Big Data Mine for rev­enue

As voice rev­enues be­gin to plateau, telcos are in­creas­ingly look­ing at big data an­a­lyt­ics to pro­vide bet­ter cus­tomer ser­vices. Voice & Data takes a look at how tele­cos are lever­ag­ing big data

Voice&Data - - CONTENT - Kr­ishna Mukher­jee x-kr­ish­nam@cy­ber­me­

‘ Data, data ev­ery­where… and ev­ery bit to be mined’ – That ’s the rhetoric of ev­ery ser­vice provider to­day. Be it struc­tured or un­struc­tured data, telcos are grad­u­ally, un­der­stand­ing the prow­ess of data and the role of big data an­a­lyt­ics in im­prov­ing their rev­enue mar­gins.

While few years back, dis­counts, cash back of­fers were the strate­gies em­ployed by telcos to woo their cus­tomers, th­ese days with voice rev­enues about to reach sat­u­ra­tion point; com­pa­nies are in­creas­ingly look­ing at of­fer­ing per­son­al­ized ser­vices with cus­tomer re­ten­tion plans, among oth­ers.

Ac­cord­ing to an­a­lyst firm, In­for­ma­tion Data Corp (IDC), over 2.7 zetabytes of data ex­ist in the dig­i­tal world to­day and 90% of all data has been cre­ated in the last two years. The tele­com in­dus­try gen­er­ates struc­tured, semi-struc­tured and un­struc­tured data from Call Logs, SMS logs, N/w log, Web Logs and so on… Min­ing th­ese logs and Call De­tail Records (CDRs) is help­ing tele­cos cap­ture and un­der­stand key cus­tomer at­tributes.

Tele­com op­er­a­tor Tata Te­le­ser­vices has been us­ing big data so­lu­tions for ad­dress­ing le­gal and reg­u­la­tory re­quire­ments which are man­dated for longer re­ten­tion pe­ri­ods. Given the grow­ing vol­ume of data and time sen­si­tiv­ity at­tached to such re­quire­ments or­ga­ni­za­tions usu­ally look to pro­cure ad­di­tional stor­age

Tele­com op­er­a­tor Tata Te­le­ser­vices has been us­ing big data so­lu­tions for ad­dress­ing le­gal and reg­u­la­tory re­quire­ments which are man­dated for longer re­ten­tion pe­ri­ods.

and servers in­cur­ring high ex­pen­di­ture.

As an al­ter­na­tive, big data so­lu­tions are de­ployed which are cost ef­fec­tive in pro­vid­ing timely in­for­ma­tion to le­gal and reg­u­la­tory teams. “While there are lot of open-source tools/so­lu­tions avail­able, to meet the re­quire­ment and given the sen­si­tiv­ity of the data, Tata Te­le­ser­vices de­ployed En­ter­prise Hadoop So­lu­tion backed by nec­es­sary se­cu­rity and support,” says Srini­vas Ral­la­bandi, head, business in­tel­li­gence, Tata Te­le­ser­vices.

Another tele­com ser­vice provider Rus­sian firm MTS is in­te­grat­ing and an­a­lyz­ing

in­for­ma­tion from call de­tail records of its cus­tomers and other company data sources on an En­ter­prise Data Ware­house (EDW) plat­form, which is help­ing it to track cus­tomers who are likely to re­duce their use of MTS.

Be­sides, by an­a­lyz­ing so­cial net­work topol­ogy, MTS is now able to iden­tify opin­ion lead­ers and in­flu­encers. MTS uses this in­for­ma­tion to support vi­ral mar­ket­ing cam­paigns to sell value-added ser­vices. To ex­e­cute the tar­get mar­ket­ing cam­paigns, the company uses the SAS Mar­ket­ing Au­to­ma­tion so­lu­tion which is in­te­grated with the Ter­a­data Ac­tive EDW and so­cial net­work anal­y­sis has proven to be more ac­cu­rate than a tra­di­tional pre­dic­tive mod­el­ing ap­proach.

How­ever, Ra­jeev Ba­tra, CIO, MTS, be­lieves op­er­a­tors are in the dis­cov­ery phase of big data. “While an­a­lyt­ics is be­ing done ex­ten­sively by all of them but big data is not used in a ma­jor way but only in bits and pieces I think, in a year or two we would see a much more big data adop­tion in the sys­tem,” says Ba­tra.

Key Fo­cus Ar­eas

In­dus­try ex­perts are of the view that the tele­com in­dus­try wants so­lu­tions around cus­tomer-ex­pe­ri­ence man­age­ment, net­work an­a­lyt­ics, churn re­duc­tion, cross sell and up­sell.

