1.
C o p y r i g h t -‐ H i m a n s h u
B a r i .
A n y
d u p l i c a t i o n
o r
d i s t r i b u t i o n
w i t h o u t
t h e
a u t h o r ’ s
p e r m i s s i o n
i s
s t r i c t l y
p r o h i b i t e d
Display
Advertising
-‐
An
Intersection
of
Science
&
Art
Author:
Himanshu
Bari
An
outsider’s
view
of
the
disruption
in
the
display
advertising
space
April
14
2. Disruption
in
the
Display
AdTech
Business
Per
a
major
publication
-‐
“The
space
between
advertisers
and
publishers,
where
there
used
to
be
a
handshake
and
a
media
plan,
is
now
densely
populated
by
various,
and
sometimes
mysterious,
products
and
services.
There
are
sell
side
platforms
and
buy
side
platforms.
There
are
ad
networks
and
ad
exchanges.
There
are
aggregators
and
optimizers.
There
are
services
so
cutting-‐edge
advertisers
aren’t
even
sure
what
they
do,
let
alone
whether
they
need
them”
What
was
the
problem
-‐ Limited
targeting
&
personalization
o Due
to
limitations
in
technology
&
data
availability,
targeting
was
done
at
very
broad
group
level
resulting
in
poor
engagement
-‐ Opacity
in
measurement
o Little
to
no
ability
to
measure
Ad
campaign
impact
to
key
business
metrics.
Also,
use
of
CPM
based
pricing
compounded
the
problem
and
resulted
in
misalignment
of
incentives
between
the
publisher
&
advertiser
-‐ Inefficient
execution
o As
seen
from
Appendix-‐1,
the
traditional
display
media
ordering
processes
is
extremely
manual
and
stitched
together
with
point
solutions
resulting
into
a
bloated
cost
structures
on
both
advertiser
and
publisher
ends
o Execution
has
traditionally
been
focused
and
customized
around
campaigns
and
there
was
little
room
to
exploit
economies
of
scale
What
has
changed
now?
-‐ Commoditization
of
technology
o Open
source
software
–
Not
having
to
spend
on
software
acquisition
really
gave
a
jump
start
to
innovation
in
the
AdTech
space
o Scale
out
computing
–
To
solve
the
problems
in
Display
advertising,
small
startups
needed
‘Google’
like
computing
paradigms
sophistication.
The
rise
of
open
source
Hadoop
&
NoSQL
technologies
gave
them
level
footing
o Cloud
–
95%
of
the
startups
in
the
AdTech
space
would
not
have
existed
without
the
availability
of
a
cheap
on
demand
infrastructure
-‐ Availability
of
‘Complete’
data
Driven
largely
by
the
commoditization
of
the
technology
used
in
storing
and
processing
data,
publishers
and
advertisers
have
access
to
data
that
is
o Granular
o Multi-‐dimensional
o Fresh
-‐ SaaS
–
Lowers
the
barrier
for
customers
to
try
new
products.
3.
Display
Lumascape…
Woow!
Opportunities
&
challenges
for
some
sections
of
the
Lumascape
PUBLISHERS
Key
Players
Too
many
to
list.
Opportunities
-‐ More
data
=
opportunities
for
price
discrimination
-‐ Better
liquidity
for
their
inventory
-‐ Monetize
first
party
data
-‐ Rich
media
–
engage
the
users
with
the
Ads..
eg.
Youtube
implemented
Ad
skip
functionality
-‐ Taking
the
one
throat
to
choke
approach
to
AdTech
4. -‐ Native
Ad
formats
(publisher
specific
ad
format
that
enables
creative
engagement
with
user)
-‐ Vertical
expansion
(starting
with
sell
side
platforms)
Challenges
-‐ Excess
inventory
&
programmatic
buying
are
putting
downward
pressure
on
CPMs
-‐ Fragmentation
of
the
Adtech
space
means
-‐
managing
multiple
ad
tech
vendor
relationships
becomes
a
big
overhead..
-‐ Too
many
dependencies
on
the
site
=
slow
down
and
brittleness
of
the
site
=
lost
revenue
-‐ Need
more
transparency
from
AdTech
vendors
-‐
most
Adtech
vendors
only
report
line
items
by
audience
segment
-‐ Lack
of
differentiation
amongst
Adtech
vendors
=
vendors
selling
dreams.
Publishers
want
to
see
case
studies
not
tech
promises.
Few
vendors
focused
on
content,
creatives,
verification,
analytics
and
audience
measurement
have
really
worked
hard
to
understand
publishers
needs
AGENCIES
Key
Players
WPP,
Omnicom,
Havas,
Publicis
etc
(see
Lumascape
for
better
list)
Opportunities
-‐ Capitalize
on
existing
brand
relationships.
