(Luận văn) consumer analytics toward development of cross selling products at retail banking an approach on big data at eximbank

t to ng hi ep MINISTRY OF EDUCATION AND TRAINING w UNIVERSITY OF ECONOMICS HO CHI MINH CITY n lo ad ju y th yi pl ua al n TO PHUC NGUYEN KHUONG n va ll fu m oi CONSUMER ANALYTICS TOWARD DEVELOPMENT OF nh at CROSS SELLING PRODUCTS AT RETAIL BANKING: z z AN APPROACH ON BIG DATA AT EXIMBANK k jm ht vb om l.c gm MASTER THESIS n a Lu n va y te re Ho Chi Minh City – 2020 th t to ng hi ep w MINISTRY OF EDUCATION AND TRAINING n lo ad UNIVERSITY OF ECONOMICS HO CHI MINH CITY ju y th yi pl al n ua TO PHUC NGUYEN KHUONG n va fu ll CONSUMER ANALYTICS TOWARD DEVELOPMENT OF oi m at nh CROSS SELLING PRODUCTS AT RETAIL BANKING: z AN APPROACH ON BIG DATA AT EXIMBANK z vb jm ht Specialization: Business Administration k Executive Business Administration : 8340101 om l.c gm Code TUTOR: ASSOC.PROF DR TU VAN BINH n a Lu n va y te re Ho Chi Minh City - 2020 th t to ng hi COMMITMENT ep I assure you this is my own research The figures and results stated in the thesis are honest and have not been published in any other works w n lo I would like to assure you that all of the help for the implementation of this dissertation was thanked and the information cited in the thesis has been traced ad ju y th yi Trainees implement the thesis pl n ua al (Sign and write full name) n va ll fu oi m TÔ PHÚC NGUYÊN KHƯƠNG at nh z z k jm ht vb om l.c gm n a Lu n va y te re th t to ng hi ep w n lo ad ju y th ABSTRACT yi Based on internal database as big data of the Eximbank, 3,527 active customers with the initial length of service more than 12 months, the method of data mining is concerned, in which the mathematic methods of K-Means of cluster, Tree Decision, and Association Analysis are applied The findings show that consumers’ characteristics using services at EXIMBANK are various, in which staffs, directors as individual customers occupy a high proportion in total customers Based on descriptive statistics, there are eight main products, such as VG (Visa Gold), VC (Visa Classis), MG (Master Gold), MS (Master Standard), VP (Visa Platinum), VV (Viva Violet Card), VA (Visa Auto Card), and others are concerned most by the customer, in which VG and VC are the two top cards used Directors are more interested in VG card, while staffs concern VC card In addition, using two cards, called the main card and the extra card, is popular This is a potential chance for the bank to develop cross-selling products, which the product bundle strategies are packed into groups Based on the method Association Analysis, propobality of using two cards or three cards at the same time of customers are derived This finding basically supports the bank to address cross-selling products of cards to customers To this, recommendations of the strategies of bundle products are suggested in the thesis, together with implementation plans pl n ua al n va ll fu oi m at nh z z k jm ht vb om l.c gm n a Lu n va y te re th t to ng hi ep w TABLE OF CONTENTS n lo COMMITMENT ad y th ABSTRACT ju TABLE OF CONTENTS yi LIST OF FIGURES pl ua al LIST OF TABLES n CHAPTER 1: INTRODUCTION RATIONALE OF RESEARCH OBJECTIVES OF STUDY RESEARCH QUESTIONS SCOPE AND LIMITATION RESEARCH METHODOLOGY n va ll fu oi m I II III IV V nh at Data collection z z Methods to analyze data vb CHAPTER 2: CONCEPTS CONCERNED AND ITS FRAMEWORK ht CONCEPTUAL FRAMEWORK jm I k Service retention gm Remaining good relationships between the customer and retail banking l.c 1.The concept of big data and its consideration in banking om Big data in Banking and its contribution a Lu Application of Big Data to investigate the customer’s spending habits 10 n Customer segmentation and review customer’s records 10 va n The sales includes other services (service cross-selling) 11 th Flexible suggesting good services to customers 12 y Personalized marketing 12 te re Building a system to record customer feedback 11 t to ng hi ep w n lo II CONCEPTS RELATED TO APPROACHED QUANTITATIVE METHODS 13 RFM 13 Customer Value Matrix Model 14 Customer lifetime value based on RFM 16 Customer segmentation 18 Cross-selling products 19 ad y th ju CHAPTER 3: BIG DATA OF BANKING INDUSTRY AND CONCEPTS APPROACHED………………………………………………………………………21 yi pl I SITUATION OF EXIMBANK AND ITS BUSINESS 21 II BIG DATA AND ITS APPLICATION IN BANKING 25 ua al n III DATA ANALYTICS 29 va n 1.PROFILE OF CUSTOMER 29 fu ll 2.RMF AND MARKET SEGMENT 33 m oi Cross – Selling strategy 38 nh Summary of chapter 43 at z CHAPTER 4: CONCLUSION AND SOLUTIONS 44 z GENERAL CONCLUSION 44 FINDING OF MARKET SEGMENTS 44 DETAILED CHARACTERISTICS BY MARKET SEGMENT 45 IDEAS OF MARKETING