Telcos can de­ploy big data so­lu­tion to ben­e­fit business and cus­tomers like mar­ket­ing, cus­tomer in­ter­ac­tions/ ex­pe­ri­ence man­age­ment, net­work man­age­ment lead­ing to in­crease in prof­itabil­ity. An up­trend can be seen in the de­ploy­ment and adop­tion of big data an­a­lyt­ics for business op­er­a­tions, pri­mar­ily in cus­tomer ex­pe­ri­ence man­age­ment where there is a need to in­te­grate struc­tured (tra­di­tional) and semi-/un-struc­tured data points. Ac­cord­ing to en­ter­prise so­lu­tions provider IBM, telcos are go­ing for two main as­pects of big data an­a­lyt­ics for their business growth –cus­tomer an­a­lyt­ics and trans­form­ing op­er­a­tion.

The company has in­vested more than $1 bil­lion in ad­vanc­ing new tele­com of­fer­ings in­clud­ing telco-re­lated ac­qui­si­tions. It has signed agree­ments with telcos such as Bharti, Voda­fone and Idea. For

Am­docs Big Data An­a­lyt­ics propo­si­tion is a holis­tic tele­com spe­cific ap­proach com­pris­ing Data Sci­ence As a Ser­vice, Ac­tion­able An­a­lyt­ics Ap­pli­ca­tions and Data Man­age­ment and Mod­ern­iza­tion. —Hadas Haran prod­uct mar­ket­ing di­rec­tor, Am­docs Big Data and An­a­lyt­ics Cou­pled with big data of­fer­ings of op­ti­mized ap­pli­ances, IBM Cog­nos Business In­tel­li­gence soft­ware com­bines op­er­a­tional and fi­nan­cial data from across the or­ga­ni­za­tion into a sin­gle source of in­for­ma­tion for re­port­ing, anal­y­sis, dash­boards and score­cards to help drive bet­ter business de­ci­sions. —Asheet Makhija coun­try leader, in­for­ma­tion man­age­ment IBM In­dia/South Asia

Voda­fone, IBM In­dia man­ages IT ser­vices, which in­clude business in­tel­li­gence, billing, and fi­nan­cial sys­tems. Be­sides, IBM has en­abled Idea to ac­cel­er­ate time-tomar­ket of new ser­vices and em­power business with an­a­lyt­ics tools, thereby im­prov­ing end-to-end cus­tomer man­age­ment and pro­vide new rev­enue streams.

“Cou­pled with big data of­fer­ings of op­ti­mized ap­pli­ances (Pure Data for An­a­lyt­ics and Pure Data for Op­er­a­tional An­a­lyt­ics), IBM Cog­nos Business In­tel­li­gence soft­ware com­bines op­er­a­tional and fi­nan­cial data from across the or­ga­ni­za­tion into a sin­gle source of in­for­ma­tion for re­port­ing, anal­y­sis, dash­boards and score­cards to help drive bet­ter business de­ci­sions,” says Asheet Makhija, coun­try leader - in­for­ma­tion man­age­ment, IBM In­dia/South Asia.

Be­sides, IBM SPSS soft­ware en­cap­su­lates ad­vanced math­e­mat­i­cal and sta­tis­ti­cal ex­per­tise to ex­tract pre­dic­tive knowl­edge that when de­ployed helps im­prove out­comes, he adds.

How­ever, ac­cord­ing to Su­nil Jose, man­ag­ing di­rec­tor, Ter­a­data In­dia: “Net­work is def­i­nitely an im­por­tant fo­cus area for CSPs, there are three broad ar­eas where big data can be used in net­work an­a­lyt­ics—ef­fi­cient roll out of the net­work which is the fo­cus for CTOs, in­creas­ing rev­enue – fo­cus for sales and mar­ket­ing teams and, im­prov­ing over­all cus­tomer ex­pe­ri­ence and brand build­ing.

Teoco di­rec­tor and In­dia coun­try man­ager Srini­vas Bhogle talks about a prod­uct called ser­vice as­sur­ance that gauges the net­work of dif­fer­ent op­er­a­tors and if there is any prob­lem it would raise the alarm and pro­vide so­lu­tion for the same.

Though most ser­vice providers are not start­ing en­tirely from scratch, hav­ing

al­ready de­vel­oped data ware­houses and re­lated business in­tel­li­gence (BI) so­lu­tions, most re­al­ize that big data an­a­lyt­ics re­quire a dif­fer­ent in­fra­struc­ture than what they have used his­tor­i­cally for data ware­hous­ing and BI. Many or­ga­ni­za­tions, there­fore, plan to invest in new so­lu­tion in­fra­struc­ture to re­al­ize the prom­ise of big data. En­ter­prise Survey Group (ESG) es­ti­mates that more than half of larger or­ga­ni­za­tions will make such in­vest­ments.

Based on ESG’s mod­el­ling of a medi­um­sized Hadoop-ori­ented big data project, the pre­con­fig­ured Or­a­cle big data ap­pli­ance is 39% less costly than a “build” equiv­a­lent doit-your­self in­fra­struc­ture. And us­ing Or­a­cle big data ap­pli­ance will cut the project length by about one-third, the company claims.

