Extracting ROI From The 
Engaged Customer 
A Portfolio Management Approach to CRM 
Keith Shields Laura Benard Jen Boyer 
Chief Analytics Officer Client Services Director Marketing Strategy Manager 
Magnify Analytic Solutions Magnify Analytic Solutions Ford Motor Company 
10/28/2014
Q. How related are the following two topics? 
Consumer Loan 
Management 
VS. CRM Marketing 
1
Q. How related are the following two topics? 
Consumer Loan 
Management 
VS. CRM Marketing 
a. Not At All Related 
2
Q. How related are the following two topics? 
Consumer Loan 
Management 
VS. CRM Marketing 
b. Somewhat Related 
3
Q. How related are the following two topics? 
Consumer Loan 
Management 
VS. CRM Marketing 
c. Very Related 
4
Both Consumer Loan Managers and CRM Managers… 
Start with a portfolio of customers 
Have access to enormous amounts of customer data 
Can manage their portfolio using predictive analytics 
Both are responsible for the long term value of their assets 
5
Why is this an important discussion for us? 
We can apply learnings from the Consumer Loan 
Industry to manage our customer portfolios… 
Spending on Marketing Analytics is expected 
to increase 72% over the next 3 years* 
Only 32% of marketing projects use analytics 
Most companies do not have the right talent 
to fully leverage Marketing Analytics 
“…77% of marketers surveyed believe data on 
customer purchase histories can improve 
marketing ROI, yet only 21% actually use it. 
Likewise, 88% believe behavioral data can do 
the same, but only 20% use it” 
Big Data and new tools are quickly 
changing this landscape 
*2014 February; The CMO Survey.org Highlights and Insights 
6
CONCEPTUAL NONSENSE FROM THE 
SCIENTIST…
Defining “Portfolio Management” 
A portfolio of consumer loans, not securities. 
• Portfolio Management, loosely, is the application of a set of 
analytically-driven collections and servicing techniques aimed at 
forecasting and maximizing a loan portfolio’s cash flows. 
• Quantifying credit risk and predicting future payment is at the heart 
of portfolio management. 
• Credit Risk and CRM seemingly dominate the Big Data landscape. 
Next slide… 
7
“Big Data” infects the CRM and Credit Risk 
disciplines more than almost any other… 
Why the pervasive interest in Big Data? 
• Largely to satisfy CRM and credit risk needs… 
Data: Information Week Analytics, Business Intelligence and Information Management 
Survey of 417 business technology professionals at companies using or planning to deploy 
data analytics, BI or statistical analysis software, October 2012 
CREDIT RISK 
NEEDS 
CRM NEEDS 
8
(Re)Defining CRM 
A portfolio of customers, not loans. 
• Portfolio Management CRM, loosely, is the application of a set of 
analytically-driven collections and servicing marketing techniques aimed 
at forecasting and maximizing a loan portfolio’s customers’ cash flows 
purchases. 
• Quantifying credit defection risk and predicting future payment 
purchases is at the heart of portfolio management CRM. 
• Incidentally, the Wikipedia definition of CRM is: 
• CRM is a system for managing a company’s interactions with current and future 
customers. It involves using technology to organize, automate and synchronize sales, 
marketing, customer service, and technical support. 
9
Portfolio Management and CRM… 
From an Analytics perspective, these are the same. The 
only difference lies in the target variable and predictors. 
