How Kellogg's Brands Work Their Digital Strategies
1. Case Study
How Kellogg’s Brands Work Their
Digital Strategies
Aaron Fetters Jeff Smith
Associate Director CMO & SVP
Global Digital Client Services
Strategy & Analytics Vizu
Kellogg’s
2. The Online Advertising Outlook for 2012
Industry Study
• A survey of more than 450 digital marketing and media
professionals conducted in late 2011 by Digiday Media & Vizu
• Participants included Brand Marketers, Media Agencies, and
Media Sellers
• Questions covered both the online brand advertising outlook for
2012 in addition to best practices for online brand advertising
3. The Online Advertising Outlook for 2012
Online considered a powerful brand-building medium
• Direct Response Advertising Growing:
– 56% of brands respondents said they will increase spending
in 2012
– 15% said they will increase spending by more than 20%
• Brand Advertising Growing Even Faster:
– 64% of brand respondents said they will increase spending
in 2012
– 22% said they will increase spending by more than 20%
4. The Online Advertising Outlook for 2012
Brand Advertising surpassing Direct Response for 1st time
Brands: What type of digital advertising do you plan to use in 2012?
Direct
Response
41%
Branding
59%
5. The Online Advertising Outlook for 2012
Some channels growing faster then others
Brands: How will brand ad spending in the following channels compare to prior year?
Mobile Advertising 69% 26%
Social Media Advertising 63% 34%
Video Advertising 57% 34%
Rich Media Advertising 29% 57%
Standard Display Advertising 20% 60%
Connected TV/IPTV 17% 37%
0% 20% 40% 60% 80% 100% 120%
Increase Stay about the same Decrease Don't Use
6. The Online Advertising Outlook for 2012
Brand’s Perspective: What would make it better?
Brands: Which of the following would lead you to increase spending on online brand
advertising? (Check all that apply)
Improved clarity around the actual return
68%
on your brand advertising investment
Ability to verify my brand advertising
created the desired result (e.g. increased 56%
Cleaning up the
awareness of my product) “Metrics
Morass” in the
Ability to use the same metrics to evaluate online medium
brand advertising effectiveness online as 53%
are used offline
Purchasing efficiency (e.g. the ability to
reach audience at scale through fewer 50%
outlets)
Ability to verify my online brand advertising
was actually delivered to my target 38%
audience
7. The Online Brand Advertising Outlook for 2012
Challenge: The online “metrics morass”
Commenting on the amount of data bandied around in
the online ecosystem, 34% of Brands stated:
“I’m drowning in data; I would prefer to focus on a
few meaningful metrics to evaluate the
performance of my campaign”
8. The Online Brand Advertising Outlook for 2012
Challenge: The online “metrics morass”
Asked to what extent they would like to leverage Asked what they typically
offline metrics in the online medium, brands say: report, Media Sellers say:
The exact same metrics, and nothing else 6% 10%
The exact same metrics, and a few
additional metrics specific to the online 55% 14%
medium
Some of the metrics from the offline
medium, some metrics specific to the 24% 30%
online medium
Just metrics specific to the online medium 15% 47%
0% 10% 20% 30% 40% 50% 60% 0% 10% 20% 30% 40% 50%
9. The Online Brand Advertising Outlook for 2012
Challenge: The online “metrics morass”
Asked specifically for the most appropriate metrics for brand advertising,
marketers responded:
Brand Lift generated as a result of the
80%
advertising
Sales generated as a result of the
57%
advertising
Interaction rates with the advertising 31%
Clickthrough rates on the advertising 29%
Shares or re-posts of the advertising 29%
Time spent / Dwell time (e.g. on a micro-
20%
site)
0% 10% 20% 30% 40% 50% 60% 70% 80% 90%
10. The Online Brand Advertising Outlook for 2012
Additional challenges
Other key things brand advertisers are looking for:
• Proof on the front end: Only 6% of brand surveyed said they “Strongly
believe” media sellers claims they can reach the custom / niche audiences
the brands are looking for.
