarding,
mountain
biking
• Many
other
areas
16. The
collaboraBon
imperaBve
• The
current
R&D
model
is
“broken”
– In
some
cases,
R&D
expenses
rising
faster
than
sales
– e.g.,
10-‐15
years
and
$Bns
to
develop
a
new
drug
• With
the
escala2on
in
R&D
costs,
collabora2on
is
becoming
an
aErac2ve
economic
solu2on
• Businesses,
research
ins2tu2ons,
government
labs,
universi2es
are
moving
towards
collabora2on
17. If
I
can
tap
into
a
million
minds
simultaneously,
I
may
run
into
one
that’s
uniquely
prepared.
Alpheus
Bingham,
Eli
Lilly
[at
the
2me]
18.
19.
20.
21.
22.
23.
24.
25. Used
by
more
than
150
corpora2ons:
Eli
Lilly,
Boeing,
DuPont,
P&G,
Colgate-‐Palmolive…
27. InnoCenBve
and
The
value
of
diversity
• Harvard
research:
166
problems
from
26
different
companies
• The
odds
of
a
solver’s
success
were
higher
in
fields
in
which
they
had
NO
formal
exper2se
– The
farther
a
challenge
is
from
the
solver’s
specialty,
the
more
likely
it
is
to
be
solved
• 75%
of
solvers
already
knew
the
solu2on
to
the
problem
– The
problem
simply
needed
a
diverse
enough
set
of
minds
to
have
a
go
at
it
The
Value
of
Openness
in
Scien2fic
Problem
Solving,
Karim
R.
Lakhani,
Lars
Bo
Jeppesen,
Peter
A.
Lohse
and
Jill
A.
PaneEa,
Technology
and
Opera2ons
Management,
January
2007
29. Example:
Ed
Melcarek
• Problem
from
Colgate-‐Palmolive:
how
to
inject
fluoride
powder
into
a
toothpaste
tube
without
the
powder
dispersing
into
the
surrounding
air
• Solu2on:
impart
an
electric
charge
to
the
powder
while
grounding
the
tube
• An
electrical
solu2on
to
a
seemingly
chemical
problem
30. People
whose
networks
span
structural
holes
have
early
access
to
diverse,
oTen
contradictory,
informaBon
and
interpretaBons
which
gives
them
a
good
compeBBve
advantage
in
delivering
good
ideas...
This...
is
creaBvity
as
an
import-‐export
business.
An
idea
mundane
in
one
group
can
be
a
valuable
insight
in
another.
Ronald
Burt,
The
Social
Origin
of
Good
Ideas
38. In
2000,
we
decided
to
stop
being
Fortress
P&G,
and
move
to
an
open
innovaBon
system
that
could
aDract
innovaBons
of
all
stripes
from
the
outside.
Great
invenBon
is
going
on
anywhere
and
everywhere
in
the
world.
[We
have]
about
8,500
researchers,
and
we
figured
there
are
another
1.5M
similar
researchers
with
perBnent
areas
of
experBse.
Why
not
pick
their
brains?
A.G.
Lafley,
CEO,
P&G
39. When
I
became
CEO
of
P&G
in
2000,
we
were
introducing
new
brands
and
products
with
a
commercial
success
rate
of
15
to
20
percent…
Today,
our
success
rate
runs
between
50
and
60
percent.
That’s
as
high
as
we
want
[it]
to
be.
If
we
try
to
make
it
any
higher,
we’ll
be
tempted
to
err
on
the
side
of
cauBon.
Over
the
same
period,
we’ve
reduced
R&D
spending
as
a
percentage
of
sales;
it
was
about
4.5%
in
the
late
1990s
and
only
2.8%
in
2007.
[We]
focused
on
creaBng…
open
innovaBon:
taking
advantage
of
the
skills
and
interests
of
people
throughout
the
company
and
looking
for
partnerships
outside
P&G.
In
essence,
we
are
building
a
social
system
with
the
purchasers
(and
potenBal
purchasers)
of
our
products,
enabling
them
to
co-‐design
and
co-‐engineer
our
innovaBons
A.G.
Lafley,
P&G’s
Innova2on
culture,
Strategy
&
Business
magazine
40. This
was
important
to
us
for
several
reasons:
First,
we
needed
to
broaden
our
capabiliBes…
Second,
building
an
open
innovaBon
culture
was
criBcal
for
realizing
the
essenBal
growth
opportunity
presented
by
emerging
markets…
A
third
reason…
had
to
do
with
fostering
teams…
For
all
these
reasons,
we
consciously
set
in
place
a
series
of
measures
for
building
an
open
innovaBon
culture…
A.G.
