Customer accounting creating value with customer analytics

that the acquisition and retention of profitable customers is crucial for SBEs toidentify the fundamental elements of their business model e.g., customer base,revenues and service cost pe

SpringerBriefs in Accounting

Series editors

Peter Schuster, Schmalkalden, Germany

Robert Luther, Bristol, UK

Massimiliano Bonacchi • Paolo Perego

Customer Accounting

Creating Value with Customer Analytics

Free University of Bozen-Bolzano

Library of Congress Control Number: 2018957984

© The Author(s), under exclusive licence to Springer Nature Switzerland AG 2019

This work is subject to copyright All rights are reserved by the Publisher, whether the whole or part of the material is concerned, speci fically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed.

The use of general descriptive names, registered names, trademarks, service marks, etc in this publication does not imply, even in the absence of a speci fic statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use.

The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made The publisher remains neutral with regard to jurisdictional claims in published maps and institutional af filiations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

It is self-evident that customers are essential to business enterprises This wasalready the case when thefirst barter transaction in history was concluded What isrelatively new is the ability of many businesses now, particularly those who charge asubscription fee for their services, to track their customers, identify their preferences,customize products to people’s tastes, and learn about their experiences and satis-faction level This wealth of information derived from the footprints of customers ofInternet service providers, media and entertainmentfirms, and insurance companies,among other sectors, radically transformed corporate customer management Butthis transformation is a work in process with lots of unanswered questions for bothcorporate managers and their shareholders

That is the reason this book on customer accounting is such a welcome addition tothe literature of management, marketing, operations research, and of course account-ing The core of the book is the introduction of the highly useful concept of acompany’s lifetime value of customers, which for many enterprises is their largestand most consequential, value-creating asset The computation of customers’ value(customer equity) and the various uses of this important metric in management andcapital market investment decisions are clearly discussed in this book The manyreal-life examples provided by the authors, both experts on the subject, demonstratethe power of this new metric and make the book fun to read

Customer value and the related measures introduced and demonstrated by theauthors are particularly important to investors, given the sharp decline in theusefulness and relevance of the traditional accounting andfinancial variables used

in investment analysis In this book, both managers and investors will find newmeasures and methods to manage customers and enhance corporate value

Who will benefit from this book? Corporate executives responsible for themanagement of their customers to create corporate value and also CFOs;financialanalysts and investors striving to value business enterprises and frustrated with thetraditional, failedfinancial measures based on accounting asset and earnings; and

v

last but not least, business students, both at the undergraduate and graduate (MBA)levels, will benefit considerably from this book in finance, marketing, and account-ing courses.

Philip Bardes Professor of Accounting

and Finance

NYU Stern School of Business

New York, NY, USA

Baruch Lev

1 Introduction 1

1.1 Customer-Centricity in a Fast-Evolving Landscape 1

1.2 Motivation and Objectives of This Book 4

1.3 Theoretical Framework: Organizational Architecture 5

1.4 Outline of This Book 8

References 10

2 Customer Analytics: Definitions, Measurement and Models 13

2.1 Customer Analytics: Definitions of CP, CLV and CE 13

2.2 CLV Formulae: Sources and Variations 16

2.3 Applications of CLV in Subscription-Based Business Settings 17

2.4 CLV Scorecard and Cohort Analysis: An Application in an SBE 19

2.4.1 The CLV Scorecard as a Performance Measurement System 20

2.4.2 Benefits of CLV Scorecard 24

2.4.3 CLV Cohort Analysis: Rationale 25

2.4.4 CLV Cohort Analysis: A Practical Illustration 28

2.5 Conclusions and Implications 33

References 33

3 Customer Analytics for Internal Decision-Making and Control 37

3.1 Review of Accounting and Marketing Literature 37

3.2 Evaluation of the Literature 41

3.3 A Case Study on the Adoption of Customer Analytics 42

3.3.1 Case Background and Research Methodology 42

3.3.2 Organizational Structure 43

3.3.3 The Performance Measurement System 45

3.3.4 The Reward System 47

3.3.5 Conclusions and Implications from the Case Study 48

3.4 An Exploratory Cross-Sectional Survey on the Adoption of Customer Analytics 50

vii

3.4.1 Sample and Data Collection 53

3.4.2 Descriptive Statistics and Univariate Analysis 54

3.4.3 Multivariate Analysis 56

3.4.4 Conclusions and Implications from the Survey 58

Appendix Chapter 3: Questionnaire 59

References 64

4 Customer Equity for External Reporting and Valuation 67

4.1 Customers as the Most Valuable (Intangible) Asset 67

4.2 Customer Franchise Is Missing in IFRS/US GAAP Financial Statements: How to Value It? 68

4.3 Describing SBEs Business Model Using Customer Metrics 69

4.4 Valuing SBEs Using Publicly Disclosed Customer Metrics: A Parsimonious Model to Estimate Customer Equity 71

4.5 Customer Equity and Stock Returns: Empirical Evidence 77

4.6 Beyond GAAP: Customer Metrics Reporting 79

References 81

5 Conclusions and Trends to Look Forward 83

5.1 Looking Back and Looking Ahead 83

5.2 Linking Online with Offline Commerce 84

5.3 Enhanced Forms of Corporate Non–financial Reporting 85

5.4 The Rising Impact of Artificial Intelligence on Modeling Customer Data 86

References 87

1.1 Customer-Centricity in a Fast-Evolving Landscape

During the Nineties, the business environment was affected by technologicaladvances resulting from“combinatorial innovations” triggered by liberalization ofthe telecommunication industry and the Internet (Varian et al.2004) Those inno-vations created the basis for many of the innovative services introduced over the pastdecade, such as cell phones, satellite radio, cable TV,financial services (e.g directbanking) and internet services (games, music, entertainment, etc.) (Libai et al.2009)

At the same time, the information technology (IT) revolution introduced nary improvements in methods of collecting, storing, analyzing, and transmittinghuge amounts of information (Varian2006,2009)

extraordi-Firms realized that this presented great opportunities to invest in IT to managecustomer relationships, since data could reveal actual customer preferences ratherthan merely their intentions, making sampling unnecessary since information oncustomer behavior became available for the entire population of customers (Gupta

et al 2006) For instance, advertising models evolved from a focus on “brandawareness” to “direct and measurable” customer acquisitions (Economist2006a,b,

2007; Epstein 2007; Epstein and Yuthas 2007; French 2007) Unlike televisionadvertising, Internet advertisers paid only when a user clicked through to their web-site, gaining a reliable measurement of customer acquisition costs (Court 2005;Laffey2007; Mulhern2009)

In recent years, firms have continued witnessing a period of transformativedevelopments that emphasize the central role of customers in all industries We

© The Author(s), under exclusive licence to Springer Nature Switzerland AG 2019

M Bonacchi, P Perego, Customer Accounting, SpringerBriefs in Accounting,

https://doi.org/10.1007/978-3-030-01971-6_1

1

provide below a few examples showing customer power and the trends shaping thefuture of marketing decisions into the next decade:

• Half of the firms listed in the DAX 30 and DJIA 30 explicitly mention in theirmission statements or company strategies the notion of value creation for cus-tomers (Kumar and Reinartz2016)

• According to a 2017Forresterreport, we are now fully within the‘Age of theCustomer’, in which newly empowered customers place elevated expectations onevery interaction they have with brands

• The 2017 Salesforce report “State of the Connected Customer”, revealed that70% of consumers now believe technology has made it easier than ever to switchbrands tofind experiences that matches their expectations

• The results of a 2016 global survey by Forbes Insights showed that firms whoincreased their spending on retention in the last 1–3 years had nearly a 200%higher likelihood of increasing their market share in the last year compared tothose spending more on acquisition

• An online survey by TECH at Harvard revealed that in 2016, increasing customerexperience received the highest priority among 908 IT decision makers at globalfirms

These latest examples clearly indicate that consumers hold far more power thanever before in today’s ultracompetitive and fast evolving business landscape Thetransition from a product-centric, transaction-focused business model to a morerelationship-oriented or customer-centric view appears as a necessary condition

to sustain long-term business performance (Sheth et al.2000; Shah et al 2006;Ramani and Kumar2008) This transition necessitates a radical shift that aggres-sively relies on interaction response capacity and customer value management(Kumar et al.2008; Ramani and Kumar2008) Interaction response capacity is thedegree to which afirm can provide successful products and services by exploitingthe feedback of a specific customer At the same time, through customer valuemanagement afirm can define and dynamically measure individual customer dataand use this information as a guiding principle for tactical and strategic resourceallocation decisions

Customer-centricfirms thus understand not only what the customer values but,more importantly, the value the customer adds to their bottom line Customer-centricity implies a carefully defined and quantified customer segmentation strategy

in which afirm’s operations aim at delivering the greatest value to the best customersfor the least cost (Sheth et al.2000; Shah et al.2006; Ramani and Kumar2008; Libai

et al 2009; Fader 2012) Shah et al (2006) and Fader (2012) emphasize thatcustomer centricity is a necessary condition for twenty-first-century firms that need

to address key strategic issues (Kumar and Rajan2012; Cokins2015) such as:

• Do we push for volume or for margin with a specific customer? How manyproducts can we sell to a specific customer?

