Customer Insight development in

Emanuele Baruffa – Vodafone Seugi - Vienna, 17-19 June 2003

Seugi 21_Vienna Pag. 1 ContentsContents::

1. Introduction

2. Customer Base Management

3. Customer Insight

4. Data Environment

5. Conclusions

Seugi 21_Vienna Pag. 2 1. Introduction

2. Customer Base Management

3. Customer Insight

4. Data Environment

5. Conclusions

Seugi 21_Vienna Pag. 3 MobileMobile telephonytelephony isis oneone ofof thethe fastestfastest growinggrowing industriesindustries inin thethe worldworld

Worldwide growth in subscribers (millions) ! 1 billion

subscribers 1480 around the world 1321

! Over 120 1152 countries have 958 mobile networks 727 ! Further acceleration 479 expected 206 87 14 34

Source: EITO 1991 1993 199 5 1997 1999 200 0 2001 2002 2003e 2004e

Seugi 21_Vienna Pag. 4 Italy:Italy: Europe’sEurope’s secondsecond biggestbiggest mobilemobile marketmarket

Country Subscribers Penetration Western European TLC market % growth by country (%) 10 9,1 9 Germany 60,300,000 84% 8 6,8 6,8 7 6,2 6,1 5,7 5,8 Italy 54,000,000 98% 6 5,6 5,6 5,4 5,0 5 UK 50,900,000 92% 4,0 4

France 39,000,000 77% 3 Spain 34,000,000 88% 2 1

0 Germany Italy UK France Spain Western Europe Sources: internal sources for Italy, Yankee Group for Source: EITO other European countries 2001/2002 2002/2003

Seugi 21_Vienna Pag. 5 PenetrationPenetration raterate inin thethe ItalianItalian marketmarket

60,000 96% 98% 100% 91% Subscribers (,000) 90% 50,000 Penetration Rate 80% 74% 70% 40,000 60% 53% 30,000 50% 36% 40% 20,000 30% 21% 20% 10,000 11% 10% 0 0% 1996 1997 1998 1999 2000 2001 2002 2003

Seugi 21_Vienna Pag. 6 ItalianItalian marketmarket sharesshares

Total subscribers on 31.03.03

17% 36%

47%

Seugi 21_Vienna Pag. 7 VodafoneVodafone ItalyItaly CustomerCustomer BaseBase

18,900,000 17,400,000 14,920,000

10,418,000

6,190,000

2,460,000 713,000

Dec. 1996 Dec. 1997 Dec. 1998 Dec. 1999 Dec. 2000 Dec. 2001 Dec. 2002

Seugi 21_Vienna Pag. 8 1. Introduction

2. Customer Base Management

3. Customer Insight

4. Data Environment

5. Conclusions

Seugi 21_Vienna Pag. 9 Customer management strategy

! Strategy: Consolidate leadership through customer base management ! Marketing Goals: meet customers’ needs and identify the best treatment to each customer at the right time through the most suitable channel at the appropriate cost. This approach helps to increase ! customer loyalty ! customer value (ARPU)

How

! create customer insight building a Customer Centric DB (CKM) that allow to: ! Identify segments for cross-selling actions " Customer Segmentation ! Identify customers with highest value potential for new high value added services such as SMS, Voice Mail " Propensity to uptake VAS ! Identify customers with highest churn probability " Churn propensity score

Seugi 21_Vienna Pag. 10 1. Introduction

2. Customer Base Management

3. Customer Insight

4. Data Environment

5. Conclusions

Seugi 21_Vienna Pag. 11 ApproachApproach forfor developingdeveloping customercustomer insightinsight

2. Gather 1. Brainstorming comprehensive customer information

3. Analyze and segment 6. Fine Tuning customer base

5. Measure 4. Data Mining results

Seugi 21_Vienna Pag. 12 CustomerCustomer ValueValue ScoreScore

% CB % Margin ! High value customers make up for majority of ~ 30% 70-80% margins High value ! Margins are relevant and customers not revenues because some high spenders also have high interconnection cost ~ 70% 20-30% Low value ! Resources should be 20-30% customers allocated proportionally to margins generated

Value is calculated on a monthly basis for each customer. Customer value is measured based on a four month average Marpu

Seugi 21_Vienna Pag. 13 ChurnChurn ModellingModelling

MODEL MACRO TARGET DATA MICRO MODEL MODEL IMPLEMEN- SEGMENTATION DEFINITON ANALYSIS SEGMENTATION BUILDING EVALUATION TATION