• PD = 1 / (1+e-z), where z = A + Bx1 + Cx2 + Dx3 + … 
• Portfolio Management: 
• PD = Probability of DEFAULT 
• x1 = credit score, x2 = days past due, x3 = loan to value ratio, etc… 
• CRM: 
• PD = Probability of DEFECTION 
• x1 = prior purchases, x2 = months since last purchase, x3 = 
unfavorable tweets, etc… 
10
Others have recognized and leveraged the 
overlap… 
11 
• Auto Pre-Approval 
• Merchant Cash Advance and Small Business Loans 
• Pier-to-Pier lending 
• Student loan servicing 
• Business Rules Engines
The Portfolio Management Paradigm 
1 
4 
Managing a loan portfolio requires that we turn impaired (high 
credit risk) loans into cash-flowing bonds… 
Customer’s loan is 
rewritten for empirically-derived 
optimal amount 
CASH FLOWS 
Customer makes 
partial payment 
$A1 
$A2 
Loan impaired, 
collections calls 
ensue 
Customer pays off 
rewritten balance 
Time 
t=0 
t=1 t=3 
• The value of this “bond” (loan) is $A1/(1+i)1 + $A2/(1+i)3 
• This paradigm applies equally to CRM. The portfolios managed by CRM 
professionals are the customer bases of the companies they serve. 
12
Adopting the PM paradigm for CRM… 
1 
5 
An engaged customer is a bond. The effectiveness of our CRM 
strategies determines the yield of that bond. 
1 
5 
Customer comes 
in for service 
Customer visits 
company website 
$A1 CASH FLOWS 
$A2 
Customer signs up 
for rewards program 
Customer purchases 
a new vehicle 
Time 
t=0 
t=1 t=3 
• The value of this “bond” (customer) is $A1/(1+i)1 + $A2/(1+i)3 
• Customer Lifetime Value (CLV) models help quantify the value of customer 
behaviors and CRM tactics. The success of CRM can be measured by the 
extent to which CLV increases, irrespective of test-control results. 
13
Marketers already recognize the need to 
view their customer base as a portfolio… 
Types of Data that Marketers Worldwide Would 
Like to Add to Their Customer Data Profile 
42% 
42% 
14 
19% 
19% 
14% 
12% 
24% 
35% 
45% 
53% 
71% 
Predictive analytics around lifetime… 
Online customer profile 
Customer service feedback 
User survey and preference data 
Social media data 
Third-party demographic data 
Sales executive insights 
Finance / customer payment data 
Order history 
Analyting is better than what we… 
0% 10% 20% 30% 40% 50% 60% 70% 80% 
In-store / agent exchanges 
(1Q2013) 
Source: CMO Council and SAS 
% of respondents 
• 71% of marketers want 
“predictive analytics around 
lifetime value” added to their 
customer data profiles… 
• Lifetime value models are 
nothing more than a forecast of 
cash flows at the customer 
level… 
• Survival analysis, vintage-level 
monitoring, and other popular 
PM disciplines are a must…
But CRM trails Credit Risk / Portfolio 
Management in the adoption of Big Data…why? 
“…77% of marketers surveyed believe data on customer purchase histories can 
improve marketing ROI, yet only 21% actually use it. Likewise, 88% believe 
behavioral data can do the same, but only 20% use it” 
15 
• Regulation 
• Accountability is “fuzzy” 
• Metrics are inexact and not directly reflective of behavior. 
• Secondary markets 
• What would CRM analytics look like if marketers were 
forced to buy, sell, and “value” their customer 
portfolios? 
• Metrics are inexact and not directly reflective of behavior.
What PM practices will help our CRM? 
• Take a longitudinal view of the customer. This is the only way to get an 
accurate outlook and valuation. Implies a need for a CLV model… 
16 
• CLV = p(sale at time 1)*E($ profit from sale) / (1 + d)1 + 
p(sale at time 2)*E($ profit from sale) / (1 + d)2 + 
p(sale at time 3)*E($ profit from sale) / (1 + d)3 + … 
• Engagement is measured longitudinally; enticement is measured cross-sectionally. 
• Quantify the impact of “mix shift” on outcomes of interest. 
• Establish “regulatory-like” rigor around model validation. 
• Understand that the two share not only a brain, but also a nervous 
system. Next slide…
ENOUGH CONCEPTUAL NONSENSE 
FROM THE SCIENTIST. NOW SOME 
PRACTICAL STUFF THE MARKETER…
Does This Change the Way We Practice CRM? 
