Attribution Marketing Mix Modeling

Report
Capabilities & Services
Copyright 2014 Marketing Productivity Group, Inc. - All rights reserved
Need Philosophy header
•
•
•
•
•
•
•
The world has changed, but the typical marketing science in place is stale and unevolved.
We are a Full Service Marketing Analytics Consultancy and our marketing science is based
on Cutting-Edge research and incorporates the new sources of social and mobile data. We
are State-of-the-Science.
Our solutions are Custom Fit to the clients needs and objectives.
We have extensive Academic and Industry Experience across the verticals and data that are
relevent to your needs.
We focus on the Profit or Objective Maximizing Solution and connect the dots between
Measuring Effect and Optimal Managerial Decisions.
We provide Transparent Results and a Dynamic Web Based Delivery.
Our Pricing is Competitive.
2
Using techniques based on sound academic research that translates into State-of-the-Science practice, we focus on
delivering insights for decision making and profit (objective) maximization.
Market Structure
Paid, Owned and Earned
Marketing Effectiveness
VAR
Marketing
Mix
Modeling
Insights for
Marketing
Decision Making
Custom
Solutions
Attribution
Single-Source
3
Test Market
Evaluation
Portfolio Price & Promo
Management
Pricing and
Assortment
Assortment
Planning
Demand
Planning
Long-Term
Effects
.
The Marketing Mix Modeling Landscape has Changed and Traditional
Models are Struggling
•
Treating online marketing as just another direct driver of sales leads to misattribution of drivers and ultimately
bad mix decisions
•
Even the best of today’s commonly used approaches have failed to keep up with current challenges
•
This following picture simply no longer holds true:
TV
•
x
Print
It’s time to
Hit Reset
Radio
Online
Sales
and begin employing techniques that capture today’s consumer and marketing
dynamics and provide integrated measurement of both online and offline marketing
tactics
4
Typical Conversion Attribution Does Not Account for the Consumer’s Full
Path to Purchase
A Single-Source Example:
Search
Incomplete Attribution Scope
Paid
Organic
Re-Market
Site
Site
Conversion/
Objective
Site
Conversion/
Objective
Media Tactics
You need to look at the whole picture.
Complete Attribution Scope
Search
Paid
Organic
Re-Market
Site
Media Tactics
5
Traditional Top Down and Bottom Up Approaches Don’t Line-up
Inconsistent Approach

The New Paradigm
- Consistent Approach
Marketing Mix Modeling –
Top Down
Insight and Decisions
Attribution Modeling
Bottom Up

Marketing Mix Modeling –
Top Down
Insight and Decisions
–
Insight and Decisions
Attribution Modeling
Bottom Up
6
–
Silo-ed Pricing and Assortment Only Reveal Part of the Story and Rely on Rigid
Assumptions
Demand Group N
Pre-defined Group
7
Sku G
…
Demand Group M
Sku R

Sku Z
Sku A
Isolated SKU
Level Approach
Pre-defined Group
Demand Group 1
…
Sku Z
Demand Group 1
We let product
attributes determine
demand groups.
•
Treating SKUs in
an isolated framework
ignores micro-economic
theory.
•
Pre-defined
Demand Groups are
based on rigid and
unverified assumptions.
...