• How can we develop profitable relationships over a long time span?

• How can we identify profitable customer segments and business processes withhigher productivity?

• Can we influence our customers to alter their behavior to interact differently (andmore profitably) with us?

In Table 1.1, we summarize the main differences between product-centric andcustomer-centric orientations after a review of several sources in marketing and man-agement literature (Sheth et al.2000; Egol et al 2004; Shah et al.2006; Kumar

In this context, disruptive developments in digital technology, Internet of Things(IoT), sensor data and the social media have accelerated the shift towards customer-centricity on an unprecedented scale and pace In a short time, firms in severalindustries have started to collect very large quantities of data from their ownoperations, supply chains, production processes, and customer interactions Thescale and diversity of customer data provide Internet-based firms such asFacebook, Google, Amazon and Netflix rich new sources of business insights,allowing firms to understand and engage with customers in novel ways to bothbetter serve them and maximize profitability Beyond a basic transaction history,companies currently track marketing interactions, clicks, web or mobile navigationpatterns, and online and offline behaviors, on their own platforms or on social media.They also receive large amounts of data from connected objects owned by customers(e.g mobile phones, tablets, tracking devices) Traditional databases cannot handlesuch volumes of information and variety of formats, but this is where‘Big Data’solutions step in We are currently witnessing a shift in the breadth and depth offirms’ customer accounting systems In this book, we use the label customeranalytics to broadly denote the metrics, processes and technologies that provide

Table 1.1 Comparison of the product-centric and customer-centric approach (source: Bonacchi and Perego 2012 )

Product-centric approach Customer-centric approach Basic philosophy Sell products Serve customers

Organizational

focus

Internally focused New product development, new account development, market share growth, and customer trelations are issues for the marketing department

Externally focused Customer relationship development, pro fitability through customer loyalty Employees are customer advocates Selling approach How many customers can we sell this

firms the insight into customers necessary to deliver offers that are anticipated,relevant and timely.

Numerous examples are emerging of the potential impact of customer analytics intraditional companies: Tesco and IBM, among other largefirms, make increasing use

of Big Data to deliver contextual insights about purchase behaviors and marketingresponse Severalfirms are also spinning up new investigative computing or datascience practices rooted in artificial intelligence (AI), deep learning and other highlydynamic and multidimensional forms of advanced analytics Half a decade ago, none

of these disrupting technologies were anywhere close to being used in daily tices In the closing chapter of this book, we will point at these developments further

prac-1.2 Motivation and Objectives of This Book

An increasing number of academic papers in marketing have examined how acustomer-centric focus can provide competitive advantages and emphasized thebenefits of providing differentially tailored responses to marketing initiatives, suchthat the contribution from each customer to overall profitability is maximized (e.g.,Verhoef and Lemon2013) The marketing literature has also started to highlight theorganizational steps and barriers critical to initiate and sustain customer centricity(Shah et al.2006; Kumar et al.2008) However, there is a dearth of knowledge aboutthe business processes with which CFOs and management accountants interact andcoordinate with other CMOs and marketing managers to monitor the attraction,conversion and retention of customers through marketing campaigns and reliance oncustomer data Interested readers should refer to recent reviews of the literaturededicated to the marketing-accounting interface (Gleaves et al.2008; Roslender andWilson2008; Kraus et al.2015)

The apparent disjunction between these two core functions emerges clearly in thedevelopments of the accounting literature on customer accounting, defined as “allaccounting techniques that measure individual customer’s and/or customer seg-ments’ contributions to firm profitability” (Holm et al.2016) On one hand, account-ing textbooks seem to cover traditional techniques of customer profitability analysisand only marginally treat contemporary topics in customer value management(Gleaves et al.2008; Bates and Whittington2009) On the other hand, the academicliterature on customer accounting is still embryonic when compared to marketing,pointing at a relevant gap between current practice and theory-driven research in thisrapidly changing business area (Guilding and McManus 2002; McManus andGuilding2008) We will provide a review of this literature in subsequent chapters

In sum, whilst the volume and complexity of customer data today require cated analytic methods that go beyond traditional measurement and reporting,accounting research and accounting textbook knowledge on these topics lag behind

sophisti-In this book, we contribute tofilling this void by examining fundamental issues,challenges and opportunities that typically a CFO or a manager in the accounting &finance function would face when dealing with customer-centricity and the role of

customer analytics in extracting valuable business insights at all stages of thecustomer lifecycle To logically map and structure the various implications, wedraw upon a theoretical framework that allows an analysis of the main leversinvolved in the implementation of a customer-centric strategy Such a conceptuali-zation, labeled‘organizational architecture’, relies on research conducted in organi-zational economics and management accounting (Wruck and Jensen1994; Brickley

et al.1995; Ittner and Larcker2001; Brickley et al.2004; Brickley et al.2009) andhas the advantage of being broadly generalizable to several business contexts andindustries In the next section, we provide a definition and a few examples of thethree components of organizational architecture relevant for customer-centricity

1.3 Theoretical Framework: Organizational Architecture

The organizational architecture framework provides the infrastructure with whichbusiness processes are deployed and ensures that the organization’s core capabilitiesare realized across business processes A key issue is ensuring that decision makersnot only have the relevant (i.e accurate and useful) information required to makedecisions, but that they must also be provided with the appropriate incentives to usethat information to achieve organizational objectives Thus, the fundamental tenetbehind organizational architecture is that value creation depends on coherenceamong three primary organizational components, namely, the assignment of deci-sion rights, the choice of performance measures, and the design of compensation andincentive systems, as depicted in Fig.1.1

The extent to which top management chooses how to design an organizationalarchitecture differs greatly amongfirms Such differences are not random but vary insystematic ways with underlying characteristics of thefirms themselves Drawing onthe contingency theory of organizations in management (Brickley et al.1995,2004;Brickley et al.2009) and management accounting research (e.g Gong and Ferreira

2014), consistent relationships and alignment among the three components shouldensure the most effectivefit with a firm’s business environment and inherent strategy(Ittner and Larcker 1997; Langfield-Smith 1997; Chenhall 2003; Widener et al

2008; Lee and Yang2011; Grabner and Moers2013) Kaplan and Norton (2004)state that“unless an organization links its strategy to its governance and operationalprocesses, it won’t be able to sustain its success” Put simply, failure to properlydesign and incorporate the three levers (hence the‘three-legged stool’ label of themodel) in internal decision-making and control systems, is likely reflected in lowerorganizational performance Previous management accounting studies recognizethat these three key organizational elements are jointly determined and complemen-tary (Nagar2002; Abernethy et al.2004; Widener et al.2008)

The role of strategy is indeed a crucial part of a contingency framework, although,

as noted by Chenhall (2003: 150),“it is not an element of the context, it is a meanswhereby managers can influence the nature of the external environment, the tech-nologies of the organization, the structural arrangement, the control culture and the1.3 Theoretical Framework: Organizational Architecture 5

management control system.” The marketing literature similarly sees in centric strategies a solution to adapt to the new competitive environment character-ized by rapid changes in technology, market forces and regulation In particular,Shah et al (2006), Fader (2012) and Cokins (2015) emphasize that customercentricity is a necessary condition for twenty-first-century firms that need to addresskey strategic issues (Kumar and Rajan2012), such as:

customer-• How many products can we sell to the customer?

• How can we develop profitable relationships?

• How can we identify profitable customer segments?