17 
We think so, especially in the following areas: 
Measuring Success 
Metrics should be more bottom-line oriented 
and exact 
Shift from basing success solely on campaign 
performance to understanding performance of 
the portfolio 
Predicting Outcomes 
Predictions should go beyond the “next 
transaction” 
All available data should be leveraged to 
proactively manage customers throughout the 
lifecycle to desired business objectives 
Influencing Behavior 
CRM becomes our “sand box” for going beyond 
understanding just correlations; to understanding 
causation as a way to change customer behavior
Predicting Outcomes 
18 
Transaction vs. Portfolio Management approach to predicting outcomes…. 
% In-Market 
Short-Term: Optimizing campaign performance to 
campaign objectives 
Segment Size 
Opportunity 
Longer-Term: Enables management of entire 
portfolio to business objectives (i.e. increasing CLV)
Influencing Behavior 
Test and learn approach will determine how we influence and change the long term 
health of our customer portfolio… 
Monitor drivers across 
the portfolio… 
Design treatments, messaging and investment 
based on customer value, individual customer 
drivers and predicted outcome 
Understand 
Drivers of 
Desired 
Outcomes 
Every CRM treatment should be analytically driven…ensuring that every CRM 
dollar spent is working to move the customer into a more valuable state 
19
Measuring Success 
Strategic 
Operational 
Tactical 
Portfolio Health 
What is the value of my customer portfolio? 
What is the mix and risk of my customer portfolio? 
Performance and Forecasting 
Do I understand both rear-ward and forward-looking performance? 
What is the aggregate impact of our CRM initiatives on improving sales? 
Dashboard and Diagnostics 
Which champion vs. challenger campaign performs best? 
Which actions influence customer outcomes both positive and negative? 
20
Thus Ends the Prepared Remarks… 
• Understand that the job of CRM is to extract repeat sales and revenue from the portfolio 
of customers. The best way to do this is make sure that customers remain engaged over 
a long period of time. 
• If a customer is a bond, then improving engagement, in effect, increases the life of the bond. 
21 
• CRM groups should measure themselves with this standard in mind. 
• Keeping customers in their “most valuable state” is a matter of advanced analytics and 
strong marketing tactics…both of which are done with an eye towards engagement. 
• The disciplines applied routinely to the management of loan portfolios are equally 
applied to CRM. Champion / Challenger tests are simply one tool in a larger toolbox. 
• Thank you for your time and attention.
“JUDGE A MAN BY HIS QUESTIONS 
RATHER THAN HIS ANSWERS.” -- 
VOLTAIRE

Magnify DMA presentation 2014

  • 1.
    Extracting ROI FromThe Engaged Customer A Portfolio Management Approach to CRM Keith Shields Laura Benard Jen Boyer Chief Analytics Officer Client Services Director Marketing Strategy Manager Magnify Analytic Solutions Magnify Analytic Solutions Ford Motor Company 10/28/2014
  • 2.
    Q. How relatedare the following two topics? Consumer Loan Management VS. CRM Marketing 1
  • 3.
    Q. How relatedare the following two topics? Consumer Loan Management VS. CRM Marketing a. Not At All Related 2
  • 4.
    Q. How relatedare the following two topics? Consumer Loan Management VS. CRM Marketing b. Somewhat Related 3
  • 5.
    Q. How relatedare the following two topics? Consumer Loan Management VS. CRM Marketing c. Very Related 4
  • 6.
    Both Consumer LoanManagers and CRM Managers… Start with a portfolio of customers Have access to enormous amounts of customer data Can manage their portfolio using predictive analytics Both are responsible for the long term value of their assets 5
  • 7.
    Why is thisan important discussion for us? We can apply learnings from the Consumer Loan Industry to manage our customer portfolios… Spending on Marketing Analytics is expected to increase 72% over the next 3 years* Only 32% of marketing projects use analytics Most companies do not have the right talent to fully leverage Marketing Analytics “…77% of marketers surveyed believe data on customer purchase histories can improve marketing ROI, yet only 21% actually use it. Likewise, 88% believe behavioral data can do the same, but only 20% use it” Big Data and new tools are quickly changing this landscape *2014 February; The CMO Survey.org Highlights and Insights 6
  • 8.
