Rel­e­vance of Master Data Man­age­ment

MDM is a vital process in data gov­er­nance:

Banking Frontiers - - Data Governance - mo­han@bank­ingfron­tiers.com

Data man­age­ment pro­fes­sion­als and se­nior busi­ness man­agers to­day are in­creas­ingly be­com­ing aware of the im­por­tance of Data Gov­er­nance both be­cause of the ex­ter­nal pres­sures such as com­pli­ance and reg­u­la­tions as well share­holder rights and priv­i­leges and the in­ter­nal busi­ness pres­sures, says C.S. Sathish Jain, a strategic and col­lab­o­ra­tive man­age­ment tech­nol­ogy con­sul­tant based in the ASEAN, and MD, ASVA Con­sult­ing.

Jain, who deals pri­mar­ily with banks and fi­nan­cial ser­vices in­sti­tu­tions in ar­eas of Master Data Man­age­ment with focus on new age fron­tiers like Blockchain, IOT and Big Data, which are dis­rupt­ing the very fab­ric of eco­nom­ics, says it must be made manda­tory for pro­fes­sion­als and man­agers to un­der­stand the me­chan­ics, virtues and on­go­ing oper­a­tions of in­sti­tut­ing data gov­er­nance within an or­ga­ni­za­tion.

“The ob­jec­tive of data gov­er­nance is pred­i­cated on the de­sire to as­sess and man­age the risks that lies hid­den within the en­ter­prise in­for­ma­tion port­fo­lio,” says Jain, main­tain­ing: “One of the crit­i­cal val­ues ob­tained via data gov­er­nance is to en­sure there ex­ists only one ver­sion of the ‘Truth’ for cus­tomer or prod­uct or any crit­i­cal data do­main and this is the com­mon point of ref­er­ence across de­part­ments, sys­tems and reg­u­la­tors, thus achiev­ing a 360-de­gree view of cus­tomer data di­men­sion.”

Hence, ‘value in the form of data’ flows across the en­ter­prise with­out any con­tam­i­na­tion or loss of in­for­ma­tion, he says, adding a ‘Master Data Man­age­ment (MDM)’ pro­gram as part of data gov­er­nance port­fo­lio will en­sure con­sis­tent data is pub­lished across lines of busi­ness.

DATA IS NEW OIL

He says suc­cess of MDM goes hand in hand with a well-de­fined ‘data gov­er­nance prac­tice im­ple­men­ta­tion’. He also main­tains that or­ga­ni­za­tions have in the past decade un­der­stood the im­por­tance of data as an as­set and hence are not cat­e­go­riz­ing data man­age­ment-re­lated pro­grams as just an­other cost cen­ter ini­tia­tive. Rather, they are be­ing pru­dent by giv­ing these ini­tia­tives the ut­most at­ten­tion and in­vest­ment.

“Data (cus­tomer and master data) qual­ity is­sues, which led to the 2008 fi­nan­cial cri­sis, have forced reg­u­la­tors to is­sue man­dates via ac­cord changes from BASEL com­mit­tee for the global bank­ing and fi­nan­cial sec­tor. Even to­day, or­ga­ni­za­tions are strug­gling to ad­dress the above,” says he.

A simple de­pic­tion of data gov­er­nance lead­ing to MDM as part of a data strat­egy jour­ney ac­cord­ing to him will be:

3 PIL­LARS

Ac­cord­ing to him, data strat­egy means:

• Align­ing busi­ness strat­egy and en­sur­ing busi­ness users, own­ers & stake­hold­ers are en­gaged right from the be­gin­ning of the jour­ney

• Align­ing with cor­po­rate ob­jec­tives and en­sur­ing al­lo­ca­tion of re­sources, gain­ing CXO ex­ec­u­tives buy-in and guid­ing prin­ci­ples for achiev­ing qual­ity data busi­ness as­set

• Since not all data is equal, it is very im­por­tant to as­sess how, and which of the data will be ex­actly used and be the spe­cific part of the data gov­er­nance jour­ney

Data Pol­icy & Stew­ard­ship en­cap­su­lates:

• Data pol­icy, def­i­ni­tions, con­trols, meta­data, own­er­ship via a stew­ard are pre-req­ui­site

• There can be more than one data stew­ard for an at­tribute or set of at­tributes- clear roles & re­spon­si­bil­i­ties

• Data pro­fil­ing will be the core de­liv­er­able in this ex­er­cise wherein fac­tors such as busi­ness rules, val­ues, ex­cep­tions, au­dit rules, own­ers, au­thor­ity, amend­ment fre­quency, reg­u­la­tory and in-coun­try re­quire­ments, etc, are de­fined and agreed Data Qual­ity Main­te­nance (DQM) & Frame­work com­prises:

• Mon­i­tor­ing and post-event stage of the data lifecycle man­age­ment will in­clude sys­tem, busi­ness process and qual­ity val­i­da­tions. In- spite of strin­gent checks, data qual­ity can take a hit due to var­i­ous other un­fore­seen rea­sons or hu­man er­rors

• DQM is an evolv­ing layer wherein over a pe­riod of 1-2 years on an av­er­age the frame­work of data qual­ity ma­tures, and this is when ROI of the en­tire data gov­er­nance will sur­face.