Following this rationale, customer-centricfirms should deliberately design anddevelop features in their organizational architecture that differentiate them fromthose typical of traditional product-centricfirms

The performance measurement system (how a firm’s performance is tualized, tracked and evaluated) involves the choice of performance measures tocoordinate the efforts of decision makers, to provide feedback to top managementfor evaluating progress toward strategic objectives and to employees for learningpurposes A critical component of the performance measurement system for

concep-Technology

− Fast pace of change

− Big Data, AI, Social media

Markets

− Rising power of the Customer

− Diffusion of services

Allocation of decision rights

Performance measures

Incentives

Financial performance

customer-centric organizations is determining how to collect customer-related data

to provide a unified, comprehensive, and organization-wide view of a firm’s tomer base, irrespective of the products purchased or channels employed by thecustomer This entails a substantial IT-related investment commitment to set up

cus-an information infrastructure for collecting, tracking, cus-and integrating data at theindividual-customer and transaction level Jayachandran et al (2005) specified howseveral information system-related activities can be integrated and allow customer-centricfirms to successfully build a viable relationship with their customers Such anintegrated database is then made accessible to those responsible for managing thecustomer relationship to analyze past performance with the goal of understanding the

“why” behind customer behavior (Shah et al 2006) One of the reasons manyorganizations struggle to deliver value from customer data is the excessive number

of possible integration points among the number of different data management andanalysis technologies In recent years, the advent of disrupting digital technologiesand Big Data has accelerated and opened up a variety of technical solutions tomeasure customer-related performance data Severalfirms today have multiple datawarehouses, data marts, data caches, and operational data stores aimed at a timelycollection of customer information

The allocation of decision-making authority (that is, who in the organization isgiven the authority to make decisions) reflects the contention that delegation andempowering people with specific knowledge is a critical determinant of organiza-tional success A typical product-centric company that is organized around func-tional silos defined by product types is not conducive to customer centricity, as eachproduct/sales manager may end up pushing different product offerings to the samecustomer without first determining what the customer’s true needs are On thecontrary, it can be posited that a customer-centric organization has its functionalactivities integrated and aligned to successfully serve its customers Thefirst stage ofthis organizational realignment is the emergence of lateral coordinating activitiesthat aim to overcome the traditional deficiencies of products or functional silos Thismay be achieved by setting up a horizontal organization structure, in which infor-mationflows are readily shared among team members (Shah et al 2006) In thiscontext, ensuring an interface between the Marketing and the Accounting andFinance (A&F) functions becomes crucial For example, more than a dozen Fortune

1000firms, such as Coca Cola, Hershey, Intel, HP, and JD Edwards, have created aspecialized function, labeled as Chief Customer Officer, to acknowledge the impor-tance of customer-centricity-related issues in the boardroom (Shah et al.2006; Rust

et al.2009) Wells Fargo has successfully realigned its organization by creating atwo-tiered sales structure whereby a relationship manager ensures an interactionorientation (external focus) and a product specialist provides the technical input forproduct development (internal focus) In this context, the interface between Market-ing and A&F is crucial to provide decision makers with relevant information oncustomer profitability

The third element of an organizational architecture refers to the formal incentiveand compensation systems (how a firm rewards its management for success).Incentive systems seek to motivate managers and employees to be more productive,1.3 Theoretical Framework: Organizational Architecture 7

to focus on organizational objectives and to learn A broad consensus from a variety

of disciplines concludes that the presence of incentives influences behavior Withregard to customer-centric organizations, firms should include selected customermetrics among the key performance indicators that are regularly reported to the topmanagement and the board Moreover, it is essential to synchronize incentive andreward systems by linking the formal evaluation of employees with customer-centricmetrics and targets For instance, sales/account managers could be rewarded forincreasing customer equity, while relationship managers could be incentivized toextend the profitable lifetime duration of the customers For example, Texas Instru-ments is reported to have successfully introduced a reward system that includes threemarketing metrics tracking the following dimensions: marketing gains for threeconsecutive years, efficient and timely services and better understanding of cus-tomers (Kumar2008b)

In sum, a customer-centric strategy should shape firms in ways that radicallydeviate from transaction- and product-centric business models The specific archi-tecture choices in the three dimensions of organizational design likely have animpact on the profitability of the firm Incorporating several customer data sourcesinto customer analytics, properly allocating decision-rights to move quickly fromdata to decision, and aligning incentives to avoid dysfunctional triggers remainfundamentally difficult tasks contingent upon the business environment, the industryand the technological developments in which afirm operates

1.4 Outline of This Book

We acknowledge that the organizational architecture (similarly to other tional design frameworks) is an abstraction of the complex interdependencies,simultaneous choices, and feedback loops found in practice However, it provides

organiza-a useful frorganiza-amework for corganiza-ategorizing the morganiza-ain orgorganiza-anizorganiza-ationorganiza-al dimensions organiza-and ness processes involved in customer-centricfirms and the main effects thereof Inthis book, we will therefore rely upon the organizational architecture to structure ouranalysis along two lines:

busi-• The current state-of-the-art academic literature: our focus will predominantly be

on accounting studies, although we will also highlight main trends andfindings inthe marketing literature;

• Practical applications or field studies that serve the purpose of illustrating withconcrete examples and research insights how customer accounting can influenceorganizations interested to shift towards customer-centricity

We will initially point to the recent developments in the dimension of mance measurement as a foundational element of the organizational architecturerequired to pursue a specific business strategy—in our setting a customer-centricstrategy The label and contents we adopt will be customer analytics to more appro-priately convey the combination of the wide range of data sources and customer

perfor-metrics with the analytic capabilities used to engage with customers To determinethe relative analytics proficiency of an organization, MIT Sloan Management Reviewdeveloped the Analytics Core Index based on the organization’s core analyticscapabilities in:

• ingesting data (capturing, aggregating, and integrating data);

• analyzing (descriptive analytics, predictive analytics, and prescriptive analytics);

• applying insights (disseminating data insights and incorporating them into mated processes)

auto-Our aim is not to dissect every analytic capability; we will focus instead onessential features that are more relevant for the typical challenges faced by a CFOsand CMOs in developing a suitable set of customer analytics

Chapter 2 provides definitions of the most widely diffused customer metrics,namely Customer Profitability (CP), Customer Lifetime Value (CLV), CustomerEquity (CE) We refer to the marketing literature that extensively covers thesemetrics and illustrate their interrelationships We point at applications in businesssettings that have a contractual, subscription-based model and mention potentialchallenges to compute CLV in non-contractual settings To illustrate the implemen-tation and impact of customer metrics in a real-world context, we provide a casestudy focused on the computation of CLV in an Internet-based, subscription-basedcompany The case presents a simulation that applies cohort analysis in an attempt tofill the void between theoretical CLV models and its implementation in practice Themain rationale is to provide CFOs and CMOs a better understanding of new andlatent customer preferences in a typical subscription-based business model bydirectly observing the customer’s purchase behavior and subsequently linking thisdata to estimate CLV andfirm performance

In Chap 3 we offer a critical evaluation of the literature in accounting thatexamined the role of customer metrics in internal decision-making and controlpurposes We draw on the relationships theorized in the organizational architectureoutlined in Chap.1to structure our selective review and emphasize key critical gaps

in our knowledge, especially vis-à-vis extant developments in the marketing ture The chapter then presents two empirical studies aimed at generating insights onthe adoption of customer metrics for internal decision-making and control purposes.The first study is a qualitative case study conducted within a subscription-basedenterprise (SBE) The second study reports a survey about the diffusion of customermetrics in a sample of SBEs In combination, the empirical evidence highlightsrelevant take-away points and current challenges about the actual use of customermetrics in performance measurement and management control systems

litera-Chapter4reiterates a common critique about currentfinancial accounting models(e.g IFRS/US GAAP), namely that they cannot capture Customer Franchise as keyvalue creator intangible asset We approach this issue by characterizing the businessmodel of subscription-based enterprises (SBEs) that offer a for-fee-per-period access

to products or services Specifically, we show how to aggregate publicly availabledata into a measure of afirm’s Customer Equity value, which incorporates the majorvalue drivers of SBEs, and empirically examine its properties We build on the idea

that the acquisition and retention of profitable customers is crucial for SBEs toidentify the fundamental elements of their business model (e.g., customer base,revenues and service cost per user, and customer turnover) We further argue thatcompanies should disclose in the Management Discussion and Analysis (MD&A)section of their annual report a set of customer metrics useful to investors, such asnew subscriber acquisitions, revenue per subscriber, customer dropouts, and cost ofcustomer acquisition.

Chapter5concludes the book and provides a glimpse on managerial, ical and institutional trends that will likely affect the way customer metrics will bedeployed to create business value in the next decade

Bonacchi, M., & Perego, P (2012) Measuring and managing customer lifetime value: A CLV scorecard and cohort analysis in a subscription-based enterprise Management Accounting Quarterly, 14(1), 27 –39.

Brickley, J., Smith, C., & Zimmerman, J (1995) The economics of organizational architecture Journal of Applied Corporate Finance, 8(2), 19 –31.

Brickley, J A., Smith, C W., & Zimmerman, J L (2004) Managerial economics and tional architecture Boston, MA: McGraw-Hill/Irwin.

organiza-Brickley, J., Smith, C., Zimmerman, J., & Willett, J (2009) Using organizational architecture to lead change Journal of Applied Corporate Finance, 21(2), 58 –66.

Chenhall, R H (2003) Management control systems design within its organizational context: Findings from contingency-based research and directions for the future Accounting Organiza- tions and Society, 28(2 –3), 127–168.

