Internet Week Yahoo! Academy: How Direct Can You Be?
1.
2. How Direct Can You Be?
Increasing sophistication of the
response generation
Dora Michail
Senior Manager, DR Marketplace Management & Data
Strategy EMEA
3.
4. Why use DR? Example of a campaign
Registration & Interest Data
Start with the
1 top of the Travel
funnel
5. Why use DR? Example of a campaign
1 Start with the top of the funnel Registration & Interest Data Travel
User Journey & User Intent
Harness user Data
intent data Flights
2
Specific insight as to NY
they traverse the
Yahoo! DR network
6. Why use DR? Example of a campaign
1 Start with the top of the funnel Registration & Interest Data Travel
2 Harness user intent data User Journey & User Intent Data Flights to NY
Advertiser Data
Re-targeting to
‘close’ Brand
3
Specific insight as interest
they traverse the
Y! DR network
Re-targeting
7. It’s a complex market....
DMP
DSP SRT RTB
ATD SSP
TPAS
DRMP
AMP
As an industry,
we thrive on hype…
8. But...
Target Reach
audience
… the core principles
DR of Direct Response
People Data remain the same.
Technology
10. But so does data....
CASE STUDY
Significant uplift in Using ad
Campaign uplift using conversions using re-targeting
additional data to
Yahoo! Direct Response vs drive targeting
untargeted run-of-network (RON) 240%
Uplift
150%
Uplift
Conversion
RON Interest Data Intent Data Advertiser Data
12. And you need the machinery....
Predictive
Campaign Global
Optimization Frequency
AD SERVING & Management
OPTIMIZATION
ALGORITHM
EXCLUSIVE &
NON-EXCLUSIVE
INVENTORY & DATA
Real Time Real Time Bidding
Bidding on Targeted Supply
on
Demographic &
Targeted Supply
Behavioural
Targeting
13. ...to address the whole funnel
Interest
Target users with the right
demographics and interests.
!
Critical to address the whole funnel
Intent
Scalable performance advertising:
Overlays intent data, cherry-
picking those ready to buy.
Decision
Intensify once key buying
signal has been given,
retarget with the right
message and frequency
SOLD
13 Yahoo! Presentation, Confidential 11/14/2011
15. Combining…
100+ data providers, 6,000+ data segments
including Yahoo!
The Challenge
Create something unique and
unduplicated for advertisers…
16. Combining…
1 First party data, e.g. advertiser retargeting pixels
2 Proprietary demographic and behavioral data
3 Third party data
Reach Engage
Relevant audiences at scale Users as individuals
17. Pre-Targeting Segments
A machine learning based Typical data management
methodology to combine multiple
sources of data into unduplicated and Demo Data
Auto Interest Forum
unique audiences. data data Finance
data Finance
data
Autos Travel data
data Gaming
Each Pre-Targeting segment is a data Generic data
group of users that have similar Interest data Offline Data
online interests and a distinct
propensity to buy certain products
(e.g. travel, telco…).
Super Super
Segment Segment
This propensity to purchase is the
first of the many targeting tools Target for Auto and Target for Telco and
Finance Dating
Yahoo! DR uses