2. WHO AM I? – VP/Managing Director EMEA
DOMINIC TRIGG
• VP Global sales & Marketing, TradeDoubler
• Ad Operations Dir, Yahoo
• Ad Director, Microsoft MSN
• Advertising head BT Internet
• 6 years Press Advertising
18 YEARS INTERNET ADVERTISING
3. WHAT IS ROCKET FUEL?
TECHNOLOGY COMPANYARTIFICIAL INTELLIGENCEBIG DATADIGITAL ADVERTISINGTRUE IMPACT
5. Programma>c
buying
Bidding
on
individual
ad
impressions
In
real
>me
For
the
opportunity…
To
show
one
specific
ad
To
one
specific
consumer
In
one
specific
context
What
is
RTB
Programma>c
Buying?
5
key
Ques>ons:
6. Effec>veness
Buy
only
consumers
that
you
want
Only
in
contexts
that
generate
impact
And
scale
up
massively
Why
is
Programma>c
so
effec>ve?
5
key
Ques>ons:
7. Branding
Direct
Response
Loyalty
Marke>ng
• Reach
&
Frequency
• Brand
equity
liX
• Purchase
intent
• Prospec>ng
• Retarge>ng
• Offline
impact
• Up
selling
• Cross
selling
• Referrals
When
can
Programma>c
be
used?
5
key
Ques>ons:
8. A
pla+orm
that
enables
1:1
Marke9ng
@
Scale
Context
3rd
party
data
1st
party
data
How
do
I
reach
my
tailored
audience
with
RTB
Programma>c?
5
key
Ques>ons:
9. The
Evolu9on
of
Digital
Ad
buying
Big
sites
big
reach!
Where
does
UU
come
from
What
do
we
know?
What
is
the
Objec>ve?
Results
set
the
algorithm
and
they
must
adapt
AGE OF DELIVERY
AGE OF TARGETING
AGE OF OPTIMISATION
Ad
Effec>veness
10. HOUSEHOLD INCOMEAGE
Let’s look at Optimisation
Possible
Combina9ons
GENDER
(7
Buckets)
(8
Buckets)
x x
(2
Buckets)
112 Combinations=
11. CITIES
TRADITIONAL
OPTIMIZATION
There
are
85
ci9es
in
Sweden.
When
combined
with
our
other
metrics
and
available
channels,
that’s
38,080
possible
combina>ons.
(112
x
85
x
4)
38,080 Combinations=
(85
Buckets)
x112Combinations
CHANNELS
(4
Plagorms)
x
12. =
504,216,244,224,000,000,000,000,000
Segments
Data segments on an Exchange
an opportunity + a problem
ATribute
#
of
Segments
Age
18
Gender
2
HHI
16
Geo
43,000
Lifestyles
100
Interests
800
ATribute
#
of
Segments
Psychographics
42
Past
Purchases
990
Age
of
Children
17
Contextual
100,000
Time
of
Day
720
Ad
Size
5
=
145,710
Segments
A
Combina9onal
Explosion!
13. SOLVING THE KNOWLEDGE PARADOX
Data
Ability
to Make
Decisions
Ideal
Actual
Opportunity
14. INTRODUCING A.I. TO THE MIX
=
500k queries
per second
8.64 million Analysts
(5,000 decisions per day)
20. &
What
do
we
mean
by
BIG
DATA?
“From
the
dawn
of
civiliza4on
un4l
2003,
humankind
generated
five
exabytes
of
data.
Now
we
produce
five
exabytes
every
two
days…
and
the
pace
is
accelera4ng.”
-‐-‐
Eric
Schmidt,
Chairman,
Google
22. “The
prac4cal
conclusion
is
that
we
should
turn
many
of
our
decisions,
predic4ons,
diagnoses,
and
judgments—both
the
trivial
and
the
consequen4al—over
to
the
algorithms.
There’s
just
no
controversy
any
more
about
whether
doing
so
will
give
us
beKer
results.”
Andrew
McAfee
Principal
Research
Scien4st,
MIT
Sloan
December,
2013
Big
Data’s
Biggest
Challenge?