•Identify •Define the event •Data Exploration •Identification of •Oversampling •Implement •Ex-post relevant market you want to – Correlation groups with • Application of model in a evaluation of segments predict between target similar churn data mining production model •Example •Example variable and behaviour techniques environment performance – Consumer – Inactive SIM explanatory – improvement – Logistic • Monthly –% of correct # Prepaid for Prepaid variables of model regression production predictions # Contracts Customers – Data accuracy –Decision tree of churn – Corporate –“Cancellation transformation –Focus on: –Neural index # Small letter” for #Trends # “Active” Networks Accounts Contract #Grouping customers •Measurement # Large Customers # High value of the goodness Accounts customers of fit (model validation on hold-out sample)

Analyse churn behavior of actual churners to predict churn behavior of the customer base

Seugi 21_Vienna Pag. 14 ChurnChurn PropensityPropensity ScoreScore

Model Lift

High Risk to churn

Medium Risk to churn 5.5

Low Risk to churn 3.2

0.5

Red Yellow Green

The likelihood to churn is estimated on a monthly basis for each VO customer. Customers flagged in RED are five times more likely to churn than an average customer.

Seugi 21_Vienna Pag. 15 Segmentation Methodological Road Map Segmentati Data Clusterin Segmentati on Data Analysis g on Mart Algorithm

Build a data Factoral Cluster analysis Application mart with all Analysis to •Started from a of segmentation identify main large number of segmentatio variables dimensions rules to •Demographic of micro-clusters the entire data segmentatio •Then reduced/ customer •Voice Traffic n aggregated base (peak-offpeak, clusters to a •Customers network, Elimination “manageble” and are assigned discounted of outliers to the tariffs) meaningful nearest •Service usage number cluster (rule: (wap, gprs, minimum sms, mplay, Cluster validation Euclidean music, …) distance with market •PropensityCustomersto are assigned to segments based on fromtheir usagecluster of research VAS uptakemobile services, attitude towards technologies andcentroid) their lifestyles

Seugi 21_Vienna Pag. 16 PropensityPropensity toto UptakeUptake VASVAS

Developed and implemented statistical models to predict on a monthly basis the likelihood of a VO customer to uptake five different Value Added Services.

SMS SMS No Users SMS

Low Users SMS INTERNET SELF CARE No registered customers on ADVANCED Vodafone Italy web site SMS High Users SMS not using Flash SMS VOICE MAIL No Users Voice WAP Mail WAP handset owners not using WAP

High propensity customers are on average three times more likely to start using a service than an average customer

Seugi 21_Vienna Pag. 17 1. Introduction

2. Customer Base Management

3. Customer Insight

4. Data Environment

5. Conclusions

Seugi 21_Vienna Pag. 18 CustomerCustomer KnowledgeKnowledge ManagementManagement SystemSystem (CKM)(CKM)

CampaignCampaign DWHDWH ((Oracle)Oracle) ManagementManagement Raw data Demographics daily + SysteSystemm Traffic monthly Service Usage Handset CKM Loyalty (Oracle) …………

DataData MiningMining ((SAS)SAS) DSSDSS AnalysisAnalysis OORRACLEACLE ++ FileFile SySyssttemem SASSAS ((Microstrategy)Microstrategy)

EEnndd UUserser DaDatata MiMininingng ((SASSAS EndEnd UserUser ((Web)Web) Miner,Miner, SASSAS Stat,Stat, ...... ))

Seugi 21_Vienna Pag. 19 CKMCKM ModellingModelling EnvironmentEnvironment

SAS Access Oracle ® SAS CKM Application Campa- Campa- gne gne Anagrafi Fase 1 Anagrafi ca ……. ca ……. D SAS Enterprise W SAS Connect ® Miner ® H Segmentation Data Model CKM CKM & Sampling Transformation Development

Rules

Fase 2 ScoScorere

Metadata SAS Warehouse Administrator ®

Seugi 21_Vienna Pag. 20 SASSAS CKMCKM ApplicationApplication

! A Graphical User interface: has been developed to interact with the Customer DataMart.

! Some of the functions of this GUY:

$ Dynamical datamart: extraction and definition of sub- universe for a given temporal interval with the maximum flexibility

$ Mining :

1. Development of a statistical model with SAS/Enterprise Miner®

2. Export of the model

! Assessment of predictive models when deployed: allow to verify performance on model (to measure degradation) and doing ex-post analysis on redemption.

Seugi 21_Vienna Pag. 21 1. Introduction

2. Customer Base Management

3. Customer Insight

4. Data Environment

5. Conclusions

Seugi 21_Vienna Pag. 22 InIn Summary…Summary…

A successful strategy needs to be based on a good understanding of customer needs by customer groups

Develop Customer Insight simple to understand in order to become a real working tool for all parts of the organization (customer care, marketing, sales). Take actions on customer insight and fine tune models to improve results over time Reasons to Fail: % Lack of strategy % Lack of data % Lack of statistical skills % Lack of commitment from Marketing CRM/Customer operations

Seugi 21_Vienna Pag. 23