Cokins, G (2015, February) Measuring and managing customer pro fitability Strategic Finance,

23 –29.

Court, D C (2005) Boosting returns on marketing investment McKinsey Quarterly, (2), 36 –47 Economist (2006a) E-commerce The Economist.

Economist (2006b, July 6) The ultimate marketing machine The Economist.

Economist (2007) Word of mouse The Economist, 77 –78.

Egol, M., Hyde, P., Ribeiro, F., & Tipping, A (2004) The customer-centric organization: From pushing products to winning customers McLean, VA: Booz Allen Hamilton.

Epstein, M J (2007) Evaluating the effectiveness of internet marketing initiatives In Management accounting guideline Hamilton: The Society of Management Accountants of Canada, The American Institute of Certi fied Public Accountants and The Chartered Institute of Management Accountants.

Epstein, M J., & Yuthas, K (2007) Managing customer value In Management accounting guideline Hamilton: The Society of Management Accountants of Canada, The American Institute of Certi fied Public Accountants and The Chartered Institute of Management Accountants.

Fader, P (2012) Customer centricity focus on the right customers for strategic advantage Philadelphia, PA: Wharton Digital Press.

French, T D (2007) Confronting proliferation in online media: An interview with Yahoo!’s senior marketer McKinsey Quarterly, (3), 18 –27.

Gleaves, R., Burton, J., Kitshoff, J., Bates, K., & Whittington, M (2008) Accounting is from Mars, marketing is from Venus: Establishing common ground for the concept of customer pro fitabil- ity Journal of Marketing Management, 24(7 –8), 825–845.

Gong, M Z., & Ferreira, A (2014) Does consistency in management control systems design choices in fluence firm performance? An empirical analysis Accounting and Business Research, 44(5), 497 –522.

Grabner, I., & Moers, F (2013) Management control as a system or a package? Conceptual and empirical issues Accounting, Organizations and Society, 38(6 –7), 407–419.

Guilding, C., & McManus, L (2002) The incidence, perceived merit and antecedents of customer accounting: An exploratory note Accounting, Organizations and Society, 27(1 –2), 45–59 Gupta, S., Hanssens, D., Hardie, B., & Kahn, W (2006) Modeling customer lifetime value Journal

Ittner, C D., & Larcker, D F (2001) Assessing empirical research in managerial accounting:

A value-based management perspective Journal of Accounting and Economics, 32, 349 –410 Jayachandran, S., Sharma, S., Kaufman, P., & Raman, P (2005) The role of relational information processes and technology use in customer relationship management Journal of Marketing, 69(4), 177 –192.

Kaplan, R S., & Norton, D P (2004) Strategy maps Boston: Harvard Business School Press Kraus, K., Håkansson, H., & Lind, J (2015) The marketing-accounting interface – Problems and opportunities Industrial Marketing Management, 46, 3 –10.

Kumar, V (2008a) Customer lifetime value – The path to profitability Foundations and Trends in Marketing, 2(1), 1 –96.

Kumar, V (2008b) Managing customers for pro fit: Strategies to increase profits and build loyalty Upper Saddle River, N.J.: Wharton School.

Kumar, V., & Rajan, B (2012) Handbook of marketing strategy Cheltenham: Edward Elgar Kumar, V., & Reinartz, W (2016) Creating enduring customer value Journal of Marketing, 80(6),

McManus, L., & Guilding, C (2008) Exploring the potential of customer accounting: A synthesis

of the accounting and marketing literatures Journal of Marketing Management, 24, 771 –795 Mulhern, F (2009) Integrated marketing communications: From media channels to digital con- nectivity Journal of Marketing Communications, 15(2), 85 –101.

Nagar, V (2002) Delegation and incentive compensation The Accounting Review, 77(2),

379 –395.

Ramani, G., & Kumar, V (2008) Interaction orientation and firm performance Journal of Marketing, 72(1), 27.

Roslender, R., & Wilson, R M S (2008) The marketing/accounting synergy: A final word but certainly not the last word Journal of Marketing Management, 24(7 –8), 865–876.

Rust, R T., Moorman, C., & Bhalla, G (2009) Rethinking marketing Harvard Business Review, 88(1), 94 –101.

Ryals, L (2008) Managing customers pro fitably Chichester: John Wiley & Sons.

Shah, D., Rust, R T., Parasuraman, A., Staelin, R., & Day, G S (2006) The path to customer centricity Journal of Service Research, 9(2), 113 –124.

Sheth, J N., Sisodia, R S., & Sharma, A (2000) The antecedents and consequences of centric marketing Journal of the Academy of Marketing Science, 28(1), 55 –66.

customer-Varian, H R (2006) The economics of internet search Rivista di Politica Economica, (IX –X),

177 –191.

Varian, H R (2009) Hal Varian on how the web challenges managers McKinsey Quarterly Varian, H R., Farrell, J., & Shapiro, C (2004) The economics of information technology: An introduction Cambridge: Cambridge University Press.

Verhoef, P C., & Lemon, K N (2013) Successful customer value management: Key lessons and emerging trends European Management Journal, 31(1), 1 –15.

Widener, S K., Shackell, M B., & Demers, E A (2008) The juxtaposition of social surveillance controls with traditional organizational design components Contemporary Accounting Research, 25(2), 605 –638.

Wruck, K H., & Jensen, M C (1994) Science, speci fic knowledge, and total quality management Journal of Accounting and Economics, 18(3), 247 –287.

Chapter 2

Customer Analytics: De finitions,

Measurement and Models

The world ’s most valuable resource is no longer oil, but data.

—The Economist

2.1 Customer Analytics: De finitions of CP, CLV and CE

In a recent discussion on the future of management accounting, Cokins (2013,2014)pointed out that cost accounting techniques like Activity-Based Costing wereconceived as causal cost tracing approaches to manage the complexity caused byincreasingly diverse types of products, services, channels and customers He labelledthe period from 1980 to date as the‘consumer era’ and suggested moving forwardinto the predictive analytics era, with a shift in emphasis from a backward-looking toforward-looking perspective of strategy and operations Cokins (2013: 25) identifiedthe expansion from product to channel and customer profitability analysis and calledfor management accounting to support the sales and marketing function tofind “thebest types of customer to retain, grow, win back and acquire” in order to maximizeshareholder value Consistently, with such a call to expand the toolkit of traditionalcost accounting techniques, a survey by Deloitte in 2016 found that more than half ofresponding North American CFOs, broadly speaking, were investing substantially in(or were planning to invest in) customer analytics, withfinance/accounting analyticsrunning a close second in terms of priority

The central tenet behind any performance measurement system is the type andsophistication of tailored business performance metrics or indicators that allowmanagers to gauge a firm’s performance against targets Literature reviews byKumar and George (2007), Villanueva and Hanssens (2007), Kumar (2008a) andPetersen et al (2009) provided exhaustive coverage of a new generation ofcustomer-metrics in the marketing literature Three core marketing-related indicatorshave been crucial in ensuring the shift towards a customer-centric strategy:

© The Author(s), under exclusive licence to Springer Nature Switzerland AG 2019

M Bonacchi, P Perego, Customer Accounting, SpringerBriefs in Accounting,

https://doi.org/10.1007/978-3-030-01971-6_2

13

Customer Profitability (CP) is defined as the difference between the revenueearned from, and the cost associated with, a customer relationship during a specifiedperiod (Smith 1993; Smith and Dikolli 1995; Foster et al 1996) This metric isusually gauged in one accounting period (e.g monthly, quarterly, half yearly and/oryearly) in which all revenues and costs have to be traced or allocated to customers.