CE­MENT THAT BONDS

Jain ex­plains that MDM is the ce­ment that bonds the en­ter­prise wide busi­ness in­tel­li­gence, work­flow and an­a­lyt­i­cal sys­tems to that of the core op­er­a­tional side of the busi­ness sys­tems and pro­cesses. He lists the causes of the 2008 cri­sis as: • Fail­ure of man­age­ments to im­ple­ment a strin­gent data strat­egy or gov­er­nance and lack of hav­ing any feed­back loop to val­i­date the data qual­ity is­sues • Reg­u­la­tors should have en­forced data gov­er­nance on en­ter­prise scale as a man­date as the bank­ing sec­tor was launch­ing very com­plex struc­tured prod­ucts with­out the un­der­ly­ing clean data in place

• Pres­ence of mul­ti­ple ver­sions of truth of cus­tomer data led to sub-stan­dard re­port­ing of fi­nan­cials

Be­sides, reg­u­la­tors could not nail these loop­holes or pit­falls due to lack of strin­gent tools and feed­back mech­a­nism to alert the in­sti­tu­tions.

Jain be­lieves that new age tech­nolo­gies can help al­le­vi­ate the data qual­ity pains. “How­ever, tech­nol­ogy can only do what it has been told to. Hence, a wa­ter­tight data gov­er­nance frame­work will only help Jain dis­cusses the im­pact of not im­ple­ment­ing data gov­er­nance in an or­ga­ni­za­tion:

Data lifecycle has key stages where data col­la­tion-consumption happens and at every stage the er­ror cor­rec­tion cost mul­ti­plies 5-10-fold in terms of cost (num­bers are for a global or­ga­ni­za­tion).

The stage 1 is data cor­rec­tion af­ter ini­tial data in­put. Data gets into the en­ter­prise’s in­ter­nal sys­tems used by the busi­ness com­mu­nity to sup­port busi­ness oper­a­tions. This set of in­gested data with is­sues gets pro­cessed and used in many busi­ness sup­port­ing sys­tems and of­ten used as part of de­ci­sion mak­ing and busi­ness sup­port­ing mod­els and cal­cu­la­tions and across line of busi­nesses.

The stage 3 is where data cor­rec­tion is done af­ter be­ing con­sumed en­ter­prise-wide and pub­lished ex­ter­nally (reg­u­la­tors etc) Data or in­for­ma­tion has hit the stage of no re­turn mean­ing the in­cor­rect data has been used across the en­ter­prise and also pub­lished to ex­ter­nal en­ti­ties.

This leads to se­ri­ous is­sues like in­cor­rect ac­count­ing state­ments, data im­pact­ing rev­enue pre­dic­tions, cost al­lo­ca­tion, etc, and cre­at­ing im­pact on cap­i­tal ad­e­quacy that could end in greater cap­i­tal in­gest.

the or­ga­ni­za­tions achieve near pris­tine data qual­ity. For ex­am­ple, blockchain tech­nol­ogy pos­si­bly can help al­le­vi­ate the doubts of in­sider fudg­ing or false re­port­ing or for that mat­ter, by­pass­ing reg­u­la­tions. MDM on a blockchain can en­sure that the en­tire spec­trum of data qual­ity checks is in place if the data strat­egy, pol­icy, ste­wards, qual­ity frame­work are done with ab­so­lute dili­gence,” says he.

CXO LEVEL IN­TER­VEN­TION

Jain says the fi­nal take-home is that given

the ad­vent of Big Data, fin­tech, blockchain and AI data gov­er­nance and master data man­age­ment should be manda­tory across busi­ness do­mains. He ex­plains: “These have be­come the core build­ing blocks, en­abling an or­ga­ni­za­tion’s pur­suit to in­no­vate, sur­vive and grow. Ul­ti­mately, the suc­cess of the above is solely de­pen­dent on the level of CXO in­volve­ment and com­mit­ment – ei­ther it can lead the or­ga­ni­za­tion to achieve near pris­tine data or end up with a tac­ti­cal silo im­ple­men­ta­tion. Data­dios!!!”

Data is not con­sumed by any ma­jor sys­tems and the cost of the cor­rec­tion is min­i­mal, but the is­sue is iden­ti­fy­ing these er­rors is near im­pos­si­ble with­out data gov­er­nance in place.

The stage 2 is data cor­rec­tion af­ter be­ing con­sumed by sys­tems.

Satish Jain avers suc­cess of Master Data Man­age­ment goes hand in hand with a well-de­fined data gov­er­nance prac­tice im­ple­men­ta­tion

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.