Or­a­cle Big Data Ap­pli­ance, in con­junc­tion with Or­a­cle Ex­a­data, and Or­a­cle Ex­a­lyt­ics, helps cus­tomers to ac­quire and or­ga­nize di­verse data types, and then an­a­lyze them along­side ex­ist­ing en­ter­prise data to dis­cover new in­sight.

IMI­mo­bile, which op­er­ates in over 60 coun­tries with over 100 op­er­a­tors, is in­vest­ing on a prod­uct called pro­file man­ager, which works in close con­junc­tion with another prod­uct called cam­paign man­ager. Pro­file man­ager is a way of ag­gre­gat­ing cus­tomer data from mul­ti­ple sources at one place with­out mak­ing changes to ex­ist­ing sys­tems. So, there is no re­place­ment to the ex­ist­ing sys­tems but it works along with what it al­ready has.

“Be­sides, we have chunk pre­dic­tion to pre­dict the prob­a­bil­ity of churn of given set of cus­tomer and if the model says that there is a high risk of churn­ing then some re­ten­tion ac­tiv­ity can be un­der­taken for the tar­geted group of peo­ple,” says Su­dar­shan Dharma­puri, vice-pres­i­dent, prod­uct man­age­ment, IMI­mo­bile.

He ex­plains that while in In­dia the an­a­lyt­ics tools are just be­ing used for re­tain­ing cus­tomers and of­fer­ing bet­ter so­lu­tions, in ma­ture mar­kets big data an­a­lyt­ics are used in a big­ger way like –in an­a­lyz­ing the health pat­tern and in con­tain­ing epi­demics.

Hadas Haran, prod­uct mar­ket­ing di­rec­tor for Am­docs Big Data and An­a­lyt­ics, elab­o­rates on his prod­uct. He says, “Am­docs Big Data An­a­lyt­ics propo­si­tion is a holis­tic tele­com spe­cific ap­proach which in­cludes Data Sci­ence As a Ser­vice. SPs strug­gle to ex­tract value from their data es­pe­cially when large data sets are in­volved. Com­bin­ing do­main ex­per­tise around net­work, mar­ket­ing, care and op­er­a­tions with data sci­ence, Am­docs DSaaS pro­vides a valu­able kick start to ser­vice providers by per­form­ing con­sul­tancy data dis­cov­ery and sit­u­a­tion spe­cific an­a­lyt­i­cal ap­pli­ca­tion de­vel­op­ment.”


Ex­perts be­lieve qual­ity is the big­gest chal­lenge as data is not at one place and the same data tells you dif­fer­ent sto­ries, so ra­tio­nal­iz­ing data is one of the key con­cerns. Ac­cord­ing to Satyakam Mo­hant, di­rec­tor and CEO, Ma Foi An­a­lyt­ics, “The chal­lenge lies in ‘va­ri­ety’ not vol­ume or ve­loc­ity. If big data had been only about the amount of data be­ing gen­er­ated or the pace at which it is be­ing gen­er­ated, one just needed to scale up ex­ist­ing data stor­age and pro­cess­ing sys­tems to take care of it.”

The main chal­lenge is in in­te­grat­ing the rel­e­vant data which ex­ists in the telco ecosys­tem but is dis­trib­uted across teams, lo­ca­tions and for­mats says VVR Kishore, head – In­dia op­er­a­tions at Mo­bileum, which has over 100 an­a­lyt­ics en­gage­ments cur­rently in In­dia and across the world.

Big Data de­ploy­ments or rather en­gage­ments typ­i­cally start with “proofof-con­cepts”. The ven­dor and telco brain­storm on cur­rent business chal­lenges and fi­nal­ize a few use cases or ex­am­ples of how cer­tain types of data can be an­a­lyzed to de­liver richer in­sights. The ven­dor then works with the op­er­a­tor to build the use case out and show in­terim re­sults. If th­ese are found to be com­pelling, the telco com­mis­sions a full-fledged de­ploy­ment where there is a big data plat­form in­stalled and in­te­grated with mul­ti­ple OSS/BSS data sources and the ven­dor be­gins work on im­ple­ment­ing a set of use cases for the telco.

“Some of the common bar­ri­ers to big data an­a­lyt­ics are the same as one would ex­pect with any new trans­for­ma­tional con­cept. First, there are or­ga­ni­za­tional bar­ri­ers such as – Who owns it? Is it IT/ CIO, business/mar­ket­ing, or a brand new ‘depart­ment’ for big data an­a­lyt­ics? How do we ramp up the skills to de­sign, es­tab­lish and run func­tional big data an­a­lyt­ics? Do we need to en­gage ex­pen­sive con­sul­tants?,” points out Ral­la­bandi.

Then, there are the in­evitable RoI wor­ries – spon­sor­ship, business case, TCO. There is also the worry that the cur­rent data base is too messy, or un­struc­tured, to lend it­self to big data an­a­lyt­ics. The fact is, this is not a lim­i­ta­tion for big data an­a­lyt­ics at all, he adds. One just needs to get started.

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.