CP belongs to the necessary toolkit that helps to make decisions about: (a) whichcustomers to select for targeting; (b) determining the level of resources to beallocated to the selected customers; and (c) selecting the customers to be nurtured

to increase future profitability (Kumar2008a)

Customer Lifetime Value (CLV) in its classical definition is the value of futurecashflow attributed to a single customer or a group of customers, discounted usingthe average cost of capital of thefirm (Kumar2008a) It is a leading informativeindicator that drives customer profitability (Kumar and Rajan2009) CLV can also

be defined in terms of profit instead of cash flow (see, Gupta and Lehmann2005) If

we assume that cashflow equals profit, CP becomes a special case of CLV with thelifetime period set at one accounting period (Gleaves et al.2008) CLV is measuredusing three main components, namely customer retention rate, margin per customer,and cost of customer acquisition CLV is a pivotal metric that is useful both forcustomer profitability analysis and in valuing companies (cf Chaps.3and4of thisbook) Customer profitability is positively associated with the forward-lookingperspective offered by CLV (Kumar and Bharath2009), in particular, when afirmhas to decide which customers to acquire/retain because CLV is the upper limit

of what one should be willing to spend to acquire/retain a customer unless one wants

to lose money CLV allows assessing which customers to nurture, with the lying tenet that management should focus on customers with high CLV Finally, theincorporation of CLV in decision-making should improve resource allocation, withmarketing resources that should strive to maximize CLV Similarly, equity valuationwill benefit because CLV offers the algorithm that helps to estimate one of the mostimportant assets of a company: the value of its customer base In fact, CLV provides

under-a vunder-aluunder-ation model thunder-at under-allows understunder-anding of the mechunder-anisms by which individuunder-alcustomer metrics (i.e ARPU, churn, cost of customer acquisition) affect afirm’ssales/earnings, and ultimately its stock return (Bonacchi et al.2015) We elaboratefurther on this topic in Chap.4of this book

Finally, Customer Equity (CE) is a combination of a firm’s current customerassets and the value of thefirm’s potential customer assets (Villanueva and Hanssens

2007) CE is defined as the sum of the CLV of all a firms’ existing and potentialcustomers In other words, CLV is a disaggregate measure of customer profitability,

while CE is an aggregate measure CE is an intangible asset of thefirm influenced

by the ability to acquire, retain, and increase the customer base (Gupta et al.2004;Kumar and Shah2009; Bonacchi et al.2015)

In sum, the key distinctions between these three concepts, which all measurecustomer value, relate to the timescale under consideration (1 year, multiple years),and to whether the analysis refers to one or all of afirm’s customers For a visualrepresentation of the inter-relation among Customer Profitability, Customer LifetimeValue, and Customer Equity, refer to Fig.2.1and Gleaves et al (2008) For the sake

of completeness, thefigure also shows the operating profit that, under the tion that all costs have been traced to customers, is the sum of the customerprofitability from all customers the firm has served within a single accounting period.According to past reviews (Kumar and George2007; Villanueva and Hanssens

assump-2007; Kumar2008a; Petersen et al.2009), the literature on CLV and other customermetrics in mainstream marketing research presently provides a rather consolidatedstream of research concerned with the development and refinement of modellingapproaches in various business settings

Thefirst modelling stream attempts to use deterministic equations in which someinputs are entered into the equation in order to calculate CLV (for a review of thesemodels see Berger and Nasr (1998)) More recently, in order to control for someendogenous parameters, researchers have proposed stochastic models to estimate

CE Inherent in all these models which try to value the long-runfinancial tion of a customer, is the expected length of the relationship The most interestingare statistical models used to predict the probability of churn (or retention) (seeVillanueva and Hanssens (2007) for a review of these models) Some researchershave also developed a parsimonious model in which the parameters can be easilyobtained, even in Microsoft Excel (Fader and Hardie2007b)

contribu-With regards to practitioner’s literature that contains applications of CLV, initialevidence is currently available in recent books such as Gupta and Lehmann (2005),Kumar (2008b) and Ryals (2008) Case studies written with pedagogical purposes

Operat ing prof it Customer Equity

Customer Prof itability

Customer Lifet ime Value

All customers

A single customer

Current accounng All Future

Fig 2.1 Classi fication of

customer metrics

2.1 Customer Analytics: De finitions of CP, CLV and CE 15

about best-practice CLV techniques are also emerging (Ofek 2002; Asis andNarayanan2007) Bendle and Bagga (2017) provide an exhaustive list of relevantcases, notes and teaching materials on CLV The next paragraph outlines theformulae applicable to compute CLV.

2.2 CLV Formulae: Sources and Variations

For exhaustive reviews of CLV models and their underlying logic, we invite thereader to refer to Jain and Singh (2010), Ascarza et al (2017) and Kumar (2018) asexcellent literature reviews of the marketing literature The definitions available onhow to compute CLV vary depending on underlying assumptions and differentnotations (Fader and Hardie 2012; Bendle and Bagga 2017) A quite commonlyused definition of CLV is the one provided by Rust et al (2009):“The customerlifetime value metric evaluates the future profits generated from a customer, properlydiscounted to reflect the time value of money” Despite the variation and at timesinconsistency across definitions, the rationale behind CLV computation thereforeresembles the Net Present Value infinance, where a constant series of cash flowsover time is discounted to take into account the time value of money (d ) Mostcommon CLV definitions therefore assume the following equation, with m the(average) contribution margin generated from a customer (or customer segment/channel) in a year or other period (cf Steenburgh and Avery2017)

A fundamental element of any CLV computation refers to the churn rate (r),

defined as the percentage of customers who end their relationship with the company

in a given period The churn rate is typically defined at the segment level, and it isimplicitly assumed that all individuals in that segment have the same probability ofending the relationship with thefirm In each subsequent period, the probability that

a customer leaves is modelled as a survival probability function that decreasesover time along the entire lifetime of the customer The series of survival probabil-ities thus determine the expected cashflows (proxied by the periodic contributionmargin) in a given period If we sum the discounted expected contribution marginsover a customer’s lifetime, for the properties of infinite geometric series we obtain asimplified version of the CLV formula that nevertheless may differ depending ontwo factors:

1 whether we include the customer’s first payment in the calculation;

2 whether the net cashflow associated with each period is “booked” at the ning or at the end of the period

begin-Different assumptions of these two factors explain the slight variation acrossavailable CLV models as conveniently summarized in Fader and Hardie (2012).Moreover, Jain and Singh (2010) emphasize that“the context of CLV measure-ment plays a key role in the methods proposed for measuring CLV and the issues thatbecome important both from a modeling point of view and the managerial point

of view By context, we mean the context of the customer-firm relationship thatgenerated the data to be used for estimating CLV From a modeling perspective, thecontext defines the data available to estimate a CLV model, and from a managerialperspective, the context defines the issues that become important in managingcustomer profitability.” Jain and Singh (2010) outlined several factors that mighthave an effect on CLV modeling and forecast Among them, the costs associatedwith the acquisition and retention of customer segments are of particular importancefor CFOs and management accountants

The next paragraph illustrates how the marketing literature carefully distinguishesdifferent business contexts and develops alternative statistical models to measureCLV in an accurate way

2.3 Applications of CLV in Subscription-Based Business

Settings

Fader and Hardie (2009) and Jain and Singh (2010) referred to different contextsthat fundamentally impact how to define and model CLV A classification ofcustomer bases usually distinguishes two dimensions: a) opportunities for trans-actions (continuous versus discrete) and b) type of relationship with customers(non-contractual versus contractual) The contractual or subscription-based businessmodel refers to the case in which a customer pays a subscription fee to have access tothefirm’s products or services The model was pioneered bymagazinesand news-paper publishers but is now used by a growing number of businesses Rather thanselling products individually, a subscription-basedfirm sells periodic (e.g monthly,yearly, or seasonal) use or access to its products or services The number and variety

of subscription-based enterprises (SBE) is fast-growing (Economist2009) tries based on the subscription model includetelephonecompanies,cable televisionproviders,cell phonecompanies,internet providers,pay-tvchannels, software andbusiness solution providers,financial service firms, as well as the online versions oftraditionalnewspapersand magazinepublishers Additions include online storage,photo sharing, social networking, and online games

Indus-The defining characteristic of an SBE is that the acquisition and departure of acustomer is observed (unlike the case of brick-and-mortar retailers, for instance) Infact, because a subscription typically involves a contractual agreement, the vendor2.3 Applications of CLV in Subscription-Based Business Settings 17

knows at any point in time the number of active customers We therefore contendthat the notions of retention and the churn rate of customers are at the heart of anycontractual or subscription-oriented business model.1The retention rate for period

t (rt) is defined as the proportion of customers active at the end of period t–1 and stillactive at the end of period t The churn rate for a given period is defined as theproportion of customers active at the end of period t–1 but who dropped out in period

t It is easy to perform this type of customer data tracking and analysis for SBEsbecause these companies always know who are still active customers and whorecently churned

Recent years have seen a dramatic increase in SBEs to offer services, such asmobile telephone, cable television, and e-banking In these types of companies, acustomer pays a subscription fee to have access to the firm’s products/services.The growth of the Internet has broadened the provision of innovative services in

a contractual setting (such as music, games, movies and e-books) and a wave ofsuccessful web start-ups (such as FriendFinder, HomeAway, LinkedIn, Pandora,Skype, Zynga) In addition, Internet SBEs are currently able to gather a massiveamount of aggregate and granular data about customer characteristics and prefer-ences through the analysis of repeated purchasing behavior transactions (Varian

Managers need to gain a more nuanced understanding of the strategic,financial,and operational implications of a subscription-based model in order to answer thecritical question: What does it take for SBEs to succeed? As highlighted in Kumarand Rajan (2009), managers need performance measurement reports that are able toconvey information useful to diagnose the health of their business and to assist them

in making strategic and tactical decisions such as:

• Which type of actual customer or future prospect should a firm retain, grow oracquire?

