Customer Mindset Metrics and Brand Performance Shuba Srinivasan Boston University "Measurable Marketing in the Path-to-Purchase NYU Stern School of Business September 28, 2012
Problem Statement Marketing Execs
Problem Statement Top Management Marketing Execs
Problem Statement Top Management Marketing Execs = f (, )
Problem Statement Top Management Marketing Execs = f (, ) New metrics of consumer behavior and activity (Paid, Owned, and Earned Media) on the Internet allow a deeper understanding of customer attitude and marketing actions
Research Objectives Does including paid, owned, and earned (POE) media add explanatory power to a sales response model that already includes marketing mix actions? *Pauwels, Srinivasan, Rutz and Bucklin (2012), The Hierarchy of Effects (HOE) Meets Paid, Earned, and Owned (POE): How Do Internet Media Work to Drive Brand Sales? Working Paper, Boston University.
Paper # 1: Research Objectives Does including paid, owned, and earned (POE) media add explanatory power to a sales response model that already includes marketing mix actions? How do POE elements interact with each other and with traditional marketing mix actions such as price, distribution and offline advertising? *Pauwels, Srinivasan, Rutz and Bucklin (2012), The Hierarchy of Effects (HOE) Meets Paid, Earned, and Owned (POE): How Do Internet Media Work to Drive Brand Sales? Working Paper, Boston University.
Research Objectives Does including paid, owned, and earned (POE) media add explanatory power to a sales response model that already includes marketing mix actions? How do POE elements interact with each other and with traditional marketing mix actions such as price, distribution and offline advertising? Are online metrics leading indicators of brand performance? *Pauwels, Srinivasan, Rutz and Bucklin (2012), The Hierarchy of Effects (HOE) Meets Paid, Earned, and Owned (POE): How Do Internet Media Work to Drive Brand Sales? Working Paper, Boston University.
CONCEPTUAL FRAMEWORK TRADITIONAL MARKETING MIX Distribution Do/Conation Brand Sales Price Advertising
CONCEPTUAL FRAMEWORK NEW MEDIA Do/Conation Brand Sales Learn/Cognitive Paid Search Clicks (Paid) Visits to own website (owned) Feel/Affective Facebook likes (earned) Facebook unlikes (earned)
INTEGRATIVE CONCEPTUAL FRAMEWORK HIERACHY OF EFFECTS TRADITIONAL MARKETING MIX & NEW MEDIA Distribution Do/Conation Brand Sales Price Advertising Learn/Cognitive Paid Search Clicks (Paid) Visits to own website (owned) Feel/Affective Facebook likes (earned) Facebook unlikes (earned)
Empirics Nielsen and Mindshare Store data on sales Distribution Advertising audit service Online metrics Purchases, Prices Distribution, Promotions Advertising Owned: websites Earned: social media Paid: Paid search 4 years of weekly data from June 2006-May 2010 for a leading brand of a low-involvement consumer packaged good; Online metrics available for a shorter time Approach: We estimate the dynamic interactions between brand performance (purchases), price, distribution, TV advertising, owned media (website visits), earned media (social media likes and unlikes), paid media (search) using Vector-Autoregressive models with exogenous variables (VARX).
TV aims to drive sales and online debate Pulsing TV Gross Rating Points 450 400 350 300 250 200 150 100 50 0 1/3/2009 4/3/2009 7/3/2009 10/3/2009 1/3/2010 4/3/2010 May lead to peaks in online visits for Major Customer Engagement Campaign 300000 250000 200000 150000 100000 50000 0 1/2/2010 4/2/2010
Though not all of it is positive for the brand e.g. negative reviews, blogs, Facebook Unlikes 100 90 80 70 60 50 40 30 UnEarned Media: Facebook UnLikes 20 10 0 1/2/2010 4/2/2010 7/2/2010 10/2/2010
RESULTS R1: TV leads online activity which leads sales Price Distribution TV Paid Search Sales Owned site visits Like Unlike
R2: Do Online Metrics Really Matter?