• Results on the back end: Nearly all respondents identified “in-market
optimization of campaigns against relevant Brand Metrics” and “Proof of
performance” as critical to maximizing ROI
11. The Online Brand Advertising Outlook for 2012
Conclusion
• In 2012, brand advertising is set to finally surpass performance based
advertising in the online medium
• That said, the outlook could be even better, if we work to clean up the
“metrics morass” in the online medium
• Use metrics that matter to maximize the return on your online brand
advertising investment
Download the full report
http://brandlift.vizu.com/knowledge-resources/research/2012-industry-outlook/
Benchmark Your Brand Advertising IQ based on this Survey
http://brandlift.vizu.com/knowledge-resources/brand-ad-iq/
12. Real Time Optimization Based on Metrics That Matter
Aaron Fetters – The Kellogg Company
13. It’s about getting the MOST
out of the $ you have
– We aren’t all the New York
Yankees (or Chelsea Football
Club)
Don’t buy players, buy wins
Identify stats indicative of
GAMES WON, vs. stats that
are easy to count
– On-base % vs. HR, RBI, etc.
– Wins Above Replacement
14. Digital Measurement – From a Brand’s Perspective
• Kellogg Company Background
• Where We’ve Been
– Challenges in Digital Measurement
• Where We Are
– Establishing Goals and KPIs
– What Have We Learned
• Where We Are Going
– Enhanced Planning
– Informed Buying
– Near-time Optimization
• What it Means for Media and Measurement Partners
16. • Attracting more of our
consumers’ media time
• Attracting more of our
Brands’ media $$$
• But filled with endless
options – our Brands are
looking for guidance
17. Though broad, historical efforts to track and measure digital have
been disjointed
1. Define a common set of Goals for
Digital Media
In the past 6 2. Establish the KPIs (metrics) that
REALLY matter to track against our
months we have Goals
3. Develop and communicate a
focused on the Digital Measurement Framework,
along with Principles for how we
opportunities to: will track Digital and capture KPI
results
18. Establishing a model for Media Effectiveness – applied initially to
Online Media
PRE-MARKET QUALITY • Branded Recognition
Does my creative have the right
to succeed? • Brand Buy Next
• % Impressions to Target
AUDIENCE DELIVERY • Frequency of Exposure
Did we efficiently reach our • % Impressions in View
Target?
• CPM
• Brand Awareness
IN-MARKET • Purchase Interest
EFFECTIVENESS • Message Association
Are we moving the needle? • Direct Response
19. Campaign targeted to a Hispanic
audience
“oh my gosh you are One publisher delivering only 1.2%
the best ever - of our Impressions to the
totally re-worked my demographic Target audience (that’s
a 98.8% waste folks)
plan just now. thank
you so However, another publisher had a
much!!!!!!!!!!!!!!!!!!!!” Targeting Efficiency of 5.1 – meaning
their Targeting ability allowed us to
reach our Target 5X better than if we
bought Run Of Network
20. Let’s take a look at Average Frequency Impressions Unique Viewers
a few of our biggest Total Campaign % on Target % on Target
campaigns from
2011
(Data provided is at an
individual
publisher, individual
campaign level)
21. Average Frequency Impressions Unique Viewers
Total Campaign % on Target % on Target
6 71.79% 61.26%
These publishers
effectively delivered to 4.6 61.55% 38.67%
our target and efficiently
balanced the exposures
22. Average Frequency Impressions Unique Viewers
Total Campaign % on Target % on Target
6 71.79% 61.26%
Conversely, this execution
is generating a great deal 4.6 61.55% 38.67%
of reach, but 75% is to
2.5 25.55% 27.52%
someone other than our
Target audience
23. Average Frequency Impressions Unique Viewers
Total Campaign % on Target % on Target
6 71.79% 61.26%
And in this case
Impressions are virtually 4.6 61.55% 38.67%
being thrown away as the
2.5 25.55% 27.52%
wrong audience is
reached over and over 15.7 21.04% 24.68%
again
24. Average Frequency Impressions Unique Viewers
Total Campaign % on Target % on Target
6 71.79% 61.26%
Publisher is doing a nice
job of finding our Target. 4.6 61.55% 38.67%
2.5 25.55% 27.52%
Did we ask for Frequency
capping with them? 15.7 21.04% 24.68%
14.5 69.57% 60.55%
25. Average Frequency Impressions Unique Viewers
Total Campaign % on Target % on Target
6 71.79% 61.26%
Incredible targeting
results from a high reach 4.6 61.55% 38.67%
portal
2.5 25.55% 27.52%
Opportunity to shift $ 15.7 21.04% 24.68%
here to reach more of our 14.5 69.57% 60.55%
target, at an increased
frequency 2.9 80.80% 80.48%
26. Near-Time
Decisions and
Optimizations
Cost per
Advertising Effective
Effectiveness Audience Cross-Media
Insights Impression
measurement and
Online optimization?