• How much should a firm invest to retain, grow and acquire customers?

• Which advertising channels are more effective and efficient?

• What is the value of the customer base (i.e the most important asset for thesekinds of companies)?

Stemming from several marketing studies, Kumar and Reinartz (2016) explained

a widely applicable approach useful to address these key dilemmas They outlined

a series of steps starting from the definition of a CLV model with available data,realizing the need for a forward-looking metric as allegedly the most criticaldecision Once this step has been achieved, the following activities only strengthen

a company’s position in enhancing a customer-centric strategy

Despite the fact that the beneficial effects of a customer-centric approach arewidely emphasized in the management accounting academic and practitioner liter-ature—see for instance a recent IMA Statement on Management Accounting oncustomer profitability management (2010) anecdotal evidence suggests that not

1 Customer churn is the standard term used among SBEs.

many firms practice such a strategy in a systematic and effective manner Wecontend that one of the reasons may be a lack of empirical evidence in academicresearch in management accounting that could generate practical guidance on how tomeasure CLV In the next paragraph, we present a real-world application of CLVmeasurement that could be useful to adopt in an SBE setting.

2.4 CLV Scorecard and Cohort Analysis: An Application

in an SBE2

In this paragraph, we focus our attention on two tools—the CLV scorecard andcohort analysis—that CFOs, management accountants and marketing managers insubscription-based enterprises can potentially use to have a better understanding ofthe way they measure and manage customer profitability We first provide somebackground information about the case company in which the tools are adopted Wethen illustrate thefirm’s performance measurement system and discuss how the CLVparadigm can be successfully implemented

Our research site, Company.net (the firm’s real name has been disguised forconfidentiality purposes), is a typical example of the fast-growing subscription-based business model in which a customer pays a subscription fee to receive aspecified number of downloads of content, namely ringtones and music MP3s.Company.net has contracts with all major record labels that supply content throughdownloads, it has agreements with the most important carriers that deliver contentonline and it bills customers on their mobile accounts Customers are contacted andacquired predominantly through paid search advertisements, using keywords such

as‘free-ringtones’ or ‘free music’.3

The industry value chain unfolds through a series

of activities illustrated in Fig.2.2 They are labeled as follows: content origination;service management; marketing and display; network delivery; customer relation-ship management; and billing

Company.net follows a typical SBE strategy which can be synthetized as follows:(1) Acquire new customers through aggressive marketing campaigns and marketingtools aimed at customer acquisition to expand its customer base;

2 This section is based on and include excerpts from pp 27 –39 of Bonacchi and Perego ( 2012 ).

3 Paid search advertising entails advertisers competing for top listing positions through bidding in ongoing auctions and then paying when users click on their advertisements, making paid search a flexible and accountable form of advertising (Laffey 2007 ) Pay per click ads are one of the most cost effective advertisement tools These are the ‘sponsored ads’ that are displayed at the top and right of the search results on Google, Facebook, Yahoo, Bing and similar search websites Payment

of these ads occurs whenever someone actually clicks on an ad, not when it gets displayed The cost per click can range anywhere from a few cents to several dollars, depending on the type of industry and keywords.

2.4 CLV Scorecard and Cohort Analysis: An Application in an SBE 19

(2) Retain existing customers by measuring the lifetime of users and estimate theirvalue with techniques that tend to stimulate user retention and minimize churnrate;

(3) Grow: after an initial period in which the user base is built and the churn rate

is stabilized, the focus shifts towards planning for an organic growth of thecustomer base, and defining the new target of customer acquisition for eachperiod (e.g month or quarter) in order to balance the churn rate and achieve thetarget growth rate

2.4.1 The CLV Scorecard as a Performance Measurement

Content

origination

Service management

Network used

to distribute VAS/music to end users (Mobile operators)

Content

Owners

Mobile Content &

Service providers

Media companies TelcosCompany.net

Revenues Cost of delivery

TV, print or online networks

Fig 2.2 Entertainment industry value chain of Company.net Source: Bonacchi and Perego ( 2012 )

including existing and future customers It is the most important asset for SBE,

influenced by the ability to acquire, retain, and increase the customer base Whereasthe marketing literature provides several methods of computing CE, we refer to twometrics that can be leveraged when evaluating the expected profitability of a firm’scustomer base:

a) Current Customer Equity (CEcur): the sum of the future profit margins ated from the customers that have already been acquired by the end of the period(Villanueva and Hanssens2007)

gener-b) Total Customer Equity (CEtot): the sum of the future profit margins generatedfrom current (CEcur) and future (CEfut) customers of the firm (Hogan et al

2002a; Kumar and Shah2009).4

From a management accounting perspective, the challenge is not only to measureCLV (and thus CE as a summation of each customer’s CLV), but especially tomanage its drivers The linkages among such drivers is crucial since the variables

of the CLV formula are interdependent rather than independent Whenever agers’ attention focus towards one direction, it becomes more difficult to ensure thatthe other metric moves in the same direction and pace The CLV scorecard is amanagerial tool that includes the interaction among CLV and its drivers as its keyattribute (Bonacchi et al.2008) Figure2.3exhibits the CLV Scorecard that identifies

man-Customer Lifetime

Value

(CLV)

CONTRIBUTION MARGIN

COST OF ACQUISITION (CoA)

ARPU

Cost of service

Attraction Conversion Cost of contact (CPC, CPA) feedback

LIFETIME

Instant churn Historic churn

LIFETIME VALUE (LTV)

Lifetime Value (LTV) and Cost of customer Acquisition (CoA), two essential CLVdrivers for Company.net.

CoA is a pretty straightforward metric, which in this setting can be estimated

in a very precise way for every single customer Company.net uses search engineadvertising that measures Cost per Click (CPC) and Cost per Acquisition (CPA)metrics rather than less traceable banner advertising (measured by Cost per impres-sion model, aka CPM), or traditional channels advertising such as magazine and TVads This evolution in advertising also implies innovative marketing strategies thathave shown new challenges in measuring their effectiveness from the early stages

of the Internet.5Since CoA is influenced by the type of marketing policies, the firstnecessary step for managerial actions implies an accurate analysis of the impact ofsuch policies on attraction and conversion rate Examples of possible indicatorsconsist of metrics that capture better targeting, better advertising, better landingpages, optimization of the flow through to checkout, more and better paymentoptions, and so on

The examination of LTV requires a sophisticated analysis of its main drivers(i.e Lifetime, Margin and YIELD) Thefirm needs to carefully examine the relation-ships between the lifetime value, synthetized by the following equation6:

Lifetime Value¼ Lifetime  Margin  YIELDwhere:

1 Lifetime is the period during which a customer remains with afirm This metric is

a function of the rate of attrition (cancellations/average users per period) over aperiod of time that subscriber-based customers‘churn out’ (unsubscribe) from thecustomer base Churn is a proxy of the customer satisfaction associated to thegoods and service provided The underlying rationale captures a fairly simplerelationship between churn per month and the number of months customers staywith the company (i.e ratio 1/churn) Thus, a 2% churn means 1/.02, or 50 monthsaverage customer duration To gain a good grasp of the churn metric, Company.net applies two approaches:

a Historic churn: the number of subscribers cancelling during the period (day,week, month) having initiated their subscription before that period, i.e join

5 Advertisers can now set a bid they are willing to pay to reach a certain typology of Facebook users These are audiences Facebook can reliably deliver thanks to demographic data collected among its users, while other ad exchanges might have to guess or infer about who fits an advertiser’s desired demographic For more details on this topic, please refer to: http://techcrunch.com/2012/09/18/ facebook-mobile-ad-network/

6 In theory, the issue of the time value of money must also be considered, usually discounted in the CLV formula, in order to get an estimate of future pro fits Discounting may be very appropriate in some businesses, particularly for firms operating in markets with a very long-cycle, high ticket retail and B2B However, it is believed that the discounted practice may confuse the analysis in a B2C where the environment is very dynamic and it is better to adopt a simpli fied approach.

date< quit date This ratio is computed as follows: subscribers that quit inperiod N/subscribers alive at the end of month N + 1 Historic churn is thefigure used to transform churn in lifetime;

b Instant churn: number of subscribers cancelling the service during the period(day, week, month) in which they initiated their subscription, i.e joindate¼ quit date Gross Addition is the number of new subscribers duringthe specified period In other words, the following ratio is calculated: sub-scribers that quit in period N/gross addition period N This metric is devised tojudge the quality and effectiveness of specific marketing and advertisingdecisions

2 Margin is the contribution margin per customer and equals Average Revenue perUsers (ARPU) less cost of service per customer For instance, a subscription to amobile service (i.e ringtones) of $10 and a service cost per customer for contentdelivery of $4 would generate a margin of $6

3 YIELD is defined as the ratio between subscribers successfully billed andcustomer candidates to be billed For instance, given a total number of 1000candidates of which 550 actually pay (for example because the sim-card of asubscriber might be empty), would generate a YIELD of 55%

Monitoring the abovementioned metrics is beneficial to evaluate strategic andtactical marketing choices—such as different advertising campaigns, the launch

of a new product or a traditional product in a new country—cross-sell and upsellcampaigns to increase ARPU, improve margin, and so on Company.net reportsinternally on LTV driver with a typical report that focuses on a product/service(music, ring tones, and other value-added services) providing the following infor-mation: country, subscriptionflight (i.e the scheduling of advertising for a period oftime), telecom operator used to deliver the service (refer to Table2.1) Comparisons

at country-level allow evaluating the success of a product already launched in othercountries Linking the type of marketing campaign to the acquired customers’ churn

is a distinctive analysis conducted to gauge the effectiveness of a subscriptioncampaign on a service/product For example, a customer acquired using GoogleAdWords could have a different churn rate than a customer acquired through Yahoo!