Lafley,
P&G’s
Innova2on
culture,
Strategy
&
Business
magazine
49. The
Diversity
Trumps
Ability
Theorem
A
randomly
selected
collec2on
of
problem
solvers
will
outperform
a
collec2on
of
the
best
individual
solvers
50. Why
is
a
community
a
more
efficient?
• BeEer
at
iden2fying
talented
people
– The
community
doesn’t
need
to
find
the
person
most
suited
for
the
task,
because…
– The
person
with
the
right
combina2on
of
talent,
willingness
and
spare
2me
will
self-‐iden2fy
for
the
task
–
and
undertake
it
without
permission,
contract
or
instruc2on
– Transac2on
costs
=
zero
• BeEer
at
evalua2ng
output
– If
the
contributor
has
overes2mated
his
or
her
own
abili2es
–
the
community
will
iden2fy
that,
too
Clay
Shirky,
Here
comes
everybody
51. CondiBons
for
diversity
to
trump
ability
• Scale
of
diversity
=
a
large
enough
pool
to
guarantee
a
diverse
array
of
approaches
• Qualified
members
(“not
just
subway
passengers”)
• Method
of
aggrega2ng
and
processing
individual
contribu2ons
• A
real
problem
(=challenging)
Clay
Shirky,
Here
comes
everybody
82. • The
mining
firm
made
its
proprietary
data
about
a
mining
site
in
Ontario
public,
then
challenged
outsiders
to
advise
where
to
dig
next
• The
par2cipants
suggested
more
than
a
hundred
possible
sites
to
explore,
many
of
which
had
not
been
mined
by
Goldcorp
–
and
that
yielded
new
gold
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97. 3
types
of
collecBve
intelligence
• Problem-‐solving
networks
• Idea
jams
• Predic2on
markets
98. • Idea
genera2on
by
employees
• Set
up
in
1996
• Annual
seed
funding
budget:
$40M
• Employees
receive
$300k-‐$500k
for
proposals
that
turn
into
business
plans
101. • 2006:
The
biggest
ever
jam
• 150k
minds
in
104
countries
• Clients,
consultants,
employees,
families
• 4
subject
areas:
transporta2on,
health,
environment,
finance
&
commerce
• 46k
ideas;
$100M
invested
in
10
of
them
102.
103.
104.
105. The
world
is
my
lab
now.
John
Kelly,
Director
of
IBM
Research,
discussing
IBM’s
collaboratories
(=open
innova2on
laboratories)
106.
107. • The
Linux
pre-‐installment
idea
was
brought
up
on
the
day
of
the
launch
(February
16
2007)
• 30k
users
quickly
gave
it
a
thumbs
up
• In
May,
Dell
launched
3
such
models
108. By
July
2009:
~12k
ideas,
~85k
comments,
~675k
vo2ngs,
354
implementa2ons
175. 3
types
of
collecBve
intelligence
• Problem-‐solving
networks
• Idea
jams
• PredicBon
markets
176. [Nearly
every
individual]
has
some
advantage
over
all
others
because
he
possesses
unique
informaBon
of
which
beneficial
use
might
be
made.
Each
member
of
society
can
have
a
small
fracBon
of
the
knowledge
possessed
by
all,
and
each
is
therefore
ignorant
of
most
of
the
facts
on
which
the
working
of
society
rests…
One
of
the
ways
in
which
civilizaBon
helps
us
overcome
that
limitaBon…
is
by
conquering
ignorance,
not
by
the
acquisiBon
of
more
knowledge,
but
by
the
uBlizaBon
of
knowledge
which
is
and
which
remains
widely
dispersed
among
individuals…
CivilizaBon
rests
on
the
fact
that
we
all
benefit
from
knowledge
that
we
do
NOT
possess.
FA
Hayek,
1974
Nobel
Prize
in
Economics.
From
The
use
of
knowledge
in
society,
1945
177. Nobody
knows
everything.
But
everybody
may
know
something.
James
Surowiecki,
The
Wisdom
of
Crowds
178.
179.
180. • Correct
on
80%
of
Oscar
nomina2ons
• Never
missed
more
than
one
top
award
since
1996
launch
181.
182.
183.
184.
185. Drawbacks
of
internal
predicBon
markets
• Not
using
real
money
– Lower
credibility
– Skewed
incen2ves
• Thin
markets
– Not
enough
trades/traders
– Not
enough
diversity
190. Marketocracy
[ini2al
incarna2on]
The
Masters
100
mutual
fund
was
comprised
of
the
leading
100
porzolios
out
of
100,000
virtual
porzolios
managed
on
the
Marketocracy
website.
Masters
100’s
performance
consistently
“beat”
the
S&P500
.