Audience
Delivery
Insights
2011 2012 2013
We have begun to build the foundation of our Digital Measurement needs. We now want
to move toward our vision of enabling near real-time decision making and optimization
based on knowledge and insights of what is working.
27. Efficiency/Effectiveness Publisher Overview
80.0%
60.0%
% of Target
40.0%
20.0%
0.0%
$0.00 $3.00 $6.00 $9.00 $12.00
CPM
•Improved DSP / Portal / Publisher evaluation based on historical performance
•Dig in deeper with Top Performers! – improve or eliminate buys far outside our KPIs
•Next step is to layer in Advertising Effectiveness
2/9/2012
28. Conversations previously NOT had with “Partner A”:
“in compliance with our 2011 RFP, we would like to
work with you to create makegoods for the following
campaigns:
•Special K Evening Snacking Q2/Q3
•Frosted Mini-Wheats
•Special K Fall Challenge
•Crunchy Nut”
Conversations needed with Partner X:
•What targeting approaches were used against Fiber
Plus and Special K but not applied to Nutri-Grain?
Publisher Performance Across Three Similar Campaigns
Average of Sum of
Impressions - Impressions -
Yahoo! X
Partner Target % on Target CPM Total
Fiber Plus OLM Q3 F25-54 79.00% $3.44 20,284
Nutri-Grain OLM Q3 F25-54 w-kids 24.80% $3.83 14,208
Special K OLM F25-54 51.10% $4.92 26,340
29. Combining Insights with costs to answer: HOW CAN I GET THE
BIGGEST BANG FOR MY BUCK?
At a campaign level we need
to drive optimization
decisions in near real-time
Media Avg. Impressions Impressions Lift in
CPM
Partner Frequency in Target in View Awareness
1 3.5 24.5% 83% 5.74 $2.44
2 3.0 16.9% 91% 0.55 $9.08
$ 3 8.4 23.5% 71% 0 $8.62
•Opportunity to optimize - $ follow performance
•Multiple publishers, each with high reach potential
•Wide variance in performance observed real-time
30. Media Partners Measurement Providers
Performance and results are the key to
Need continued research and analysis of
increased investments – help us get
which metrics matter!
more!
Rigorous and diligent pursuit of Research
Market Mix Models will reflect the
Quality
quality of what goes into them.
33. Real Time Optimization Based on Metrics That Matter
Budweiser Tackles a New Approach to
Content"
Aaron Fetters – The Kellogg Company
34. It’s about getting the MOST
out of the $ you have
– We aren’t all the New York
Yankees (or Chelsea Football
Club)
Don’t buy players, buy wins
Identify stats indicative of
GAMES WON, vs. stats that
are easy to count
– On-base % vs. HR, RBI, etc.
– Wins Above Replacement
35. Digital Measurement – From a Brand’s Perspective
• Kellogg Company Background
• Where We’ve Been
– Challenges in Digital Measurement
• Where We Are
– Establishing Goals and KPIs
– What Have We Learned
• Where We Are Going
– Enhanced Planning
– Informed Buying
– Near-time Optimization
• What it Means for Media and Measurement Partners
37. • Attracting more of our
consumers’ media time
• Attracting more of our
Brands’ media $$$
• But filled with endless
options – our Brands are
looking for guidance
38. Though broad, historical efforts to track and measure digital have
been disjointed
1. Define a common set of Goals for
Digital Media
In the past 6 2. Establish the KPIs (metrics) that
REALLY matter to track against our
months we have Goals
3. Develop and communicate a
focused on the Digital Measurement
Framework, along with Principles
opportunities to: for how we will track Digital and
capture KPI results
39. Establishing a model for Media Effectiveness – applied initially to
Online Media
PRE-MARKET QUALITY • Branded Recognition
Does my creative have the right
to succeed? • Brand Buy Next
• % Impressions to Target
AUDIENCE DELIVERY • Frequency of Exposure
Did we efficiently reach our • % Impressions in View
Target?
