This document discusses research and how it can inform brand strategy and marketing actions. It describes how a Vice President of a consumer goods company sees her role as selling soap through research. The document advocates using a wide range of data sources and digital analytics to understand customer behavior, brand engagement, and media impact in real-time. It proposes applying experimental design and other research techniques to digital marketing problems to help develop loyal customer segments and attribution strategies. Finally, it promotes an open-source community called ARO to bring together professionals from various analytics fields.
Creating Low-Code Loan Applications using the Trisotech Mortgage Feature Set
1130 omma metrics joel rubinson
1. Joel Rubinson
February 21, 2012
Joelrubinson@gmail.com
Blog.joelrubinson.net
Twitter: @joelrubinson
The ideas expressed are
the property of
Rubinson Partners, Inc. 1
2. Donna is a researcher (now called
insights)
Donna is Vice President,
Consumer & Market
Insights, Unilever, Americas
When I asked what she does for a
living, she said…”I sell soap.
Research is just what I do”.
2
Rubinson Partners, Inc.
3. Research is about informing brand strategy
and marketing action in a grounded
way…evidence, analytics, insights
◦ How do I grow brand sales effectively and
efficiently?
◦ How do I create brand loyalty?
◦ How do people use my brand?
◦ How do people shop?
◦ How do target consumers consume and respond
to media?
Rubinson Partners, Inc.
4. A digital river of information
•fed by numerous tributaries
•Reflecting changing societal values, lifestyle trends, and brand
engagement
•Mostly naturally occurring, on-going 24/7
•Some tributaries are “man-made” and proprietary like a brand
tracker
•Start research with river rather than projects
4
Rubinson Partners, Inc.
5. http://www.youtube.com/watch?v=BYfbjZ
mk2VQ
5
Rubinson Partners, Inc.
6. The value of forensic analysis
◦ Massive databases of actual behavior,
native to the process, have been
proven to have great insights
value…DunnHumby frequent shopper
data.
◦ Digital is native to the shopping
process for offline purchases (as well
as TV is native to online).
Must be studied
Impact must be measured
Rich in forensic insights
Instead of an attribute rating on “dinner
ideas”, ask..when people search for
“dinner ideas” where do they wind up?
6
Rubinson Partners, Inc.
7. Brand tracking needs to position brands to
succeed in the digital future
7
Rubinson Partners, Inc.
8. While valuable for brand imagery and traditional media, brand
survey tracking is too slow, sample sizes are insufficient, and
recall is inaccurate for owned and earned media.
◦ The promise of digital is real time, and you must sense the
naturally occurring digital signals to respond instantly.
With brand tracker data, all
measures are strongly
correlated to market share.
Social media and digital data
look nothing like that.
Ex: although Starbucks is only
twice as large as Dunkin
Donuts, it has more than 10X
the buzz
8
Rubinson Partners, Inc.
9. Marketing is losing a handle on how much brand
communication is out there
◦ Paid
◦ Owned
◦ Earned
Hard to measure
◦ Marketing mix modeling usually done at a market level
◦ Owned media is way undervalued
◦ Earned media impressions is way larger than social
media comments (but much less than Facebook would
lead you to believe!)
◦ And what about search impressions?
9
Rubinson Partners, Inc.
10. We need to learn from each other
◦ Advanced analytic tools used by research can
benefit digital analytics
Loyalty modeling
Game theory statistics
Conjoint analytics
◦ Understanding digital possibilities will benefit
research analytics
10
Rubinson Partners, Inc.
11. Share of Requirements distribution for a 10% share brand; avg.
category purchase cycle of 12x/year
% category 70%
buyers
60%
50%
40%
30%
20%
10%
0%
Share of
25% 50% 75% requirements
12. Share of Requirements distribution for a 10% share brand; avg.
category purchase cycle of 12x/year
% category 70% 1. Which loyalty
buyers
60%
segments grow and
which decline as the
brand grows?
50%
40% 2. Where do Facebook
30% fans come from?
3. How can digital
analytics help to
20%
10%
develop strategies
0% for less loyal
25% 50% 75%
Share of groups?
requirements
measure fringe Only OK preferred “engaged”
% Buyers 12 5 4 5
Your brand share 2 34 60 90
of group
% sales 19 17 21 43
13. • Applied in marketing
research analytics to TURF
analysis problems and
regression modeling to
determine relative value of
different predictor variables
in the face of
multicollinearity
• A possible tool for
attribution analysis?
14. Experimental design used to understand the
relative impact of an attribute on choice
◦ Example:
Brand X, Y, Z
Protein, no protein
Different price levels (low, medium, high)
Packaging (resealable, not resealable)
Couldn’t this work for display advertising?
◦ Image, no image
◦ Rich, not rich
◦ Ad unit format
◦ Media placement approach
14
Rubinson Partners, Inc.
15. Judah and I see great value in an open source
community for all that we call “arrow” (Analytics
Research Organization), http://www.thearo.org/
◦ ARO ("arrow") brings together professionals globally
and locally from marketing research, digital analytics,
modelers, predictive analytics, business analytics and
other experts who are interested in consumer and
shopper data, big data, digital analytics, traditional
media analytics, media mix modeling and other forms
of research and analysis.
15
Rubinson Partners, Inc.
Editor's Notes
Quincy md
Game theory
People and other organizations. Open source sergei--