197. The
free
soTware
movement
• Free
as
in
free
speech,
not
free
beer
(liberty,
not
price)
• Started
well
before
Linux
was
created,
but
it’s
probably
the
most
widely-‐known
example
198. Some
lessons
• Given
a
large
enough
beta-‐tester/co-‐developer
base,
almost
every
problem
will
be
fixed
quickly
– “Given
enough
eyeballs,
all
bugs
are
shallow”
(Linus's
Law)
• More
users
find
more
bugs
–
because
adding
more
users
adds
more
different
ways
of
stressing
the
program
• The
next
best
thing
to
having
good
ideas
is
recognizing
good
ideas
from
your
users.
Some2mes
the
laEer
is
beEer.
Eric
Steven
Raymond,
The
Cathedral
and
the
Bazaar
255. We
think
of
our
members
as
an
army
of
eyes
and
ears.
But
we’re
not
asking
them
to
be
journalists.
The
phrase
‘ciBzen
journalism’
makes
about
as
much
sense
as
‘ciBzen
denBst.
Leonard
Brody,
CEO,
NowPublic.com
256.
257.
258.
259.
260.
261.
262.
263.
264.
265.
266.
267. InvesBgaBve
reporBng
milestones
• Ft
Myers
News-‐Press:
contribu2ng
to
serious
journalis2c
inves2ga2ons
– 2006
housing
development,
bid-‐rigging
• Talking
Points
Memo
and
the
firing
of
state
aEorneys,
March
2007
– Awarded
the
George
Polk
Award
for
Legal
Repor2ng
for
“tenacious
inves2ga2ve
repor2ng”
– When
Dept.
of
Jus2ce
dumped
3,000
pages
of
documents
on
the
press,
site
members
divided
the
pile
into
50-‐page
slices
and
made
stunningly
quick
work
of
the
subject
318. • The
compe22on
generated
1Bn
impressions
(=
media
investment
of
$36M)
• During
and
aIer
the
compe22on,
the
ad
garnered
600M
views
• Cost
of
the
winning
ad:
$12.80
388. With
Lego
Factory
we
can
expand
beyond
our
100
in-‐house
designers
to
marvel
at
the
creaBvity
of
more
than
300,000
designers
worldwide.
Mark
Hansen,
Director,
Lego
Interac2ve
Experiences
389.
390.
391.
392. • Lego
provides
a
community-‐like
environment
where
users
can
share
their
Lego
experience
and
the
company
can
get
feedback
as
well
as
new
ideas
• In
Lego
Creator,
users
upload
their
own
crea2ons;
other
users
vote,
and
Lego
turns
the
most
popular
ones
into
real
products.
• The
company
brings
‘high-‐spenders’
to
the
more
advanced
Brickmaster
program
394. Dozens
of
hardware
inventors
around
the
world
have
begun
to
freely
publish
their
specs.
There
are
open
source
MP3
players,
VOIP
phone
routers,
mobile
phone,
laptop…
Clive
Thompson,
Wired
412. • Iden2fying
and
measuring
landforms
(craters,
ridges,
valleys)
–
in
order
to
find
evidence
of
water
• As
a
pilot,
an
already-‐categorized
dataset
was
put
online
–
88k
images
– Within
a
month
all
were
categorized
accurately
by
the
community
– Took
a
professional
geo-‐scien2st
two
years
413.
414.
415.
416.
417.
418.
419.
420. • Observa2ons
grown
10-‐fold
in
a
period
of
5
years
• “You
don’t
need
a
PhD
to
count
the
birds
in
your
backyard”
• Ornithology,
like
astronomy,
is
by
now
highly
dependent
on
amateurs
for
gathering
and
siIing
through
raw
data
428. Outsourcing
patent
review
• The
system
is
broken
– On
average,
2.5
years
between
filing
and
decision
– Backlog
of
1M
patent
applica2ons,
~470k
in
2007
alone
– 5,500
examiners
–
only
20
hours/applica2on
– Patent
parking
– Overlapping,
dubious
patents
(e.g.,
system
for
crea2ng
a
note
related
to
a
phone
call
–
MicrosoI)
• Solu2on:
open
the
review
process
to
public
comment
488. 4.
Crowd
filtering
Passively
and
ac2vely
filtering
the
exponen2ally-‐increasing
catalogue
of
the
Web
489. Mass
amateurizaBon
has
created
a
filtering
problem
vastly
larger
than
we
had
with
tradiBonal
media,
so
much
larger
that
many
of
the
old
soluBons
are
simply
broken.
The
brute
economic
logic
of
allowing
anyone
to
create
anything
and
make
it
available
to
anyone
creates
such a
staggering
volume
of
new
material
that
no
group
of
professionals
will
be
adequate
to
filter
[it].
Mass
amateurizaBon
of
publishing
makes
mass
amateurizaBon
of
filtering
a
forced
move.
