1. Attribution is challenging in a post-cookie mobile world where most consumer data and time is spent within apps on mobile devices rather than browsers. Device IDs also frequently change, making tracking difficult.
2. Attribution needs to consider all aspects of delivering an ad, including time, location, audience, device, app, and consumer behavior within closed networks. The focus should be on measuring what drives results rather than just publisher credit.
3. A case study by Pandora and Kellogg's showed that an audio campaign led to a 7% sales lift and over 3x return on ad spend, with most incremental sales impact coming from audio impressions compared to video and display.
Digital Transformation 'Before and After' seminar 10th February, Edinburgh
Adtech SF 5-21-15 Pandora_Kellogg_NCS FINAL
1. 1CONFIDENTIAL AND PROPRIETARY 1CONFIDENTIAL AND PROPRIETARY
THE LAST OF
THE LAST CLICK
ATTRIBUTION
Thursday, May 21, 2015
3:00 PM - 3:50 PM
Results Track
ADAM PAULISICK
SVP, Marketing & Strategy
Nielsen Catalina Solutions
@NCSolutions
@Paulisick
HEIDI BROWNING
SVP, Strategic Solutions
Pandora
@pandorapulse
@Heidi_Golightly
2. 2CONFIDENTIAL AND PROPRIETARY 2CONFIDENTIAL AND PROPRIETARY
THE NOT SO MOBILE FRIENDLY
HISTORY OF ATTRIBUTION
First Touch Last Touch Linear Recency
4. 4CONFIDENTIAL AND PROPRIETARY 4CONFIDENTIAL AND PROPRIETARY
WE ARE ENTERING THE POST-COOKIE WORLD
86% of mobile time is spent in apps, NOT the browser
Share of media time spent per day by US adults (%)
People are shifting from desktop to mobile
14% BROWSER 86% APPS
13.4
19.2
23.3
20.7
19.2 18.0
Mobile
Desktop/
laptop
Source: Flurry Analytics, ComScore, NetMarketShareSource: eMarketer
Social Messaging
9.5% (other)
2012 2013 2014
Gaming 32%
Facebook
17%
Twitter 1.5%
Entertainment 4% (other)
Utilities 8%
Productivity 4%
News 3%
Others 3%
Google Browsers 5%
Others 2%
Apple Safari 7%
YouTube 4%
5. 5CONFIDENTIAL AND PROPRIETARY 5CONFIDENTIAL AND PROPRIETARY
§ Without cookies, anonymous users are identified only
by mobile advertising IDs (MAIDs)
§ Devices are changed on average every 18 months,
making MAID an inconstant identifier
§ Most consumer data is siloed within apps
§ Cookie-based desktop data is still unplugged for mobile
Ad targeting in mobile is challenged by Advertisers and Agencies
WHAT DOES MOBILE MEAN FOR ATTRIBUTION?
6. 6CONFIDENTIAL AND PROPRIETARY 6CONFIDENTIAL AND PROPRIETARY
DELIVERING
ON MOBILE ROI
Attribution now applies to time, location,
audience, device, app, etc… but could
also be a consideration in closed networks
for consumer (listener) behavior
POV Shift: Attribution is not just about
publisher but every element of delivery
7. 7CONFIDENTIAL AND PROPRIETARY 7CONFIDENTIAL AND PROPRIETARY
WHAT GETS THE CREDIT?
More importantly, what’s the measure?
Start from the sale
and work backwards
8. 8CONFIDENTIAL AND PROPRIETARY 8CONFIDENTIAL AND PROPRIETARY
KELLOGG’S POP-TARTS
STRATEGY AND
OBJECTIVES
• Deliver relevant content to young adults
of a household so they consider
Pop Tarts as a breakfast option
• Accelerate lapsed buyers
• Grow sales among existing buyers
10. 10CONFIDENTIAL AND PROPRIETARY
13-17
Mobile and
cross-platform
Content ninjas with
short attention spans
Ever changing
listening habits
WHO ARE THEY HOW WE REACH THEM WHERE WE REACH THEM
ENGAGING YOUNG ADULTS
Audio
Display
Branded Radio
Mobile
Tablet
Web
13. 13CONFIDENTIAL AND PROPRIETARY
1. The end is the starting point
2. Show me you know me
3. Build an attention strategy
4. Invoke the power of audio
CAPTURE ATTENTION AND
DRIVE MOBILE ROAS
15. 15CONFIDENTIAL AND PROPRIETARY
HOW DO WE MEASURE INCREMENTAL SALES
Campaign Complete Exposed & Unexposed HHs matched
on 500+ variables to isolate impact
Audience Data NCS Purchase Data
Measure Sales Impact by
focusing on purchasing
patterns
Historical household/to show it’s not just a random
purchase and accounts for all other media.
16. 16CONFIDENTIAL AND PROPRIETARY
DIFFERENCE BETWEEN SALES & INCREMENTAL SALES
Your Ad Drove
Incremental Sales
Unexposed
Households
Exposed
Households
Exposed and Unexposed Comparison Accounts
for 52 Weeks Purchase History & Demographics
+
Typical Household
Purchase
TOTAL SALES
17. 17CONFIDENTIAL AND PROPRIETARY 17CONFIDENTIAL AND PROPRIETARY
DELIVERING ON MOBILE ROAS
+4%
7% LIFT
In Incremental Sales
3X+ ROAS
Exposed
Unexposed
18. 18CONFIDENTIAL AND PROPRIETARY 18CONFIDENTIAL AND PROPRIETARY
HOUSEHOLD
IMPRESSIONS
66%
16%
15%
3%
49%
28%
19%
2%
INCREMENTAL
SALES IMPACT
AUDIO
VIDEO
DISPLAY
2 + OVERLAP
KEY FINDINGS: BY MEDIA
19. 19CONFIDENTIAL AND PROPRIETARY 19CONFIDENTIAL AND PROPRIETARY
21%
19%
60%
17%
70%
13%
MOBILE DESKTOP
KEY FINDINGS: BY DEVICE
HOUSEHOLD
IMPRESSIONS
INCREMENTAL
SALES IMPACT
2 + OVERLAP