AMARPREET KALKAT on the im­pact of so­cial in­tel­li­gence on busi­nesses.

Today, the num­ber of peo­ple on so­cial me­dia is ris­ing ev­ery sec­ond and the vol­ume of con­ver­sa­tions is es­ca­lat­ing rapidly. So­cial in­tel­li­gence is the tool of choice for busi­nesses to de­ci­pher the di­verse data—to iden­tify op­por­tu­ni­ties and avoid blun­ders.

The Smart Manager - - Front Page - AMARPREET KALKAT IS CEO OF FR­ROLE INC.

So­cial in­tel­li­gence is the prac­tice of gath­er­ing data from so­cial me­dia chan­nels and an­a­lyz­ing it to make crit­i­cal busi­ness de­ci­sions. Its com­mon use is to mine cus­tomer sen­ti­ments, in­ter­ests, and con­ver­sa­tions in or­der to sup­port mar­ket­ing and busi­ness op­er­a­tions. Al­though so­cial in­tel­li­gence is a rel­a­tively un­ex­plored ter­ri­tory, it is grad­u­ally emerg­ing as a key fa­cil­i­ta­tor of change. It is com­pelling mar­keters to trade gut feel­ings with ac­cu­rate data-driven in­sights to de­vise smarter strate­gies, and achieve ef­fec­tive proac­tive de­ci­sion-mak­ing. It is about the mar­keters’ will­ing­ness to be a part of so­cial con­ver­sa­tions and not merely as­sem­ble so­cial data. This wealth of in­for­ma­tion has much to of­fer—to an ar­ray of seg­ments in­clud­ing con­sumer tech­nol­ogy brands, dig­i­tal mar­ket­ing and me­dia agen­cies, and me­dia com­pa­nies. Can ‘so­cial’ en­tirely dis­place con­ven­tional meth­ods of mar­ket­ing re­search and in­tel­li­gence gath­er­ing? The an­swer is a big ‘yes’. So­cial in­sights are slowly be­com­ing the most im­por­tant piece of in­for­ma­tion for C-level ex­ec­u­tives and board mem­bers who are seek­ing the best pos­si­ble ba­sis for their de­ci­sions.

brief his­tory

While the ear­li­est so­cial lis­ten­ing prod­ucts can be traced back to 2000-2002, most of the prod­ucts com­monly avail­able today be­long to the 2005-2008 pe­riod. Many of these (first gen­er­a­tion of so­cial data prod­ucts) have un­der­gone in­cre­men­tal im­prove­ments and a few have taken a mean­ing­ful next-gen­er­a­tion leap. Many of the sec­ond-gen­er­a­tion (2009-2012) prod­ucts are fo­cused on tak­ing so­cial con­tent and so­cial data to out­side in­ter­faces such as web­sites, TV, out­door dis­plays, etc. , and can be termed as the ‘so­cial en­gage­ment’ ven­dors.

The third gen­er­a­tion (from 2012 to the present), on the one hand, in­cludes so­cial alert providers such as Datam­inr and Banjo; on the other, in­cludes true so­cial in­tel­li­gence ven­dors such as IBM Wat­son and Fr­role. A small num­ber of first-gen­er­a­tion ven­dors such as Crim­son Hexagon and Brand­watch can also be listed here, given the non-lin­ear prod­uct progress be­ing made by them.

present sta­tus

The so­cial in­tel­li­gence mar­ket today stands at an in­ter­est­ing junc­ture. While it has been grow­ing at 30% an­nu­ally, and is poised to be a $5bn mar­ket by 2020,1 Gart­ner has put so­cial an­a­lyt­ics in the trough of dis­il­lu­sion­ment.

The re­al­ity prob­a­bly is some­where in the mid­dle. So­cial data con­tin­ues to an­swer more and more cus­tomer ques­tions for mar­keters and there­fore con­tin­ues to be­come more use­ful. Led by the next-gen­er­a­tion prod­ucts, the ven­dors con­tinue to build bet­ter tech­nol­ogy and al­go­rithms so that they can an­swer more ques­tions with more pre­ci­sion. So­cial in­tel­li­gence is grad­u­ally emerg­ing as a key fa­cil­i­ta­tor of change in the world of busi­ness. It is com­pelling mar­keters to trade gut feel­ing with ac­cu­rate data-driven in­sights to de­vice smarter strate­gies and achieve ef­fec­tive proac­tive de­ci­sion-mak­ing.

