MMFSL uses de-du­pli­ca­tion to gain a sin­gle view of cus­tomer data

De-du­pli­ca­tion ex­er­cise has en­abled MMFSL to group com­mon cus­tomers across Mahin­dra line of businesses. This pro­vides a sin­gle view of cus­tomers, en­abling the firm to iden­tify black­listed cus­tomers and im­prove cus­tomer ac­qui­si­tion ef­fi­ciency

InformationWeek - - Contents - BY VARUN HARAN tweet @varunph Varun Haran varun.haran@ubm.com

De-du­pli­ca­tion ex­er­cise has en­abled MMFSL to group com­mon cus­tomers across Mahin­dra line of businesses. This pro­vides a sin­gle view of cus­tomers, en­abling the firm to iden­tify black­listed cus­tomers and im­prove cus­tomer ac­qui­si­tion ef­fi­ciency

wo decades ago, Mahin­dra and Mahin­dra Fi­nan­cial Ser­vices Ltd. (MMFSL) com­menced its jour­ney in the ru­ral non-bank­ing fi­nance in­dus­try, with a vi­sion to trans­form ru­ral and semi-ur­ban In­dia into a self-re­liant fi­nan­cial ecosys­tem. Given the firm’s so­cially in­clu­sive busi­ness model ser­vic­ing over 2 mil­lion cus­tomers in ru­ral and semi-ur­ban In­dia, there was a need for a so­lu­tion that would in­cor­po­rate a process of check­ing en­tity de-du­pli­ca­tion, iden­ti­fy­ing sim­i­lar­i­ties of ex­ist­ing en­ti­ties and man­u­ally or au­to­mat­i­cally group­ing same en­ti­ties as one clus­ter (group).

MMFSL’s de-dupe process, chris­tened eParichay, ful­fils these busi­ness needs. The eParichay de-du­pli­ca­tion im­ple­men­ta­tion aims to es­tab­lish a sys­tem for gen­er­at­ing a sin­gle view of the cus­tomer by iden­ti­fy­ing mul­ti­ple re­la­tions held by the cus­tomer across all Mahin­dra line of businesses.

The en­tire fo­cus of the so­lu­tion is aimed at en­hanc­ing MMFSL’s dif­fer­en­ti­a­tion at the end-cus­tomer by mak­ing data qual­ity the bedrock of im­proved cus­tomer ser­vice. “The so­lu­tion helps MMFSL ac­quire new busi­ness more ef­fi­ciently while re­duc­ing the cost of ac­qui­si­tion and im­prov­ing com­pli­ance,” says Suresh A Shan­mugam, Head – BITS at MMFSL.

THE DE-DU­PLI­CA­TION PROCESS

MMFSL em­ploys sev­eral types of dedu­pli­ca­tion pro­cesses viz. set match (one time process), prime match (re­al­time process), and off­line process. The one time process de­fined by MMFSL for eParichay in­volves tak­ing the ex­ist­ing data of all en­ti­ties (cus­tomer, em­ploy­ees, deal­ers, ven­dors, etc.), and run­ning them through an in­tel­li­gent fuzzy match­ing en­gine for clus­ter­ing. The re­al­time process in­volves pick­ing up new cus­tomer data sub­mit­ted in real-time and run­ning it through an in­tel­li­gent fuzzy match­ing en­gine.

The fuzzy match­ing en­gine that is cen­tral to this ef­fort takes the data sub­mit­ted and clus­ters the en­ti­ties as MPC or auto match­ing of most prob­a­ble clus­ter and LPC, which is man­ual ver­i­fi­ca­tion and clus­ter­ing (if nec­es­sary), done for the least prob­a­ble clus­ter. Records are stored for all en­ti­ties with clus­ter (group) ref­er­ences, if clus­tered, and a golden key for each grouped en­tity is cre­ated, which is used for cus­tomer ser­vic­ing through a sin­gle win­dow.

CASH­ING IN ON THE BEN­E­FITS

Prior to eParichay, the process of ver­i­fy­ing whether a new cus­tomer had any ex­ist­ing loans from MMFL was very dif­fi­cult due to high de­pen­dence on man­ual pro­cesses. Al­most all the cus- tomers in a new loan were en­tered as a new cus­tomer as cross-ver­i­fi­ca­tion was not pos­si­ble, mak­ing it easy for even a guar­an­tor or co-ap­pli­cant to get a fresh loan. It was ex­tremely dif­fi­cult to judge the qual­ity of the cus­tomer as a re­jected cus­tomer of one branch could ap­proach an­other to avail loan.

MMFSL over­came this by mak­ing a com­mon data­base of group cus­tomers and other en­ti­ties e.g. loan cus­tomers, deal­ers, bro­kers, sales­men, em­ploy­ees, etc. A process was im­ple­mented to cleanse this com­mon data­base for ap­prox­i­mately 5.5 mil­lion en­ti­ties. A blan­ket one time de-du­pli­ca­tion was per­formed to group the com­mon en­ti­ties. MMFSL there­after im­ple­mented a de-dupe process for new ad­di­tions, whereby any new en­tity was first checked for data ac­cu­racy, then checked for pres­ence in the data­base and, if avail­able, grouped as a sin­gle cus­tomer. This en­abled a process to iden­tify black­listed, non-per­form­ing as­sets, out­stand­ing and non-starter cus­tomers at the time of loan en­try.

For MMFSL, the de-du­pli­ca­tion ef­fort has been a win­ner in many ways. Ac­cord­ing to Shan­mugam, “When most of the re­puted fi­nan­cial in­sti­tu­tions could not achieve suc­cess in clus­ter­ing and group­ing en­ti­ties, we could achieve this suc­cess­fully and come out with a sin­gle win­dow view for all cus­tomers on web por­tal with this process.”

Cus­tomer ac­qui­si­tion ef­fi­ciency for MMFSL has im­proved as has cus­tomer loy­alty. MMFSL is now able to tag cus­tomers ba­sis risk and no mis­rep­re­sen­ta­tion of re­jected cus­tomers is pos­si­ble.

Suresh A Shan­mugam

Head – BITS at MMFSL

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.