STRATEGY 46 SOLUTIONS 46 k jm ht vb gm I II III IV V om l.c PRODUCT STRATEGY 46 PRICE STRATEGY 48 a Lu PROMOTION STRATEGY 49 n PROCESS STRATEGY 50 va n VI EMPLICATION OF STATEGY 48 th APPENDIX y Web link: te re REFERENCE t to ng hi ep w LIST OF FIGURES n lo ad Figure 2.1: Customer value matrix and its description 15 y th ju Figure 2.2: Customer value matrix and its description 16 yi Figure 2.3: The diagram of proposed model 18 pl ua al Figure 3.1: Share of investment in big data analytics by sector 26 Figure 3.2: Share of investment in big data analytics by sector 26 n n va Figure 3.3: Position of customer 29 ll fu Figure 3.4: Type of products that consumers concerned 30 oi m Figure 3.5: Length of service of active customers 31 nh Figure 3.6: Type of customers using services at the bank 32 at Figure 3.7: Type of products that consumers concerned 33 z z Figure 3.8: Matrix of LOS and Recency 34 vb ht Figure 3.9: Market segments based on RFM 36 k jm Figure 3.10: Result of Tree Decision of five market segments 38 gm Figure 3.11: Result of Association Analysis and its potential bundles 39 Figure 3.12: Characteristics of potential bundle product between MS and VA 40 l.c om Figure 3.13: Potential bundle strategy (MS-VA) by customer income 41 a Lu Figure 3.14: Characteristics of potential bundle product among VP, MS, and VA 41 n Figure 3.15: Potential bundle strategy (VP-MS-VA) by customer’s income 42 n va Figure 3.16: Characteristics of potential bundle product among MG, VV, and VA 43 th Figure 4.2: Marketing strategies suggested to the bank 46 y Figure 4.1: Ranges of account balance of customers 45 te re Figure 3.17: Potential bundle strategy (MG-VV-VA) by customer’s income 43 t to ng Figure 4.3: Type of product 47 hi ep Figure 4.4: Top prestigious banks in Vietnam 48 w n lo LIST OF TABLES ad y th ju Table 1.1: Information of variables concerned Table 2.2: Presentation of customer value matrix 15 49 pl 14 ua yi Table 2.1: Recency, Frequency and Monetary Score Description al n Table 4.1: Interest rate of top ten prestigious bank in Vietnam va 53 n Table 4.2: Plan of strategic implication ll fu oi m at nh z z k jm ht vb om l.c gm n a Lu n va y te re th t to ng hi CHAPTER 1: INTRODUCTION ep RATIONALE OF RESEARCH w I n lo Rapid development of information technology in the banking industry has created a big concern that the banks must think of how to explore internal data to serve competitive strategies In fact, currently, most banking institutions, insurances and financial services are attempts to adopt a new approach toward data mining to the development and innovation of services that they provide to customers Like most other industries, big data analytics are going to be a major change to support business units to generate campaigns for a short term and long term strategies to attract more new customers, retain existing ones and fight against the competitors ad ju y th yi pl ua al n Investing a big data system cause simulation that data mining is much concerned in the banking industry, because it supports to extract valuable information form huge amounts of data Particularly it contributes into finding out consumer behavior and present a market situation picture of a firm However, it is not easy to explore big data of firms Data scientists developed quantitative methods as mathematic approaches, such as descriptive and predictive analytics, etc n va ll fu oi m nh at Nowadays, banks have realized the importance of customer relation, this is one of successful factors However, challenges of how to retain most profitable customers and to reduce a churn rate are problematic To solve this problem, consumer behavior should be investigated and analyzed, which big data are a worth resource and quite helpful to measure and predict consumer behavior correctly Therefore, the power of data is to derive utility across various spheres of their functioning, product across selling, regulatory compliance management, risk management, and customer service management z z k jm ht vb gm om l.c As we know, particularities of banks’ activities generate a huge amount of data from unstructured data, such as transaction history, customer to the unstructured data such as the customer’s activities on the website, or the mobile banking application on social networks However, how to explore big data available is the most problem Although there are not few banks who recognize that and want to turn the big data available to the most effective weapon for the market competition, they are seemly facing problems of new system, skills, and so on n a Lu n va y te re th With changes in integration policies of Vietnam through information technology