Unraveling Multimodality with Large Language Models.pdf
How To Get Low CPI High Value Users Via Mobile Ad Networks by Naghi Prasad
1. How to get low CPI and High Value Users via
mobile ad networks
Naghi Prasad
Co-Founder, Deep Forest Media
March 2014
2. 2
Agenda
• Introduction
to
paid
user
acquisition
• Measure
Life
Time
Value
of
Users
• Increase
Life
Time
Value
of
users
• Decrease
the
cost
of
acquiring
high
value
users
3. 3
User Acquisition Dilemma
• Increasingly, revenues and downloads are decoupling.
• Rage
of
Bahamut
is
not
a
top
500
iOS
app
by
download
but
#14
app
in
revenue
in
US,
DARPU
~$0.80!
• Clash
of
Clans
–
one
of
the
top
grossing
iOS
apps
is
not
even
in
top
50
in
downloads
Getting the right user is the game!!
4. 4
Net positive return on users
• Life TimeValue (LTV) of user
• Profit per user. = LTV – CPI
Cannot optimize for what you don’t measure
Measure
5. 5
LTV Calculation – Too Late to Affect Ad Campaign
Days
Leading Indicators for LTV prediction
LTV and Leading Indicators
Measure
7. 7
Freemium versus paid apps
• Freemium is the dominant model
• Is Freemium a good model for your app?
• Paid apps are hard to get through paid acquisition
Reduce
costs
8. 8
Campaign types
• CPI
• Minimal
risk
• Hard
to
get,
at
volume
• CPC
• Conversion
risk
still
on
the
advertiser
• CPM
• Maximum
risk;
maximum
reward
• Willingness
to
experiment
and
take
losses.
Reduce
costs
9. 9
Supply Sources
• Over 100 supply sources
• Native
platforms
(Facebook,
….)
•
Ad
Networks
• RTB
Exchange
• Types of inventory
• Native
platforms
•
Incentivized
vs
non-‐incentized
• Video
(incented
video,
non-‐incented
install
vs
non-‐incented
video)
• Pay
per
Engagement
• Rich
Media
Reduce
costs
10. 10
Retargeting
• Very useful technique for mature apps
• RTB
exchanges
ideal
for
this
•
Work
with
a
platform
that
keep
your
data
siloed
and
targets
on
your
behalf.
• Any
kind
of
data
advantage
you
have
falls
in
this
bucket.
Reduce
costs
11. 11
Deep Forest Media
Easy
§ One click
buying
Transparent
§ One-stop unified
reporting
Automated
§ Algorithmic
spend
optimization
advertisers mobile users