The na­ture of the in­sight is chang­ing. De­scrip­tive in­sights like num­ber of men­tions and sen­ti­ment are giv­ing way to ac­tion­able in­sights like what caused the change in men­tions or sen­ti­ment. While pre­dic­tive and pre­scrip­tive

in­sights are still to ma­te­ri­al­ize in the real sense of the word, the next-gen­er­a­tion ven­dors are fo­cus­ing heav­ily on it and build­ing it as a true dif­fer­en­tia­tor.

In­te­gra­tion of so­cial and non-so­cial data is be­com­ing a re­al­ity. Ven­dors such as IBM Wat­son al­low the user to im­port var­i­ous kinds of data and then build and vi­su­al­ize cor­re­la­tions be­tween them. At Fr­role, we help cus­tomers cre­ate so­lu­tions on top of the prod­uct that can com­bine var­i­ous kinds of data, and pack­age it in a man­ner that ex­actly meets their needs. A global agency has en­gaged Fr­role to build a data man­age­ment plat­form that can com­bine so­cial data with ad spend data, Google search, sales, and other mar­ket­ing ex­pense data. Back in 2013, Gart­ner lamented in their re­port, Who's Who in So­cial Me­dia An­a­lyt­ics?, “Most of the ven­dors sup­port use cases that fo­cus ex­clu­sively on the anal­y­sis of so­cial data, rather than the po­ten­tially more im­pact­ful in­sights that may be de­rived by an­a­lyz­ing both so­cial data and other en­ter­prise data.” It is fi­nally chang­ing.

There are two kinds of ven­dors pro­vid­ing peo­ple in­tel­li­gence—spe­cial­ized ven­dors such as StatSo­cial and Peo­ple Pat­tern, and full-stack ven­dors such as Brand­watch, NetBase, Crim­son Hexagon, and Fr­role which pro­vide both peo­ple in­tel­li­gence and topic in­tel­li­gence/so­cial lis­ten­ing.

next-gen­er­a­tion so­cial in­tel­li­gence

Fr­role is one such dis­rup­tive so­cial in­tel­li­gence startup. We pro­vide hard-to-ob­tain con­sumer in­sights to mar­keters and prod­uct own­ers, com­bin­ing our ex­per­tise in the area of ma­chine learn­ing with our ca­pa­bil­i­ties to an­a­lyze mil­lions of uni­ver­sal data sets in real time. For all these data sets, we per­form ex­ten­sive anal­y­sis across se­man­tic, meta­data, and sta­tis­ti­cal di­men­sions by lever­ag­ing stan­dard and cus­tom-built al­go­rithms around ML, NLP, NER, and clus­ter­ing to do this anal­y­sis. While most so­cial an­a­lyt­ics/in­tel­li­gence prod­ucts pro­vide re­sults based on statis­tics and the first level of NLP, we go deeper into build­ing se­man­tic con­text for each topic and ty­ing it up with in­for­ma­tion avail­able in the gen­eral and his­tor­i­cal data sets. The next-gen­er­a­tion so­cial in­tel­li­gence com­pa­nies are fo­cus­ing on ‘peo­ple

in­tel­li­gence.’ The beauty of this is to un­der­stand the users be­hind so­cial con­ver­sa­tions be­yond just so­cial lis­ten­ing. Hence, an­a­lyze de­mo­graph­ics, psy­cho­met­rics, brand pref­er­ences, pur­chase be­hav­ior, and con­tent affin­ity of peo­ple based on what they post on so­cial me­dia. Us­ing de­rived in­sights, it helps mar­keters to mea­sure real­time shifts in their au­di­ence char­ac­ter­is­tics, an­a­lyze past buy­ing be­hav­ior, and pre­dict fu­ture pat­terns. Also, these deep in­sights, which even in­clude ‘mood anal­y­sis’ of the tar­geted au­di­ence, can po­ten­tially serve as fod­der for crit­i­cal busi­ness de­ci­sions, both short term and long term. This is a sig­nif­i­cant next step for a brand as mood anal­y­sis can tell a brand how a con­sumer is likely to act on a cer­tain feel­ing, some­thing that brand mar­keters care for deeply.

In early 2016, one of our clients, a global smart­phone maker, used so­cial in­sights to de­rive its new prod­uct launch and go-to-mar­ket strat­egy in emerg­ing mar­kets such as In­dia. The range of an­a­lyt­i­cal tech­niques has ex­ploded, and com­pa­nies must tap new ar­eas of ex­per­tise to stay ahead of the game. A brand’s cus­tomers are speak­ing about the brand, its com­peti­tors, their likes and dis­likes, and many other in­ter­est­ing things. Why would that brand still com­pletely rely on age-old mar­ket re­search tac­tics, and not em­brace these or­ganic, un­bi­ased, multi-seg­mented, real­time in­sights about their in­dus­try, com­pe­ti­tion?