system, together with fire competition, more and more banks in Vietnam have paid more attention to investing a huge money for big data system and people capacity For example, EXIMBAK is one of the banks, who is willing to pay money to structure data t to ng hi ep system very early It realizes benefits of data warehouse to support business strategies In fact, the benefits of internal brought not the raise of the internal management efficiency, but also help increase the competitive advantages, maximize profits w Currently, the banks must be flexible for their plan toward activities of innovation, in term of capturing needs of customers and improving satisfaction and retention of customer By the way, CRM is quite helpful and recognized for acquisition and retention of customers As a result, the bank can get more opportunism to make long lasting and profitable relationships with customers n lo ad ju y th yi To investigating big data system of the bank, Eximbank has offered a program of building capacity for staffs, who are directly related to data analytics and business development In addition, the department of database is established to exploit internal database, which the role of an employee is placed in the ecosystem talent development However, everything is seemly concerned and exploited more and more toward data mining, predictive analytics on customer behavior pl n ua al n va ll fu OBJECTIVES OF STUDY oi m II Presenting characteristics of consumers who are using services of banking - Analyzing history of transaction behavior of consumers - Identifying and developing market segments through consumer behavior - Developing developments of cross-selling products to increase benefits of customers toward customer retain at nh - z z k jm ht vb RESEARCH QUESTIONS l.c There are some questions concerned as follows gm III om (1) What are characteristics of consumer toward service usage at banking? n va (3) How is the market segment developed? a Lu (2) How customers take transaction at the bank? n (4) What are cross-selling product developments to increase benefits of customers to retain them? y te re th 42 t to ng hi ep Continuously, the figure 4.12 is a message to tell us that the potential bundle strategy can be considered on customers belong the best group and the frequent group with the income level less than 300 million/month w Figure 3.15: Potential bundle strategy (VP-MS-VA) by customer’s income n lo ad ju y th yi pl n ua al n va ll fu oi m nh at To Rule ID10: There are 2,666 customers who are concerning on using cards of MG, VV and VA This means once the customer is using the cards of MG and VV, they are going to use VA This is happened at confidence of 95.01% and presents a positive association among them As resulted in figure 13, there is a serious risk that can happen to the bank The customers belong to the Rule ID7, mostly they are defined as “Lost Relationship”, particularly for the market segments of best group and shopper group As resulted in figure 3.17 A suggested bundle strategy with the association of MG, VV, and VA can focus on the customers who belong to the frequent group with the average income of less than 400 million VND per month z z k jm ht vb om l.c gm n a Lu n va y te re th 43 t to ng Figure 3.16: Characteristics of potential bundle product among MG, VV, and VA hi ep w n lo ad ju y th yi pl n ua al va n Note: CR= Close Relationship; ER = Establishing Relationship LR = Lost Relationship; PR = Potential Relationship ll fu m oi Figure 3.17: Potential bundle strategy (MG-VV-VA) by customer’s income at nh z z k jm ht vb om l.c gm n a Lu n va Summary of chapter: Based on mathematic method, five market segments are developed Particularly, with applying Association Analysis, the findings present potential cross-selling products Although there are many potential bundle strategies, five of them are the best feasible, such as, the cross-selling product of MS and VA (Rule ID2), of VP, MS and VA (Rule ID7), and MG, VV, and VA (Rule ID10) These associations are basic for the next chapter to think of solutions to meet consumers’ demand and retain them to be close relationship y te re th 44 t to ng hi ep CHAPTER 4: CONCLUSION AND SOLUTIONS w n lo GENERAL CONCLUSION ad I ju y th Based on data mining of 3,527 customers from source of big data, the mathematics methods of K-Means cluster, Decision Tree, Association Analysis are taken into account of Initially, through descriptive statistics, the general picture of actual situation of the bank is presented Four main groups of customers are using the service of Eximbank, such as, directors, managers, staffs, and some of workers These customers have a long time of