Shoot
for
the
broader
campaign
&
brand
goals…
-‐ Making
up
that
lost
margin
with
more
modern
data
services
that
require
fewer
but
more
expensive
full-‐time
employees
and
global
data
services
through
partnerships
with
companies
like
Adobe.
The
pitch
goes
something
like
this:
While
it
might
be
cheaper
for
clients
to
deal
directly
with
ad-‐tech
vendors,
their
savings
will
be
eaten
up
by
the
time
and
resources
needed
to
cobble
together
a
global
ad-‐tech
ecosystem
on
their
own.
-‐ Wave
the
white
flag
with
DSPs
-‐
Agencies
often
lay
out
rules
of
engagement
for
ad-‐tech
partners
and
map
out
how
the
two
sides
can
team
up
to
win
new
clients.
One
former
agency
trading-‐desk
executive
said
his
firm
and
others
would
offer
quid
pro
quos
to
DSPs.
If
a
DSP
wanted
to
be
the
trading
desk's
primary
automated
ad-‐buying
tool
or
wanted
the
agency
to
guarantee
to
funnel
$1
million
a
month
in
ad
spending
through
its
system,
the
trading
desks
would
ask
the
DSP
for
introductions
to
new
clients
or
to
give
the
agency
first-‐look
at
any
inbound
client
inquiries
5. Challenges
-‐ DSP
vendors
going
direct
to
brands
-‐-‐-‐
Creating
an
interesting
dynamic
where
agencies
are
trying
to
build
DSP
like
capabilities
and
DSP
vendors
are
trying
to
build
service
functions
with
agency
like
capabilities
-‐ The
operations
teams
in
the
agencies
that
do
repetitive
tasks
are
the
most
under
threat.
The
creative
folk
that
strategize
and
do
campaign
planning
are
not
and
are
exactly
the
skills
that
the
DSP
companies
need.
-‐ Brand
Clients
can
cheaply
implement
and
oversee
many
ad-‐tech
functions
themselves,
endangering
the
fees
they
paid
to
full-‐time
agency
employees
for
overseeing
those
functions.
Razorfish
has
cut
fees
it
charges
clients
for
ad-‐
tech
services
such
as
dashboard
software
and
cloud-‐based
hosting
services
10%
to
20%
-‐ Brands
dictating
which
DSPs
they
want
to
work
with
and
agencies
have
to
go
with
that.
DMP
Key
Players
-‐
Oracle(Bluekai),
Adobe(AudienceManager),
Knotice,
nPario,
X+1,
Lotame
Current
focus
-‐ Ingest
&
normalize
data
from
search,
display,
email,
CRM,
site…etc.
-‐ Insight
generation
–
analyze
data
and
create
‘custom
audiences’
for
targeting
-‐ Deliver
the
audiences
to
DSPs
and
perhaps
content
management
systems
Opportunities
-‐ Mobile
tracking
&
targeting
–
Cookies
don’t
work
on
mobile
means
blindness
to
everything
that
is
not
browser
based.
-‐ Users
need
customer
targeting
and
measurement
across
the
lifecycle
from
acquisition
to
retention.
To
become
true
‘Customer
data
platforms’
(CDPs),
DMPs
they
need
to
be
o Integrated
with
marketing
automation
systems.
o Integrated
with
customer’s
internal
CRM
systems
-‐ Every
DMP
vendor
seems
to
have
different
strengths
along
the
core
feature
areas.
Given
the
immaturity
of
current
adoption
and
the
range
of
future
use
cases,
we
will
inevitably
see
more
use
case
&/or
vertical
specialization
amongst
DMP
vendors
Challenges
-‐ DSP
vendors
looking
to
acquire
DMP
like
capabilities
-‐ Organizational
resistance
in
rolling
out
a
DMP
–
lack
of
clear
ownership
&
need
for
rejiggering
of
marketing
ops
teams
6.
DSP
&
RE-‐TARGETING
Key
Players:
Turn
&
MediaMath
are
big
rivals.
TellApart
(CPA
based
pricing)
&
Criteo
(
CPC
based
pricing),
Adroll,
Triggit
are
retargeting
focused.,
Opportunities
-‐ Growing
popularity
of
CPA
based
pricing
–
Leads
to
better
cost
transparency
and
alignment
of
incentives.
-‐ Expand
into
the
direct
buys
–
eg.
AppNexus
just
announced
it
Tixt
platform
that
automates
the
RFP
creation
process.
-‐ Capturing
Broader
Advertising
Budgets.
Eg
–
from
Criteo
S1
-‐
To
date,
a
majority
of
our
revenue
has
been
derived
from
delivering
advertisements
to
users
who
have
expressed
an
intent
in
one
of
our
clients’
products
or
services,
with
the
objective
of
driving
a
sale
based
on
that
intent.