My Display Ads, or Facebook Furthermore, customers acquired from a specificmobile operator (e.g AT&T, Vodafone, or T-Mobile) could have different attitudes

to churn and a diverse YIELD profile

Table 2.1 Analysis of churn data: example of Company.net ’s internal report

Dimension

Metric Instant churn Historic churn Yield Margin Country

Subscription flight

Telco

Source: Bonacchi and Perego ( 2012 )

2.4 CLV Scorecard and Cohort Analysis: An Application in an SBE 23

These drivers can be considered leading, forward-looking indicators and must bemonitored on a daily basis to verify that the company’s results fit with the strategicobjectives of the business Whereas company presentations often display animprovement of all customer metrics overtime, however, this is unlikely to playout in reality For instance, if a firm attempts to raise ARPU (price), it willpredictably increase churn Similarly, whether a firm intended to grow faster byspending more on marketing, CoA would likely rise as inevitable consequence.Churn may increase as well, because a more aggressive marketing campaign willlikely capture customers of a lower quality.

In sum, the availability of accurate customer data and the peculiar nature of thebusiness examined (in which all transactions are made online and the logfiles ofeach transaction are constantly and accurately recorded) represents a prerequisite for

a timely assessment and effective management control of these metrics in order toassess afirm’s marketing strategy

2.4.2 Benefits of CLV Scorecard

The main benefit associated with the CLV scorecard is the opportunity to analyzepast customer behavior and understand the drivers of CLV For example, it may bevaluable to study the differences between CLV or COA of an average customeracquired through Facebook or Google ad campaigns In other words, the analysis ofthe CLV drivers allows managers to test cause-and-effect relationships betweenmanagerial actions and outcomes of such actions in terms of both customer metrics(Churn, CoA, Margin, Yield) andfinancial results (CLV and CE)

In the context of this case setting, we observed that Company.net regularlymonitors the following objects or units of analysis:

• specific marketing campaigns in place In fact, different channels contribute withdifferent results in terms of CLV For example, customer acquired throughGoogle or Facebook could have a tendency to churn differently than thosecoming from Yahoo or Bing The analysis should exploit the retention behavior(instant and historic churn) of customers acquired through these two channels inorder to understand which lifetime those users project

• different countries in case the same service is offered internationally Thisinformation is quite useful when a service is introduced in a new country, sincedata gathered in a similar country provides an appropriate benchmark Forexample, the pay-back of marketing investments in Brazil could be a goodproxy to infer what would happen if the firm invested by offering the sameservices in Chile

• different time frames to monitor different consumer behavior In this way, a daily

or hourlyfine-tuning of a marketing campaign should be performed, for example,

by changing the bid price on the keyword advertising.7In the paid search model(Laffey 2007), monitoring the behavior of click-throughs via paid search isessential, because it provides a precise measurement of the success of theadvertising method in terms of achieving the objectives set forth Data collectedfrom such tracking should then be fed back into the process to make performancereviews more effective It cannot be excluded, for example, that poor qualityprospects are attracted (perhaps through the use of the wrong keywords), or thatclicks from some sources work better than others Furthermore, monitoring can

be used to ensure that afirm is not excessively paying for clicks; for instance,being in second or third position may generate as much business as being infirstposition, where the listing position is determined by how much an advertiser isprepared to pay for a keyword or phrase

• different customer characteristics Although top Company.net managers areinterested in computing the lifetime value of their customers, they are similarlykeen on identifying the drivers of a profitable duration in their customer-firmrelationships (Reinartz and Kumar2003) More specifically, the telecommunica-tion carrier can be considered a proxy of income Moreover, a longer subscriptionlength is a proxy of customer satisfaction, since it is well known in the literaturethat the longer a subscriber remains with the company, the lower the probabilitythey are going to churn (Gupta and Lehmann 2005; Fader and Hardie 2007a,

2010) Conceptually, the following can be thus formally posited:

Profitable lifetime duration

¼ f telephone operator; length of subscription; advertising campaignð Þ

2.4.3 CLV Cohort Analysis: Rationale

Churn rate is the main driver of customer lifetime and churn data are crucial to judgethe ex-post success of a customer acquisition campaign The marketing literaturehowever raises serious concerns about the typical approach of projecting a historicchurn rate into the future to infer the lifetime of the acquired customer and then theirprofit stream (i.e LTV) Gupta et al (2004) demonstrate that the widespread methodfor the conversion of retention rate to expected lifetime (1/churn rate) and then thecalculation of present value over thatfinite time period overestimates lifetime value.Fader and Hardie (2007a, 2010) showed that the retention rate (defined as theopposite of the churn rate) is an increasing function of time Therefore, the longersubscribers are with the company, the lower the probability that they are going to

7 When a user searches for a speci fic keyword, the order of the results the user obtains is determined

by current bids in the auction Payment is made by advertisers each time a user searches for a term and then clicks on their link.

2.4 CLV Scorecard and Cohort Analysis: An Application in an SBE 25

churn Researchers account for two main reasons that explain this phenomena(Blattberg et al.2008) As the customer uses the product or services, he/she increasespreference or satisfaction or is locked-in and that causes the decreasing churn rate.Another explanation of an increasing retention rate may be due to cross-sectionalheterogeneity in individual retention probabilities across customers in their prefer-ence for a product/service (e.g., Gupta and Lehmann2005, 29–31) The heteroge-neity problem can be particularly severe if the company shows substantial growthand thus manyfirst-time customers whose churn can be different from older cus-tomers, thus distorting the calculation of how much repeat purchase behavior willoccur in the future.

Company.net’s Business Intelligence unit has recognized and addressed thesecrucial issues The BI unit applies a heuristic methodology to estimate LTV fromfuture acquired customers by projecting into the future the data of each acquiredcustomers’ cohort with a data-mining approach Cohort analysis has been used bystatisticians for decades However, recent advancements in data collection andprocessing power have made cohort analysis a viable technique for online businesses

to study customer loyalty trends, predict future revenues, and monitor churn Themost popular cohort analysis (and the one we present in this case study) involvessegmenting customer groups based on a“join date.” The month, week, or day thenbecomes the user’s “cohort,” meaning each cohort is the cluster of users who joined

in that same time period The pivotal metrics used in this analysis is called Marginper Thousand Customers (MpK) and focuses on projected margins.8

Company.net estimates across cohorts an average MpK for the last N months ofactual data and then, after a normalization based on previous observations, it projects

it into the future months The metric is calculated with different windows, namelydaily, weekly (preceding 7 days) and monthly (preceding 30 days) The CLV iscomputed with the following assumptions: 1) LTV is based on future profit (seedistinction between cash flow and profit Gleaves et al 2008); 2) CLV does notdirectly consider the CoA (see Pfeifer et al.2005).9Formally CLV is obtained bysubtracting CoA from LTV, i.e acquisition costs are not included as part of lifetimevalue However, customer acquisition cost is often displayed alongside a customer’sLTV In this way, Company.net gains insights about whether an unprofitablecustomer whose CLV (LTV minus CoA cost) is negative is due to high costs ofacquisition rather than a low LTV

8 Using the same rationale at Company.net also computes Revenue per Thousand Customers (RpK) focusing on projected revenues instead of margin In the following example, we will concentrate our attention on the MpK evolution, however the same report can be adopted for RpK, since in fact gross margin is a percentage of the revenue (in this example 65%) and the difference between Rpk and MpK is merely a question of scale Our choice to present MpK is because with this metric we are allowed to calculate CE.