For a long time, sen­ti­ment anal­y­sis has been the stan­dard ap­proach for un­der­stand­ing how con­sumers feel about a brand. How­ever, this man­ner of anal­y­sis sticks to ag­gre­gat­ing pos­i­tive, neg­a­tive, and neu­tral sen­ti­ments; in a world of col­ors, that is like hav­ing only black, white, and grey as the avail­able col­ors. It makes for a pretty dull and il­leg­i­ble world, whether it is the real world or the world of con­sumer be­hav­ior. With mood anal­y­sis, we have in­tro­duced a next-gen­er­a­tion al­go­rithm that helps you un­der­stand how con­sumers feel about a par­tic­u­lar brand, and how they are likely to act on their feel­ings.

Sen­ti­ment anal­y­sis has been the stan­dard ap­proach for un­der­stand­ing how con­sumers feel about a brand.

the fu­ture

So­cial In­tel­li­gence in the fu­ture is no more go­ing to be about un­der­stand­ing what your cus­tomers are say­ing about you on so­cial me­dia. It is go­ing to be about tak­ing what peo­ple said on so­cial me­dia to un­der­stand your cus­tomers—whether those cus­tomers are on your web­site, in your store, or sit­ting at their home. Lead­ing ven­dors are al­ready bring­ing in tech­nol­ogy that would help you un­der­stand the de­mo­graph­ics of a par­tic­u­lar group of peo­ple who avail them­selves of a spe­cial dis­count promo, or un­der­stand the needs of a group of cus­tomers in your CRM who walked into a cer­tain store on a cer­tain day. So­cial in­tel­li­gence is on its way to be­com­ing truly in­te­grated and is no more go­ing to be lim­ited to so­cial me­dia con­ver­sa­tions.

It is also be­gin­ning to tread into the do­main pre­vi­ously owned by mar­ket­ing re­search providers. Back in 2012, a sem­i­nal re­port by McKin­sey, The So­cial Econ­omy, said, “We be­lieve 100 per­cent of the cur­rent ($31B) mar­ket­ing an­a­lyt­ics mar­ket can be cap­tured through so­cial tech­nol­ogy”.

While this has stayed as a wish­ful thought for long, ven­dors like us now fi­nally pro­vide fea­tures that al­low users to ask on-the-fly ques­tions from real-time so­cial data. You need to pro­vide a few train­ing ex­am­ples for the an­swers you want to val­i­date, and the ma­chine learn­ing al­go­rithms kick in to au­to­mat­i­cally un­der­stand enough from those ex­am­ples to start an­swer­ing any ques­tions a mar­keter might have. This oblit­er­ates the need to do ex­pen­sive, in­fre­quent, mi­cro-sam­pled fo­cus groups and con­sumer sur­veys, al­low­ing mar­ket­ing re­search to be­come both real time and real.

So­cial in­tel­li­gence is chang­ing the way we con­duct mar­ket an­a­lyt­ics, of­fer­ing re­li­able data points to build au­di­ence pro­files, seg­ments, and prod­uct in­no­va­tions. With the in­te­gra­tion of so­cial in­tel­li­gence and cus­tomer data, the con­ven­tional meth­ods of mar­ket­ing re­search and an­a­lyt­ics are be­com­ing ob­so­lete. Gone are the days of ex­pen­sive fo­cus groups and sur­veys. Now mar­keters can use so­cial data to un­der­stand what mo­ti­vates an au­di­ence and con­sol­i­date the in­tel with other in­for­ma­tion (i.e., cus­tomer data) for richer in­sights.

Cus­tomers today are more than just cus­tomers, they are the last mile for mar­keters. Us­ing so­cial me­dia, they not only in­flu­ence the buy­ing de­ci­sions of oth­ers but can also make or break a brand’s over­all per­cep­tion in the long run. For brands to be dy­nam­i­cally com­pet­i­tive, it is im­per­a­tive they know their cus­tomers in real time. In today’s ag­gres­sive mar­ket, or­ga­ni­za­tions need to cap­ture max­i­mum cus­tomer in­for­ma­tion and an­a­lyze it ef­fec­tively to dis­cover pat­terns, trends, and other vi­tal clues. So­cial me­dia net­work­ing ac­tiv­ity is gen­er­at­ing big data—and these grow­ing sources are the new fron­tier for cus­tomer in­tel­li­gence.

So­cial in­tel­li­gence in the fu­ture is no more go­ing to be about un­der­stand­ing what your cus­tomers are say­ing about you on so­cial me­dia.

hype cy­cle for dig­i­tal mar­ket­ing, 2015

Fig­ure 1 : Fr­role So­cial In­tel­li­gence Dash­board

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.