R3: How Much Do Online Metrics Matter? Price 20 %; Distribution 60%; T V Advertising 5%; Paid 2%; Owned 10%; Earned 3%
R4: HIERACHY OF EFFECTS LONG-TERM ELASTICITIES Distribution Price 0.07 2.73 0.04-3.53 TV Advertising 0.001 0.03 Sales 0.10 Paid Search (paid) 0.72 0.09 Visits to own site (owned) 2.42 0.05 2.72 0.25 0.37 0.81 Facebook likes (earned) Facebook unlikes (earned) 0.12 0.18 0.15-0.01
HIERACHY OF EFFECTS LONG-TERM UNIT CHANGES Distribution Price 0.89 1,000K 0.00001-200K 0.00007 TV Advertising 137 Sales 9.00 Paid Search (paid) 5.21 0.01 Visits to own site (owned) 0.53 0.23 0.08 8.25 0.37 40.25 Facebook likes (earned) Facebook unlikes (earned) 0.87 0.0045 1.80-319
From Insight to Opportunity: Paid, Owned and Earned Media and Brand Performance Insight Do online metrics add power to explain sales? YES! 15% of sales explained by online metrics Paid, owned and earned (POE) media empirically shed new light on the hierarchy-of-effects (HOE) theory New activity-based POE metrics have lower tracking costs and provide an opportunity for early warning signals Supports our intuition that new internet metrics could be a way to account for some of the long-term effects of marketing, e.g., brandbuilding effects Caveat Need Opportunity Individual brands may respond differently Extend academic work to more contexts Overall, our study should help strengthen marketers case for building share in customers hearts and minds as measured through customer online engagement via the POE metrics.
Paper #2 SCHEMATICALLY* Marketing Actions Advertising Brand Health Indicators Awareness Firm Performance Promotions Responsiveness Stickiness Consideration Conversion Brand Sales Brand Profit Distribution Stickiness Liking Product Stickiness Mindset Route Potential Mindset Route *Relevance criteria are in italics Transactions Route
Example 2: Proposed Approach What is the financial value of an extra consumer attitude point? Key financial metric is the sales conversion of that extra point
Example 2: Proposed Approach What is the financial value of an extra consumer attitude point? Key financial metric is the sales conversion of that extra point What are the drivers of sales conversion? Potential (distance from the maximum, e.g. 100% awareness) Stickiness (staying power in absence of further stimulation) Responsiveness (attitude lift from marketing actions)
Example 2: Proposed Approach What is the financial value of an extra consumer attitude point? Key financial metric is the sales conversion of that extra point What are the drivers of sales conversion? Potential (distance from the maximum, e.g. 100% awareness) Stickiness (staying power in absence of further stimulation) Responsiveness (attitude lift from marketing actions) How to measure sales conversion and its drivers? Equations for attitude response and transactions (sales) response Testing if attitude change is needed in order to achieve transactions response Differences between high-involvement and low-involvement products Econometrics: cross-nested and hierarchical linear mixed effects models
Results On a large dataset (7 years, monthly data) of multiple brands, categories in the consumer packaged goods Lower-funnel metrics have higher conversion, but are less sticky. Consumer attitude change matters more in high-involvement category Upper funnel is more important for sales conversion in the highinvolvement category Generalization: sales move with the square root of consumer liking Patterns are stable and brand-specific Responsiveness to marketing (lift) is predominantly brand-specific Resource allocation differs by brand *Hanssens, Pauwels, Srinivasan, Vanhuele and Yildirim (2012), Consumer Attitude Metrics for Guiding Marketing Resource Allocation, Working Paper, Boston University.
Results Why collect expensive and ambiguous mindset metrics? The use of these data has to outperform a straight (transactions-only) marketing mix model 1. Forecast test: one-month and 12-month ahead predictions 2. Recommendations test Diagnose brands recommendations observe managerial follow-up new diagnostic one year later
From Insight to Opportunity Customer mindset metrics can be important intermediate performance variables for guiding marketing spend Need to measure mindset effects together with transactions effects. In so doing, proof of sales conversion is key. Potential, stickiness and lift determine the relative importance of mindset metrics Differences emerge across involvement categories and across brands. Brand strategy matters! Interesting EGs can be derived, e.g. sales = f [ sqrt (liking) ] Mindset metrics allow for superior long-run predictions of brand business performance
The Marketing Strategy Research Journey CLOSING COMMENTS *Acknowledging Delaine Hampton, MSI Talk