•
Sku A
N Pre-defined
Demand Groups
Demand Category Modeling

Mobile and Social Media have Reshaped Consumer Behavior
Mobile and social media allow consumers to interact with a brand like never before. Now, customers enter into an openended relationship with a brand or product, where they cycle around considering, evaluating, and advocating. Many share
their experiences online. The new modeling approaches need to incorporate this type of dynamic behavior.
consider
bond
evaluate
advocate
enjoy
buy
new media make the “evaluate” and “advocate” stages increasingly relevant to
the process.
8
Source: Edelman (2010) "Branding in the Digital Age" in Harvard Business Review.
Updated Analytics for a New World, Based on Cutting-Edge Research – Solutions for Dynamic
and Interconnected Effects. We are the pioneers driving the advances in marketing science.
•
In Marketing Mix Modeling, Vector Autoregression (VAR) provides a proven tool for explaining the "network" of direct
and indirect effects of paid, owned, earned, and shared media inherent in the consumer decision process.1,2
•
Marketing Response Functions (MRFs) trace the attribution of marketing actions across the dynamic network quantifying
how the evaluate and advocate stages influence repurchase or other potential consumer behavior on an ongoing basis.1
•
Our Conversion Attribution approach addresses the impact and interaction in media impressions across media channels
(or platforms) on sales, using single-source data.3
•
The Cutting Edge Approach models spatial competition within a model of product demand to measure the CategoryLevel Pricing and Assortment impact on sales.4
1 Pauwels
(2004) "How Dynamic Consumer Response Shapes Long Term Marketing Effectiveness" in Marketing Science.
et. al (2012) "Beyond Likes and Tweets" a Stern Center for Measurable Marketing working paper.
3 Cohen, Bollinger, and Lai (2013)“Measuring Asymmetric Persistence and Interaction Effects of Media Exposures Across Platforms” a Wharton Consumer Analytics Initiative Working
Paper.
4 Cohen (2009) "Assessing the Impact of Retailer Store Brand Presence on Manufacturer Brands in an Equilibrium Framework“ in International Journal of Industrial Organization
2 Stacey
9
The MPG partners have founded and helped successfully lead several of the key marketing analytics suppliers,
agencies, and academic departments, as well as participated as decision-makers at client-side organizations. At
the same time, our advisors are on faculty and worked at leading universities and have a proven track-record as
thought leaders on cutting-edge approaches to solving marketing problems.
10
Over the past twenty-five years, the members of the MPG team have been able to apply our
techniques to numerous clients across multiple sectors in every corner of the globe.
Packaged Goods
Durables
Automotive
Financial Services
Pharmaceuticals
Traditional & Online Retail
Telecom
Travel and Entertainment
Restaurants
11
We Know the Modeling Needs of Different Verticals
• CPG – store purchases / online supermarket – Marketing Mix Modeling and Pricing and Assortment.
• Insurance – Marketing Mix Modeling with the objective of optimizing prospective sales and Custom Consumer
Sentiment Modeling.
• Media Agency– Marketing Mix Modeling with multiple objectives including optimizing ratings and traffic to a
network and driving sales from ads on a network to a product.
• Media Provider – Marketing Mix Modeling with the focus of measuring the effectiveness of specific campaign.
• Pharma – Marketing Mix Modeling with the objective of optimizing sales.
• Auto – Marketing Mix Modeling with the objective of optimizing sales.
• Finance –
• Online Travel – Marketing Mix Modeling with the objective of driving web traffic and mobile device usage.
12
Solutions Fit to Your Needs with the Latest Science
• Specializing in VAR: Cutting-edge approach to handling direct and indirect tactics across paid, owned, earned, and shared
in a dynamic and interconnected environment.
Media Tactics
Example of a Two Equation System:
Indirect
Previous Period’s Behavior
Direct
Indirect
Earned
Indirect
Earned
Direct
Indirect
Owned
Earned
Owned
Earned
Conversion/
Objective
Conversion/
Objective
Previous Period’s Behavior
• Attribution: Cutting-edge approach to handling direct and indirect tactics across paid, owned, earned, and shared in a
dynamic and interconnected environment at the consumer level.
• Pricing and Assortment: Spatial-competition demand models.
• Custom Model-Driven Solutions: We are a fully-equipped and highly-experienced statistical and econometric team. We
have the custom solution to your problem.
13
Transparent Results and Output
Quantification of the path-to-purchase for each marketing variable
Decomposition of sales (or other KPIs)
TV
Paid
Sales
44.9%
Base
Trade
29.1%
TV
8.7%
Print
2.6%
Paid Search
2.5%
FSI
1.5%
1.2%
0.3%
1.2%
0.3%
Company
website
4.6%
Satisfaction
0.1%
Quality
0.3%
Google (Natural)
2.7%
Brand
Owned Earned
Health
Online Display
Brand Posts
Buzz
Recommend
TRADE
Breakdown of the direct and indirect effects for each marketing variable
Models quantify sales response over time
14
We Will Guide You Through the Results and Focus on Profit (or Objective)
Maximizing Actions
• Marketing Mix Modeling: Budget allocation across media tactics from the framework of
profit or marketing objective optimization.
• Attribution – Budget allocation across media tactics from the framework of profit or
marketing objective optimization at the consumer level. Our Top-Down and Bottom up