Convincing
People
NOT
to
Trust
Their
Judgment
23. Facebook likes per year 1 Trillion
Google searches per year 2.2 Trillion
Est. sand grains in West Texas desert 2.8 Trillion
Rocket Fuel consumer data points 3.5 Quadrillion
THE EXPLOSION OF CONSUMER DATA
24. THE
MARKETER’S
DILEMMA
“There
is
no
point
in
collec.ng
and
storing
all
this
data
if
the
algorithms
are
not
able
to
find
useful
pa7erns
and
insights
in
the
data….”
26. The
Future
In
addi>on
to
being
able
to
process
more
data
in
a
smaller
>me
frame,
AI-‐powered
solu>ons
can
quickly
iden>fy
which
data
points
are
significant
to
performance,
and
eliminate
the
ones
that
don’t
maker.
27. Making this stuff matter
àSUCCESS for Customers by combining
BIG DATA with ARTIFICIAL INTELLIGENCE
28. IN THE AGE OF
TARGETING…OPTIMISATION…
DEMOGRAPHIC A
BEHAVIOURAL
SEGMENT B
CONTENT
CATEGORY C
AI …
29. A DAY IN THE LIFE OF THE ADDRESSABLE CONSUMER
7:35 AM 9:20 AM 11:30 AM 12:05 PM
2:15 PM5:30 PM11:00 PM 8:00 PM
CONTEXT
IS CRITICAL
30. PURCHASE
INTENT
AWARENESS
FAVORABILITY
CONSIDERATION
CUSTOMERS
LOYALISTS
Full Funnel + Cross-Channel Campaign
What
makers
is
the
UU
and
their
rela>onship
to
the
campaign
31. AUTOMATED SELF-LEARNING
Age/Gender
Occupa>on
Income
Ethnicity
Purchase
Intent
Online
Purchases
Offline
Purchases
Browsing
Behavior
Site
Ac>ons
Zip
Code
City/DMA
Search
Sites
Search
Categories
Recency
Search
Keywords
Web
Site/Page
Referral
URL
Site
Category
Bizographics
Social
Interests
Lifestyle
ROCKET FUEL
x
+
-‐
-‐7
+17
X
-‐2
+8
+14
X
-‐9
-‐13
-‐12
X
+19
+13
X
+11
X
X
X
+25
+6
X
-‐7
+17
-‐2
+28
X
+11
X
X
-‐9
+14
+17
+19
+8
+11
X
X
+17
-‐23
+6
X
+17
-‐7
X
-‐2
-‐13
-‐12
X
+13
+6
X
X
X
-‐9
X
+17
X
+19
+8
+14
+18
-‐23
+17
-‐12
+11
-‐9
+8
+14
X
+11
-‐13
-‐12
+11
X
X
-‐7
+17
+8
+18
X
+11
X
-‐12
-‐10
+6
+14
X
+8
+11
-‐10
+13
+28
+6
+13
+19
X
+11
-‐10
+13
-‐12
+17
X
-‐7
+8
X
60
11MM+
Features
Posi9ve
Lii
Marginal
Lii
Nega9ve
Lii
+8
+13
+11
-‐9
+11
32. [
+
]
FLOW OF AVAILABLE
IMPRESSIONS ON EXCHANGES
IMPRESSION PROPENSITY SCORE
Likelihood to drive desired objective
-2.42
1.25
2.11
1.26
-2.78
1.256
-1.809
-2.42
1.25
2.11
-1.26
2.78
0.586
-2.009
1.25
2.11
-1.26
2.78
1.56
0.00
IDENTIFYING
MOMENTS
OF
INFLUENCE
+
Applying
learnings
at
the
impression
level
[
+
]
33.
34. INSIGHTS
INTERFACE
Giving access to campaign
insights in real-time, including:
» Personal login details
» Supporting multiple client campaigns
» Quick overview across campaigns
» All key metrics and trends at a glance
» Insights updated every 10 minutes
» Insights across 1000’s of data points
» Compare two metrics interactively
» Live calculation of top customers
35. Agency/Client
Then
adding
a
real
world
Support
Structure
Dedicated,
Named
Account
Manager
Analysts
Team
Opera9ons
Team
Account
Mgmt
Engineering
and
Research
Corp
Mgmt
Day
to
Day
Campaign
Management
Performance
Review
Escala>on
Support
Structure
Availability
24/7/365