    CONCEPTUAL NONSENSE FROMTHE SCIENTIST…
  • 9.
    Defining “Portfolio Management” A portfolio of consumer loans, not securities. • Portfolio Management, loosely, is the application of a set of analytically-driven collections and servicing techniques aimed at forecasting and maximizing a loan portfolio’s cash flows. • Quantifying credit risk and predicting future payment is at the heart of portfolio management. • Credit Risk and CRM seemingly dominate the Big Data landscape. Next slide… 7
  • 10.
    “Big Data” infectsthe CRM and Credit Risk disciplines more than almost any other… Why the pervasive interest in Big Data? • Largely to satisfy CRM and credit risk needs… Data: Information Week Analytics, Business Intelligence and Information Management Survey of 417 business technology professionals at companies using or planning to deploy data analytics, BI or statistical analysis software, October 2012 CREDIT RISK NEEDS CRM NEEDS 8
  • 11.
    (Re)Defining CRM Aportfolio of customers, not loans. • Portfolio Management CRM, loosely, is the application of a set of analytically-driven collections and servicing marketing techniques aimed at forecasting and maximizing a loan portfolio’s customers’ cash flows purchases. • Quantifying credit defection risk and predicting future payment purchases is at the heart of portfolio management CRM. • Incidentally, the Wikipedia definition of CRM is: • CRM is a system for managing a company’s interactions with current and future customers. It involves using technology to organize, automate and synchronize sales, marketing, customer service, and technical support. 9
  • 12.
    Portfolio Management andCRM… From an Analytics perspective, these are the same. The only difference lies in the target variable and predictors. • PD = 1 / (1+e-z), where z = A + Bx1 + Cx2 + Dx3 + … • Portfolio Management: • PD = Probability of DEFAULT • x1 = credit score, x2 = days past due, x3 = loan to value ratio, etc… • CRM: • PD = Probability of DEFECTION • x1 = prior purchases, x2 = months since last purchase, x3 = unfavorable tweets, etc… 10
  • 13.
    Others have recognizedand leveraged the overlap… 11 • Auto Pre-Approval • Merchant Cash Advance and Small Business Loans • Pier-to-Pier lending • Student loan servicing • Business Rules Engines
  • 14.
    The Portfolio ManagementParadigm 1 4 Managing a loan portfolio requires that we turn impaired (high credit risk) loans into cash-flowing bonds… Customer’s loan is rewritten for empirically-derived optimal amount CASH FLOWS Customer makes partial payment $A1 $A2 Loan impaired, collections calls ensue Customer pays off rewritten balance Time t=0 t=1 t=3 • The value of this “bond” (loan) is $A1/(1+i)1 + $A2/(1+i)3 • This paradigm applies equally to CRM. The portfolios managed by CRM professionals are the customer bases of the companies they serve. 12
  • 15.
    Adopting the PMparadigm for CRM… 1 5 An engaged customer is a bond. The effectiveness of our CRM strategies determines the yield of that bond. 1 5 Customer comes in for service Customer visits company website $A1 CASH FLOWS $A2 Customer signs up for rewards program Customer purchases a new vehicle Time t=0 t=1 t=3 • The value of this “bond” (customer) is $A1/(1+i)1 + $A2/(1+i)3 • Customer Lifetime Value (CLV) models help quantify the value of customer behaviors and CRM tactics. The success of CRM can be measured by the extent to which CLV increases, irrespective of test-control results. 13
  • 16.
    Marketers already recognizethe need to view their customer base as a portfolio… Types of Data that Marketers Worldwide Would Like to Add to Their Customer Data Profile 42% 42% 14 19% 19% 14% 12% 24% 35% 45% 53% 71% Predictive analytics around lifetime… Online customer profile Customer service feedback User survey and preference data Social media data Third-party demographic data Sales executive insights Finance / customer payment data Order history Analyting is better than what we… 0% 10% 20% 30% 40% 50% 60% 70% 80% In-store / agent exchanges (1Q2013) Source: CMO Council and SAS % of respondents • 71% of marketers want “predictive analytics around lifetime value” added to their customer data profiles… • Lifetime value models are nothing more than a forecast of cash flows at the customer level… • Survival analysis, vintage-level monitoring, and other popular PM disciplines are a must…
  • 17.