approach are consistent.
• Pricing and Assortment: Demand models, allocation across media tactics, and optimized
assortment.
• Custom support across objectives.
• CFO anaylsis – KPMG expertise on addressing financial
• Bill Harvey
15
16
E. Craig Stacey, PhD, Chief Executive Officer, is a recognized expert in the area of marketing productivity analysis with a special emphasis on marketing mix modeling and
Jim Friedman, need title, brings 30 years of entrepreneurial and leadership experience to MPG including successfully starting and selling two companies in the marketing
successfully starting and selling two companies in the marketing analytics space. Most recently, he served as Global Executive Director and Chief Strategy Lead for the
Global Executive Director and Chief Strategy Lead for the Marketing Analytics practice of Accenture Marketing Sciences. Previously, Jim served as President and CEO of
Sciences. Previously, Jim served as President and CEO of ImmediateFX Holdings, Inc., a technology-based marketing consultancy delivering continuous marketing measurement
marketing consultancy delivering continuous marketing measurement solutions. At ImmediateFX, he was the founding partner and driving force in transforming IFX from an
founding partner and driving force in transforming IFX from an industry vision to the leading provider of continuous marketing measurement solutions, helping companies
marketing measurement solutions, helping companies improve their marketing performance and Return on Investment across the consumer goods, pharmaceutical, retail and
Investment across the consumer goods, pharmaceutical, retail and service industries.
Stephen Dubuque, Chief Operating Officer, is the managing partner at MPG whose 25 years of business background includes both custom and syndicated research. Prior to
background includes both custom and syndicated research. Prior to joining MPG, Steve was a Senior Vice President, Director of Research at Universal McCann. In addition,
President, Director of Research at Universal McCann. In addition, Steve has worked on both the client and research supplier sides at Kraft, Advanis, and Information Resources,
Koen Pauwels, Chief Academic Advisor, is Professor of Marketing at Ozyegin University, Istanbul and Honorary Professor at the University of Groningen. He received his
Professor at the University of Groningen. He received his Ph.D. from UCLA, where he was chosen “Top 100 Inspirational Alumnus” out of 37,000 UCLA graduates. Applying
Inspirational Alumnus” out of 37,000 UCLA graduates. Applying his award-winning research to companies across 3 continents, Koen just published his first book ‘It’s not the
continents, Koen just published his first book ‘It’s not the Size of the Data, It’s How You Use it: Smarter Marketing with Analytics and Dashboards”.
17
John Liska, Chief Technial Officer, has over 15 years of experience in high tech and software with an emphasis on cloud-based large-scale data enterprises within Retail and
on cloud-based large-scale data enterprises within Retail and CPG. Prior to joining MPG he was the Sr. Director of Integration Services at DemandTec. In this role he
Michael A. Cohen, PhD, is the Chief Marketing Scientist and technology engineer at Marketing Productivity Group as well as a fellow of NYU Stern's Center for Measurable
Group as well as a fellow of NYU Stern's Center for Measurable Marketing. His areas of expertise are media and pricing strategy, corporate social responsibility, consumer
pricing strategy, corporate social responsibility, consumer packaged goods, quantitative methods in marketing, and econometrics. He received his Ph.D. from the University of
econometrics. He received his Ph.D. from the University of Connecticut, where his dissertation was awarded, by the Food Distribution Research Society, as the best dissertation
the Food Distribution Research Society, as the best dissertation in food marketing in 2009. Dr. Cohen has served as a Marketing Professor at NYU’s Stern School of Business.
Winston Bradley, Senior Economist, came to MPG 2013 as a specialist in time-series econometrics. Prior to joining MPG, Winston worked as a quantitative analyst on the
joining MPG, Winston worked as a quantitative analyst on the energy trading floor of Southern Company, the U.S.'s second largest power utility, where he developed a platform
second largest power utility, where he developed a platform for programmatic electricity trading and gross margin reporting, as well as a first-of-its-kind automated daily
reporting, as well as a first-of-its-kind automated daily counterparty load forecasting model. Additionally, he gained retail banking analytics experience while working as a data
retail banking analytics experience while working as a data miner for BBVA-Compass. Winston has a master's degree in economics from the University of Alabama.
Jake Katz, Chief Marketing Officer, has over 10 years of years experience with data analysis and associated applications in academics and business. Prior to joining the MPG, he
applications in academics and business. Prior to joining the MPG, he launched an e-commerce concept company and worked as a quantitative analyst at Brevan Howard UK's
18
References – as many and as prestigious as
possible
19
Contact Us
Craig Stacey PhD, Founding Partner
www.mproductivity.com
E: [email protected]
P: +1.212.796.0863 ext. 7002
C +1.404.202.7367
Jake Katz, Director of Business Development
E: [email protected]
P: +1.212.796.0863 ext. 7003
C +1.203.464.0729
20

similar documents