We
are
beginning
to
leverage
the
Criteo
Engine,
data
assets
and
proprietary
knowledge
to
help
businesses
achieve
longer
term
business
objectives,
such
as
customer
retention,
brand
awareness
and
preference
shift,
in
order
to
drive
sustained
sales
growth
over
time.
-‐ Focus
on
Mobile
-‐
Opportunity
to
significantly
expand
inventory
and
reach
as
well
as
address
the
growing
user
audience
and
content
consumption
on
mobile
devices
-‐ Specialization
by
vertical
&
re-‐targeting
type
(Search
retargeting,
site
retargeting,
CRM
retargeting)
–
eg.
Sojern
targets
travel
-‐ Consolidation
between
traditional
DSPs
and
the
new
breed
of
re-‐targeting
vendors
-‐ Expansion
in
BRICS
countries
by
US
&
EMEA
DSPs
Challenge
-‐ Media
reach
becoming
a
commodity
as
almost
all
DSPs
offer
connectivity
to
all
the
major
exchanges
-‐ Lack
of
unified
view
-‐
While
your
DSP
does
mid
funnel
prospecting,
your
personalized
retargeting
vendor
mops
up
the
conversions.
Try
this:
ask
your
personalized
retargeting
vendor
where
these
conversions
came
from.
Ask
them
what
the
effect
of
different
prospecting
campaigns
has
on
retargeting
campaigns
-‐ Inventory
acquisition
risk
–
Overdependence
on
some
exchanges.
Eg.
Criteo
get
30%
of
its
inventory
from
Google
&
AppNexus
-‐ DMPs
starting
to
offer
DSP
capabilities
7. BRANDS
-‐ Need
transparency
in
brand
measurements
metrics.
Google
eg.
Is
stressing
a
lot
on
providing
more
brand
metrics
around
the
Ads
that
are
shown
through
its
network
-‐ Not
just
Audience
BUT
Context
&
Content
is
becoming
key
-‐ Need
to
aggregate
tooling
around
the
following
cycle
o TARGET
(Granular
audience
across
devices
in
the
right
context
with
the
right
content)==>
MEASURE
(reach,
relevance,
conversion
etc)
==>
ANALYZE
(how
the
targeting
improved
the
measures)
==>
OPTIMIZE
(
learn
from
previous
campaigns)
==>
REPEAT…
-‐ Starting
to
build
their
own
cross
device
first
party
datasets
DATA
AGGREGATORS
-‐ First
party
data
is
still
the
treasure
trove..
Need
third
party
primarily
for
prospecting
and
augmenting
-‐ Trying
to
create
the
real
‘360
view’
of
consumers
-‐ Need
matching
technologies
that
work
cross
device
-‐ Seeing
increasing
competition
from
DMPs
(
See
Acxiom’s
2013
10K
filing)
-‐ Matching
cross
device
is
an
opportunity
-‐ No
cookies
on
mobile
means
the
dynamics
of
data
and
matching
change
a
lot
REGULATION
Opportunities
-‐ Overall
what
is
needed
is
a
way
for
consumers
to
o Opt-‐out
o Correct
information
o Know
who
holds
what
data
and
demand
that
data
from
them
o Keep
personal
data
away
from
job
&
financial
decision
makers
-‐ Consumers
are
worried
about
privacy
but
they
want
to
‘share
within
reason’
See
the
trends
around
‘influencer
marketing’
Consumers
are
happy
to
share
information
on
brands
they
like..
-‐ Plenty
to
learn
from
analogies
from
stock
trading
&
credit
bureaus.
Challenges
-‐ Being
overly
draconian
or
cost
prohibitive
to
implement
-‐
Govt.
looking
to
regulate
consumer
data
like
credit
information
see
Rockefeller
Data
bill.
But
the
bill
has
the
risk
of
being
overly
draconian
or
not
being
implementable
at
all.
-‐ Aligning
interests
from
various
groups
amidst
intense
lobbying
8.
M&A
Trends:
The
M&A
activity
exploded
in
2011
with
with
60
deals
valued
at
a
total
of
$4.5
billion.
It
shrunk
a
bit
in
2012
but
has
started
to
pick
up
steam
again.
M&A
in
the
display
advertising
space
will
happen
across
the
following
three
dimensions
1. Mutual
consolidation
a. The
large
DMPs
&
some
DSPs
will
be
the
acquirers
more
often
than
not.
b. ‘Data
assets’
will
be
one
of
the
key
magnets.
Eg.
Criteo’s
acquisition
of
Ad-‐X
was
a
lot
about
Ad-‐X
has
amongst
best
databases
around
mobile
app
tracking
in
the
industry
and
in-‐app
is
where
consumers
spend
majority
of
their
time
on
mobile.
c. Many
customers
find
managing
multiple
vendor
relationships
hard.