9 When it comes to make informed prospecting decisions, there are at least two ways of considering acquisition spending (Pfeifer et al 2005 ) Either not include acquisition spending and compare the Lifetime Value (LTV) to CoA; or alternatively, include acquisition spending in the speci fication of customer value, correctly labeled as CLV, and compare the value of CLV to zero.

The following steps are taken in order to compute the MpK metric:

1) Collect gross additions (acquired subscribers) for each cohort and thecorresponding number of billings (subscribers who effectively pay) in eachperiod It is worthwhile noticing that gross addition is always higher than thenumber of subscribers billed for two reasons First, some of the subscribers quit(this phenomenon is measured by churn rate) Second, it is not possible to chargeall the candidates subscribers This phenomenon is measured by the YIELD.2) Multiply the number of billed subscribers for the Margin per customer in order todetermine MpK

3) Estimate for each considered period of time the average MpK among the differentcohorts By averaging across cohorts, one can determine an average MpK at theend of 1 month, 2 months, and so forth (Eq.2.1)

t¼ period of time ¼ cohort

n¼ total periods where data is available

#_Billing¼ paid subscribers for each cohort

The estimation also requires a normalization of the data (i.e outliers are notconsidered in the average) In most cases, outliers are due to problems in thefirm’s information system that produces the log files and also based on previousexperience they can be easily identified

4) Project average MpK on the lifetime of the customer in order to estimate howmuch margin can be obtained from the acquisition of 1000 customers today Inother words, with this step it is possible to estimate the tail of the MpK for thefuture (Fig.2.4)

As the cohorts mature, there are fewer datapoints to average across, and hence thepotential for error increases However, it is still a useful exercise to assess the futuremargin associated with the acquisition of a thousand customers How far into thefuture the estimation can be extended depends on the type of business

At Company.net, the following metrics are computed across different timeranges:

In theory, CLV models should estimate the value of a customer over theirlifetime However, in manyfirms, including Company.net, 3 years is considered areasonable estimate of the horizon over which the current business environment2.4 CLV Scorecard and Cohort Analysis: An Application in an SBE 27

(with regard to technology, competition, etc.) would not substantially change(Kumar2008a; Kumar and Rajan 2009) Next, we apply the previously outlinedequation and develop an example about how a typical cohort analysis is conductedfrom thefirm’s log file to compute the MpK’s drawing on the rationale explained inthis section.

2.4.4 CLV Cohort Analysis: A Practical Illustration

Our example assumes that Company.net launched an online service in a newcountry Thefirm’s customer acquisition and collection are described in Table2.2.Thefirm acquired a first cohort of 25,016 customers in the first month and thesecustomers produced the following stream of payments: 8614 billing events in thefirst month, 13,437 in the second month, and so on Another cohort of 38,862customers was acquired in the second month, which generated a separate stream

of billing events (15,086 thefirst period, 27,082 the second, and so on)

For the calculation of MpK, it is necessary to complete the scenario with ARPU

as an input variable For the example covered here, we hypothesize an ARPU of $10which is considered to be‘booked’ at the beginning of the contract period, with acost of service of $4.50 (i.e., resulting Contribution Margin of $6.50) We alsohypothesize a customer acquisition cost of $11 per customer (this figure is anestimate derived from the ratio of marketing cost/gross addition) Company.netraise the following questions raise at the end of May:

1) What is the expected contribution, all things being equal, of the next 1000customers we will acquire?

2) What is the marketing investment’s pay-back period?

3) If we stop investing in this market/service, what is the expected residual value ofthe customers currently on the firm’s books (i.e customer equity of currentcustomer base)?

To properly answer these questions, they develop a report which follows theMpK rationale Thefirst step in preparing the report is to transform billing events inmonetary terms by multiplying billing events by margin and scaling the result by onethousand customers (Table 2.3) For example, the January figure ($2238) wasobtained by multiplying 8614 (January’s billing) with $6.50 (margin) and finallymultiplying by 1000/25,016 in order to scale by one thousand customers

The next crucial step requires the computation of an average contribution marginestimated for each period The question is how much margin should be expectedfrom one thousand customers acquired today Given the above-mentioned Eq (2.1),

it follows that one thousand costumers acquired today are expected to provide agross margin of 2905.65 in the next month (average of 2238; 2523; 3828; 3457;2482); 3908 in the second month (average of 3491; 4530; 4245; 3367), and so forth(refer to Table2.4) As the cohorts mature, there are fewer data points to average

Table 2.2 Customer acquisition (Gross addition) and collection (billing events)

Period of time

Cohort Gross addition Number of customers billed

Note: Figures were disguised by a constant multiplier for con fidentiality reasons

Source: Bonacchi and Perego ( 2012 )

Table 2.3 Evolution of gross margin for each cohort

Period of time

Cohort Gross addition Gross margin X # customers billed scaled by thousand customers

Note: Figures were disguised by a constant multiplier for con fidentiality reasons

Source: Bonacchi and Perego ( 2012 )

2.4 CLV Scorecard and Cohort Analysis: An Application in an SBE 29

over Hence, for the subscribers that started in month one, we have 5 months of data,for the subscribers that started in month two, 4 months of retention data are available,and so forth The number of actual data (i.e., number of cohorts) depends on dataavailability, with the usual range from a minimum of two to a maximum of

12 months of past data Regarding the normalization of the data, the analysis ofthe standard deviation allows detecting in which cases the average is biased fromoutliers In those cases, the outliers are removed from the average

A typical pattern found in Company.net is that after an initial period, month MpK tends to level off With such a pattern, one can extrapolate forwardusing the same month-on-month decrease across several months The followingalgorithm is applied to forecast the evolution of MpK for the cohort whose data isnot available (in our example from the cohort number six onwards): drop fromprevious month multiplied by an x% of decrease to account for a decrease in churnrate (in this example, the percentage of decrease is supposed to be 95% for eachmonth) In our example, we have 5 months of data and then we extrapolate forwardusing the same month-on-month decrease on the basis of previous experience for thesubsequent 31 months Stated differently, MpK of period six (1988) is obtained asfollows: 2420.61 (19%  95%) This method allows projecting cohort’s MpKinto the future, estimating the distribution’s tail (as exhibited on the right side ofFig.2.4) As already mentioned, at Company.net this estimation procedure usuallydoes not exceed 36 months when forecasting MpK

month-by-Further, Table2.5provides an answer to thefirst two questions previously stated,namely: one thousand customers acquired today are supposed to produce a margin of

$31,121 (2906 + 3908 + + 123 + 118) during the next 36 months This result isobtained with the following algorithm

36

MpKt

Table 2.4 Projection of MpK in the future

Time period MpK Standard deviation Rate (%)

Thisfigure must then be compared with CoA in order to obtain a CLV of $20,221(31,121–11,000), as illustrated in Table2.5.

We obtain the pay-back of the marketing campaign from Table2.4 summingMpK up to $11,000 (i.e., 2906 + 3908 + 3601 + 2982) This is quite relevantinformation whenfirms have to optimize the resources spent on customer acquisitionand evaluate the risk of a marketing investment However, the value of futurecustomers and the payback period alone are not sufficient to obtain a full picture.The other piece of information required to make rational decision in this businesssetting is Customer Equity (CE)

In order to answer the third question, a further step is needed to estimate the CE

of current customers extending the actual data of each acquired cohort (from 1 to 5)with the estimation for the future This means that data in Table2.6Column A must

be projected into the future using data from Table2.6Column B Hence, for eachcohort we sum up the future periods (i.e for cohort 1 the period 6 to 36; for cohort

2 period 5 to period 36; for cohort 3 period 4 to 36; for cohort 4 period 3 to 36; andfor cohort 5 period 2 to 36) The last step is to un-scale the data in order to considerthe real gross addition for each cohort, since in fact all the calculations made sofar refer to 1000 customers The result is a CE of the current customer base of

$5,328,426 (Table2.6), representing the value embedded in the acquired base

customer-The PMS of Company.net, due to the normalization of the data, is also able tosimulate the effect of other factors that are not under the control of the company,such as competitors’ actions, regulation changes, or technological discontinuities,any of which may affect the consumer behavior in both conversions and retention.This is particular important when it is necessary to neutralize the effect of theabovementioned factors from the metric utilized as the base for managers’ bonuses

As an example, consider the case of a problem in the logfiles transmitted from thetelecommunications carriers to Company.net for a specific cohort

Table 2.5 Typical report to support decision-making

Note: This example of report highlights the relation among cost of acquisition and Lifetime Value (estimated using MpK) This type of reports are crucial in order to assess the contribution of new subscribers to value creation

Source: Bonacchi and Perego ( 2012 )

2.4 CLV Scorecard and Cohort Analysis: An Application in an SBE 31