• CPM
• Brand Awareness
IN-MARKET • Purchase Interest
EFFECTIVENESS • Message Association
Are we moving the needle? • Direct Response
40. Campaign targeted to a Hispanic
audience
“oh my gosh you are One publisher delivering only 1.2%
the best ever - of our Impressions to the
totally re-worked my demographic Target audience (that’s
a 98.8% waste folks)
plan just now. thank
you so However, another publisher had a
much!!!!!!!!!!!!!!!!!!!!” Targeting Efficiency of 5.1 – meaning
their Targeting ability allowed us to
reach our Target 5X better than if we
bought Run Of Network
41. Let’s take a look at Average Frequency Impressions Unique Viewers
a few of our biggest Total Campaign % on Target % on Target
campaigns from
2011
(Data provided is at an
individual
publisher, individual
campaign level)
42. Average Frequency Impressions Unique Viewers
Total Campaign % on Target % on Target
6 71.79% 61.26%
These publishers
effectively delivered to 4.6 61.55% 38.67%
our target and efficiently
balanced the exposures
43. Average Frequency Impressions Unique Viewers
Total Campaign % on Target % on Target
6 71.79% 61.26%
Conversely, this execution
is generating a great deal 4.6 61.55% 38.67%
of reach, but 75% is to
2.5 25.55% 27.52%
someone other than our
Target audience
44. Average Frequency Impressions Unique Viewers
Total Campaign % on Target % on Target
6 71.79% 61.26%
And in this case
Impressions are virtually 4.6 61.55% 38.67%
being thrown away as the
2.5 25.55% 27.52%
wrong audience is
reached over and over 15.7 21.04% 24.68%
again
45. Average Frequency Impressions Unique Viewers
Total Campaign % on Target % on Target
6 71.79% 61.26%
Publisher is doing a nice
job of finding our Target. 4.6 61.55% 38.67%
2.5 25.55% 27.52%
Did we ask for Frequency
capping with them? 15.7 21.04% 24.68%
14.5 69.57% 60.55%
46. Average Frequency Impressions Unique Viewers
Total Campaign % on Target % on Target
6 71.79% 61.26%
Incredible targeting
results from a high reach 4.6 61.55% 38.67%
portal
2.5 25.55% 27.52%
Opportunity to shift $ 15.7 21.04% 24.68%
here to reach more of our 14.5 69.57% 60.55%
target, at an increased
frequency 2.9 80.80% 80.48%
47. Near-Time
Decisions and
Optimizations
Cost per
Advertising Effective
Effectiveness Audience Cross-Media
Insights Impression
measurement and
Online optimization?
Audience
Delivery
Insights
2011 2012 2013
We have begun to build the foundation of our Digital Measurement needs. We now want
to move toward our vision of enabling near real-time decision making and optimization
based on knowledge and insights of what is working.
48. Efficiency/Effectiveness Publisher Overview
80.0%
60.0%
% of Target
40.0%
20.0%
0.0%
$0.00 $3.00 $6.00 $9.00 $12.00
CPM
•Improved DSP / Portal / Publisher evaluation based on historical performance
•Dig in deeper with Top Performers! – improve or eliminate buys far outside our KPIs
•Next step is to layer in Advertising Effectiveness
2/9/2012
49. Conversations previously NOT had with “Partner A”:
“in compliance with our 2011 RFP, we would like to
work with you to create makegoods for the following
campaigns:
•Special K Evening Snacking Q2/Q3
•Frosted Mini-Wheats
•Special K Fall Challenge
•Crunchy Nut”
Conversations needed with Partner X:
•What targeting approaches were used against Fiber
Plus and Special K but not applied to Nutri-Grain?
Publisher Performance Across Three Similar Campaigns
Average of Sum of
Impressions - Impressions -
Yahoo! X
Partner Target % on Target CPM Total
Fiber Plus OLM Q3 F25-54 79.00% $3.44 20,284
Nutri-Grain OLM Q3 F25-54 w-kids 24.80% $3.83 14,208
Special K OLM F25-54 51.10% $4.92 26,340
50. Combining Insights with costs to answer: HOW CAN I GET THE
BIGGEST BANG FOR MY BUCK?
At a campaign level we need
to drive optimization
decisions in near real-time
Media Avg. Impressions Impressions Lift in
CPM
Partner Frequency in Target in View Awareness
1 3.5 24.5% 83% 5.74 $2.44
2 3.0 16.9% 91% 0.55 $9.08
$ 3 8.4 23.5% 71% 0 $8.62
•Opportunity to optimize - $ follow performance
•Multiple publishers, each with high reach potential
•Wide variance in performance observed real-time
51. Media Partners Measurement Providers
Performance and results are the key to
Need continued research and analysis of
increased investments – help us get
which metrics matter!
more!
Rigorous and diligent pursuit of Research
Market Mix Models will reflect the
Quality
quality of what goes into them.