Clay
Shirky,
Here
comes
everybody
490. The
acBvity
of
the
10%
[who
filter]…
is
as
valuable
to
any
online
community
as
the
acBons
of
the
[1%
of]
‘supercontributors’.
Bradley
Horowitz,
former
VP
of
Advanced
Development
Division,
Yahoo
517. If
an
appliance
manufacturer
finds
a
reviewer
on
buzzilions.com
saying
that
his
oven’s
door
melts
on
the
self-‐cleaning
cycle,
then
the
manufacturer
has
a
quality
problem,
not
a
review
problem.
Groundswell
518. Amazon
has
branded
itself
by
gathering
informa2on
from
consumers
–
and
then
returning
it
to
them
in
the
form
of
services
such
as
product
recommenda2ons,
sales
ranking
and
client
reviews.
519. Up
to
23
collabora2ve
features
on
any
Amazon
product
page
Picture
source
needed
529. Web
2.0
companies...
build
systems
that
get
beDer
the
more
people
use
them...
The
architecture…
is
such
that
users
pursuing
their
own
“selfish”
interests
build
collecBve
value
as
an
automaBc
byproduct.
An
architecture
of
parBcipaBon.
Tim
O’Reilly
530. The
crowd
produces
mostly
crap.
The
crowd
finds
the
best
stuff.
The
rise
of
Crowdsourcing,
Wired
Magazine
531. The
filtering
sequence
has
been
reversed
From
Filter-‐then-‐Publish
To
Publish-‐then-‐Filter
Clay
Shirky,
Here
Comes
Everybody
562. 5
types
(some
overlap)
1. Collec2ve
intelligence:
what
the
crowd
knows
– Solu2on
networks,
idea
jams,
predic2on
markets
2. Crowd
crea2on:
what
the
crowd
creates
– The
1%
rule
3. Crowdtasking:
what
the
crowd
does
4. Crowd
filtering:
what
the
crowd
thinks
– The
10%
rule
5. Crowdfunding:
what
the
crowd
finances
Jeff
Howe,
Crowdsourcing
[except
for
#3]
563. Crowdsourcing
rules
1. Pick
the
right
model
out
of
the
5
(or
a
combina2on)
2. Pick
the
right
crowd
(=target
audience)
3. Offer
the
right
incen2ves
– Personal
glory,
sense
of
community
–
or
even
cash
4. Ask
not
what
the
crowd
can
do
for
you,
but
what
you
can
do
for
the
crowd
– Crowdsourcing
works
best
when
the
individual/company
give
the
crowd
something
it
wants
– Create
a
horizontal
rela2onship
within
the
community.
It
could
be
more
important
than
a
ver2cal
rela2onship
between
the
company
and
the
individual
contributors.
They
want
to
talk
to
each
other
5. Keep
the
pink
slips
in
the
drawer
Jeff
Howe,
Crowdsourcing
564. Crowdsourcing
rules
[cont.]
6. Keep
it
simple
and
break
it
down
– Be
clear
on
what
you
want
your
contributors
to
do
– Get
the
division
of
labor
right
– “While
crea2ve
capacity
and
judgment
are
universally
distributed
in
a
popula2on,
available
2me
and
aEen2on
are
not”
(Yochai
Benkler)
– “Because
everyone
already
knew
what
an
encyclopedia
entry
was”
(Jimmy
Wales,
when
asked
why
Wikipedia
has
done
so
well)
7. The
dumbness
of
crowds,
or
the
benevolent
dictator
principle
– Have
someone
there
to
greet
them
when
they
show
up
– Someone
needs
to
guide,
and
some2mes
decide
8. Remember
Sturgeon’s
Law
9. Remember
the
10%
rule,
the
an2dote
to
Sturgeon’s
Law
10. The
community’s
[almost]
always
right
– Don’t
try
to
control
the
discussion
–
provide
the
plazorm
for
it
565. If
you’re
not
conducBng
an
exercise
like
that
in
your
organizaBon,
you
risk
missing
the
boat
on
a
sea
change
that’s
transforming
business.
You
must
overcome
natural
organizaBonal
resistance
to
the
idea
of
relinquishing
significant
control
to
people
outside
the
company.
Even
without
knowing
your
business,
I’d
be
willing
to
bet
that
contribuBon
systems
can
address
one
or
more
of
the
business
challenges
you
face
beDer
than
the
methods
you
currently
use.
Your
company
probably
has
advantages
that
start-‐ups
can
only
dream
of:
exisBng
customers,
traffic
to
your
website…
Naturally,
adopBng
those
methods
is
easier
when
compeBtors
have
beaten
you
to
the
punch...
But
what
if
you
want
to
lead
your
rivals?
ScoE
Cook,
Intuit
Founder
&
Chairman,
The
contribu2on
revolu2on,
Harvard
Business
Review