Single
throat
to
choke
adds
simplicity
d. Several
agencies
are
looking
to
build
their
own
end
to
end
AdTech
stacks.
DSPs
with
integrated
DMP
capabilities
and
Ad
serving
capabilities
seem
like
their
targets
2. Traditional
enterprise
tech
companies
jumping
in
–
a. Their
core
tech
businesses
are
rotting
due
to
the
commoditization
of
the
infrastructure
tech
and
shifting
of
value
at
higher
level
of
the
stack.
As
a
result
these
companies
are
looking
for
greener
pastures
b. They
already
started
with
marketing
automation
acquisitions
–
DMPs,
DSPs
&
data
aggregators
are
next
c. As
advertising
moves
from
being
an
‘Art’
to
being
a
science,
the
business
&
operating
models
look
familiar
&
attractive
to
the
traditional
tech
companies.
d. Their
stocks
are
cheap
BUT
they
are
awash
in
cash.
As
a
result
they
can
pretty
much
buy
into
any
space
that
is
being
disrupted
by
startups.
As
you
can
see
from
Appendix-‐2,
IBM,
ORCL,
MSFT
(
new
CEO
might
like
marketing?),
SAP
&
Salesforce
put
together
have
more
than
$200B
in
cash/current
assets
e.
3. Large
internet
companies
expanding
portfolios
a. Primarily
looking
to
vertically
integrate
and
provide
a
unified
customer
experience.
Adobe
is
the
closest
to
building
a
complete
stack.
Adobe
has
set
the
standard
in
terms
of
having
the
most
integrated
platform.
But
the
key
risk
to
manage
in
doing
so
is
not
creating
a
platform
that
becomes
so
generic
that
it
doesn’t
solve
any
use
case
really
well..
9. b. Companies
like
Google,
Apple,
Facebook
already
possess
a
lot
of
leverage
as
they
can
easily
tweak
the
browser
,
social
and
mobile
platforms
that
provide
the
underpinnings
of
the
data
that
is
driving
the
display
ad
revolution
c. Expected
to
boost
acquisitions
in
the
mobile
ad
targeting
area.
Appendix
1. Complicated
media
ordering
process
2. Traditional
tech
awash
in
cash
10.
3. Expanding
DSP
use
cases
4. Overview
of
the
ecosystem
buckets
and
the
blurring
of
lines
http://www.adexchanger.com/data-‐driven-‐thinking/the-‐new-‐digital-‐ad-‐
ecosystem/
5. List
of
key
players
in
every
bucket
of
the
display
advertising
space
http://www.lumapartners.com/lumascapes/display-‐ad-‐tech-‐lumascape/
6. The
rise
of
programmatic
–
Programmatic
RTB
is
on
the
rise
but
it
is
still
a
small
sliver
of
the
budget.
Bulk
of
the
money
is
still
in
DIRECT
buys..
so
now
folks
are
creating
programmatic
DIRECT
11. http://www.the-‐makegood.com/2014/03/18/ad-‐tech-‐investors-‐are-‐
wasting-‐millions-‐on-‐buyer-‐interfaces/
7. The
introduction
of
DEALID
–
allows
to
capture
the
nuances
of
a
person
to
person
negotiation
in
a
deal
the
agencies
then
take
their
dealID
that
they
struck
with
the
publisher
and
use
that
to
bid
for
the
publishers
inventory
in
the
open
exchange..
this
allows
the
exchange
to
better
match
the
inventory
from
the
publisher
to
the
agencies
needs.
Who
uses
them
-‐
Demand-‐side
platforms
like
Turn,
Invite
and
Mediamath.
And
private
exchanges
like
those
offered
by
Google,
Pubmatic
and
The
Rubicon
Project
and,
soon,
AppNexus.
8. Device
graph
on
mobile
(vendors
like
Tapad
&
AdMobius)
These
platforms
use
a
method
of
audience
targeting
often
called
"probabilistic"
identification,
designed
to
overcome
the
cookie
limitations
of
the
mobile
channel
by
building
detailed
profiles
linked
to
individual
device
characteristics.
Per
AdMobius
-‐
"We
are
ingesting
multiple
different
types
of
IDs,
never
the
original
UDID,
never
the
original
device
ID,"
Grigorovici
told
AdExchanger
in
2012.
"We
index
everything
in
our
database
in
terms
of
our
own
AdMobius
ID….
We
essentially
stitch
together,
if
you
will,
multiple
different
non‑personal
identifiable
IDs."
9. Current
&
future
DMP
use
cases
(Source:
Winterberry
consulting)
Current:
12.
Future:
10. Making
Native
ad
formats
work
http://www.forbes.com/sites/groupthink/2014/03/10/making-‐native-‐
advertising-‐work-‐for-‐you/