engagement in the bank’s service, which there are 60% up the length of stay (LOS) of customer stay with the bank more than years, in which around 11% with LOS of years up yi pl n ua al n va ll fu Mostly customers pay more attention to using two card, so called the main card and extra card The use of two cards at the same time is an expression of the needs of users This result also should have noted the growing problem of cross-selling strategies that exploit all market demands in the future, especially for the income groups and have high positions oi m at nh z These cards are grouped into eight main groups, such as VG (Visa Gold), VC (Visa Classis), MG (Master Gold), MS (Master Standard), VP (Visa Platinum), VV (Viva Violet Card), and VA (Visa Auto Card) However, VG and VC are both popular and concerned the best Customers concerning top two products, e.g VG and VC, which VG is seemly more interested by directors and VC is more interested by staffs In addition, customers who are using VG have greater income than that of VC z k jm ht vb l.c FINDINGS OF MARKET SEGMENTS gm II om With the usage of K-means analysis, in which three elements of R, F, and M are enclosed in cluster analysis Based on that five segments are developed The uncertain group is little high of 27.5%, while the customers are the best group accounting for 8.5% In addition, still many customers come to the bank for small transactions n a Lu n va The results bring an important message, the best group is dominated by customers who are directors and managers, while the customers belong the uncertain group are mainly workers, this is consistent with practices, due to the income level of workers are low and unstable Results are good for the bank to create appropriate strategies and stimulate more transaction values in the future Oppositely, the message also confirms that workers y te re th 45 t to ng hi ep and staffs are uncertain ones, this is not surprised, because these people are low income, particularly for worker they are often unstable job In addition, the findings of market segment show that credit card customers and personal consumption customers account for the high percent of expenditure of customer’s transaction This means that these customers sometimes have a huge transaction value w n lo ad ju y th III DETAILED CHARACTERISTICS BY MARKET SEGMENT yi The churn group and uncertain group can be display as a risk group Once characteristics of this group are investigated more detailed, business opportunities are improved and risk limitation is displayed The reason of this consideration is to address right strategies to improve the business To investigate more detailed, the method of “Tree Decision” is employed There are five market segments played as the target variable combined with length of stay (LOS) and balance account of active customers in the study Measures of LOS and balance account are defined as follows pl n ua al n va fu As stated previously, the four ranges of LOS, e.g range of >12 – 31 months (37.51%), range of >32 – 45 months (20.73%), range of >45 – 52 months (16.87%), range of >52 – 65 months, and range of > 65 – 81 months (11.43%) - Measure of account balance of active customers is based on balance of the study period of 12 months Three ranges (Low, Average, High) of account balance are grouped in figure 5.1, which definition of the low account balance is < million VND accounting for 37,51%, that of the average account balance as – 20 million VND accounting for 35,47% and that of the high account balance as 20 – 1,420 million VND accounting for 27,02% ll - oi m at nh z z k jm ht vb om l.c gm Figure 4.1: Ranges of account balance of customers n a Lu n va y te re Source: Result of analytics on internal data of the bank th 46 t to ng IV IDEAS OF MARKETING STRATEGIES hi ep As found, characteristics of consumption with five market segments are a high reliability, because these results are measured on the consumption history of customer recorded in the system To effectively exploit the market, the marketing strategy of the bank needs to create typical campaigns to meet the customer’s demand w n lo ad To gain marketing strategy development successfully, strategies suggested should be associated with five market segments found Suggestions of the marketing strategy development are offered as bellows (figure 4.2) ju y th yi Figure 4.2: Marketing strategies suggested to the bank pl n ua al Marketing strategy’s suggestion n va fu ll Product strategy oi m Process strategy at nh Price strategy Promotion strategy z z ht SOLUTIONS vb V k jm Customer segmentation derived is based on characteristics of transactions This finding contributes to