    But CRM trailsCredit Risk / Portfolio Management in the adoption of Big Data…why? “…77% of marketers surveyed believe data on customer purchase histories can improve marketing ROI, yet only 21% actually use it. Likewise, 88% believe behavioral data can do the same, but only 20% use it” 15 • Regulation • Accountability is “fuzzy” • Metrics are inexact and not directly reflective of behavior. • Secondary markets • What would CRM analytics look like if marketers were forced to buy, sell, and “value” their customer portfolios? • Metrics are inexact and not directly reflective of behavior.
  • 18.
    What PM practiceswill help our CRM? • Take a longitudinal view of the customer. This is the only way to get an accurate outlook and valuation. Implies a need for a CLV model… 16 • CLV = p(sale at time 1)*E($ profit from sale) / (1 + d)1 + p(sale at time 2)*E($ profit from sale) / (1 + d)2 + p(sale at time 3)*E($ profit from sale) / (1 + d)3 + … • Engagement is measured longitudinally; enticement is measured cross-sectionally. • Quantify the impact of “mix shift” on outcomes of interest. • Establish “regulatory-like” rigor around model validation. • Understand that the two share not only a brain, but also a nervous system. Next slide…
  • 19.
    ENOUGH CONCEPTUAL NONSENSE FROM THE SCIENTIST. NOW SOME PRACTICAL STUFF THE MARKETER…
  • 20.
    Does This Changethe Way We Practice CRM? 17 We think so, especially in the following areas: Measuring Success Metrics should be more bottom-line oriented and exact Shift from basing success solely on campaign performance to understanding performance of the portfolio Predicting Outcomes Predictions should go beyond the “next transaction” All available data should be leveraged to proactively manage customers throughout the lifecycle to desired business objectives Influencing Behavior CRM becomes our “sand box” for going beyond understanding just correlations; to understanding causation as a way to change customer behavior
  • 21.
    Predicting Outcomes 18 Transaction vs. Portfolio Management approach to predicting outcomes…. % In-Market Short-Term: Optimizing campaign performance to campaign objectives Segment Size Opportunity Longer-Term: Enables management of entire portfolio to business objectives (i.e. increasing CLV)
  • 22.
    Influencing Behavior Testand learn approach will determine how we influence and change the long term health of our customer portfolio… Monitor drivers across the portfolio… Design treatments, messaging and investment based on customer value, individual customer drivers and predicted outcome Understand Drivers of Desired Outcomes Every CRM treatment should be analytically driven…ensuring that every CRM dollar spent is working to move the customer into a more valuable state 19
  • 23.
    Measuring Success Strategic Operational Tactical Portfolio Health What is the value of my customer portfolio? What is the mix and risk of my customer portfolio? Performance and Forecasting Do I understand both rear-ward and forward-looking performance? What is the aggregate impact of our CRM initiatives on improving sales? Dashboard and Diagnostics Which champion vs. challenger campaign performs best? Which actions influence customer outcomes both positive and negative? 20
  • 24.
    Thus Ends thePrepared Remarks… • Understand that the job of CRM is to extract repeat sales and revenue from the portfolio of customers. The best way to do this is make sure that customers remain engaged over a long period of time. • If a customer is a bond, then improving engagement, in effect, increases the life of the bond. 21 • CRM groups should measure themselves with this standard in mind. • Keeping customers in their “most valuable state” is a matter of advanced analytics and strong marketing tactics…both of which are done with an eye towards engagement. • The disciplines applied routinely to the management of loan portfolios are equally applied to CRM. Champion / Challenger tests are simply one tool in a larger toolbox. • Thank you for your time and attention.
  • 25.
    “JUDGE A MANBY HIS QUESTIONS RATHER THAN HIS ANSWERS.” -- VOLTAIRE