prepare a framework for segmenting customers based on their transaction for estimating future value of different segments of customers in the bank, particularly for the retail banking scope In addition, results of empirical analysis could be a guideline for marketing strategies, developing sales programs to launch new products for each group, and stimulating consumption of the most valuable customer group om l.c gm a Lu PRODUCT STRATEGY n In general, there are many kinds of products that the bank is doing As depicted in figure 5.3, there are 15 products to provide the customers However, as grouped previously, main products are discussed, e.g VG (Visa Gold with 24.24%), VC (Visa Classis with 22.31%), MG (Master Gold with 12.08%), MS (Master Standard with 11.31%), VP (Visa Platinum with 9.27%), VV (Viva Violet Card with 6.89%), and VA (Visa Auto Card with 3.40%) The rest of products is labeled as other groups accounting for 10.49% (included: JP = JCB Platinum; JC = JCB Standard; JETSTAR = JCB Credit Jetstar; JG n va y te re th 47 t to ng hi ep = JCB Gold; VB = Visa Business; MW = Master One World; PB = Master Passbook Card) w Figure 4.3: Type of product n lo ad ju y th yi pl n ua al n va ll fu oi m at nh vb VISA GOLD VISA CLASSIS MASTER GOLD MASTER STANDARD VISA PLATINUM VISA VIOLET CARD VISA AUTO CARD JP = JCB Platinum; JC = JCB Standard; JETSTAR = JCB Credit Jetstar; JG = JCB Gold; VB = Visa Business; MW = Master One World; PB = Master Passbook Card EEFA = UAFA Card k jm ht om l.c gm VG: VC: MG: MS: VP: VV: VA: OTHER: z Note of product type: z Source: Result of analytics on internal data of the bank n a Lu va n For the best group, products of MG, VP and VG should be introduced to customers with the income class larger than 75 million VND/month (see figure 5.4), which directors and managers are the target market y te re th However, the product strategy also think of cross-selling products with a bundle between VP and VA and a bundle between MG and VA Which the customers belong to the ‘close 48 t to ng hi ep relationship’ with the average income level of less than 500 million VND per month, particularly they are segmented in the best group and the frequent group w n PRICE STRATEGY lo ad According to the Vietnam report, top ten prestigious banks in Vietnam in 2018 are presented in figure 5.4, they are Vietcombank, VietinBank, BIDV, Techcombank, ACB, MBBank, VPBank, Agribank, SHB and Sacombank The ACB is ranked at the fifth position As a result, competitions among banks in Vietnam are fierce, so each bank has a price strategy As presented in table 4.1, the ACB sets a highest interest rate of 5.40% for customers who deposit money with larger amount of 10 billion VND of the term of months ju y th yi pl ua al n Price strategy should be applied on different types of customer and based on price levels of competitors presented in figure 5.3 Suggested price strategies are the following n va To customers who want to borrow money for consumption, e.g buying car, motorbike, so on the interest rate can be 9% for months and 9.5% for 12 months To customers who borrow money for family business, the interest rate can 7.8% applied for short term of months ll fu - oi m - nh at Figure 4.4: Top prestigious banks in Vietnam z z ht vb Point of Media coding Point of survey k jm Financial point om l.c gm n a Lu n va y te re th 49 t to ng hi Source: Vietnam Report – link: https://vietnambiz.vn/so-sanh-lai-suat-10-nganhang-uy-tin-nhat-viet-nam-thang-112018-108736.html ep w n lo ad ju y th Table 4.1: Interest rate of top ten prestigious bank in Vietnam yi Deposit (VND) n ll fu m k jm ht 18 months 6.60% 6.70% 6.70% 6.60% 6.70% 6.80% 6.90% 7.00% 7.055 7.10% 7.15% 7.20% 6.90% 7.10% 7.10% 7.20% 7.30% 7.30% 6.80% 7.30% 7.40% 7.20% 24 months 6.60% 6.80% 6.80% 6.60% 6.70% 6.80% 6.50% 6.60% 6.65% 6.70% 6.75% 6.80% 7.50% 7.10% 7.10% 7.20% 7.30% 7.30% 6.80% 7.30% 7.40% 7.30% gm Sacombank vb 10 z Agribank SHB z at MBBank VPBank Interest rate of banks by term 12 months months months 5.50% 5.50% 6.60% 5.50% 5.50% 6.80% 5.50% 5.50% 6.80% 5.90% 5.90% 6.60% 6.00% 6.00% 6.70% 6.10% 6.10% 6.80% 5.80% 5.80% 6.50% 5.90% 5.90% 6.60% 5.95% 5.955 6.65% 6.00% 6.005 6.705 6.05% 6.05% 6.7% 6.10% 6.10% 6.80% 6.00% 600% 7.20% 6.50% 6.50% 7.00% 6.50% 6.50% 7.005 6.50% 6.50% 7.00% 6.60% 6.605 7.00% 6.60% 6.60% 7.00% 5.50% 5.60% 6.80% 6.80% 6.905 7.00% 6.90% 7.00% 7.10% 6.10% 6.20% 6.90% nh months 4.80% 4.80% 4.80% 4.80% 4.90% 5.00% 5.10% 5.20% 5.25% 5.30% 5.35% 5.40% 5.30% 5.20% 5.20% 5.30% 5.40% 5.40% 4.80% 5.30% 5.30% 5.305 oi ACB month 4.40% 4.50% 4.50% 4.70% 4.80% 4.90% 4.80% 4.90% 4.95% 5.00% 5.05% 5.10% 4.70% 5.10% 5.105 5.10% 5.20% 5.20% 4.50% 5.10% 5.10% 4.70% va