In the age of im­me­di­acy and ac­tion, are we miss­ing the point when it comes to us­ing data ef­fec­tively?

Campaign Middle East - - PARTNER CONTENT - By Farah Moum­neh Farah Moum­neh is head of an­a­lyt­ics and strat­egy at OMD

Do you re­mem­ber what you were do­ing on Jan­uary 15, 2009? If you were a man named Ch­es­ley “Sully” Sul­len­berger, then you would have been crash-land­ing a plane onto the Hud­son River, sav­ing the lives of every sin­gle one of the 155 pas­sen­gers aboard your air­craft.

It was an ex­tra­or­di­nary feat, and rightly played out as the heroic sur­vival tale in the me­dia. How­ever it was what came af­ter which re­ally struck a chord; the way in which Sully had plot­ted his course of ac­tion. In­stead of al­low­ing data to com­pletely drive and ex­e­cute his de­ci­sions, he used it to frame his think­ing, thus al­low­ing his in­tu­ition and decades of ex­pe­ri­ence to de­ter­mine the right con­clu­sion.

This ac­count re­mains one of the best ex­am­ples when it comes to demon­strat­ing the cor­re­la­tion be­tween hu­man learn­ing and gen­uine in­tel­li­gence in us­ing data. It’s true that we have be­come ac­cus­tomed to plac­ing the value of data above our own in­stincts at times, con­sid­er­ing it a more re­li­able re­sult than one pow­ered by emo­tion alone. How­ever, in today’s dig­i­tal era of fake news, trans­parency con­cerns and brand fraud, we’re com­ing full cir­cle. Hu­mans, it would seem, are in vogue once again.

There’s long been an ar­gu­ment that data is killing cre­ativ­ity. How­ever, it shouldn’t be about pitch­ing them against one an­other; more about us­ing each of their strengths to turn the com­bined ben­e­fits into so­lu­tions for clients. We’ve high­lighted be­fore how the hu­man el­e­ment is a key com­po­nent when it comes to brand strat­egy and shouldn’t be un­der­es­ti­mated, es­pe­cially given the dis­rup­tive ter­rain we find our­selves op­er­at­ing in today.

As bud­gets shrink, the temp­ta­tion to fo­cus on the lower end of the fun­nel is ev­i­dent, re­ly­ing on more tan­gi­ble data, such as trans­ac­tional and au­di­ence data, to make de­ci­sions. As a re­sult, brand and per­for­mance teams of­ten find them­selves at op­pos­ing ends of the spec­trum, strain­ing to work to­wards the same goal.

When it comes to us­ing data ef­fi­ciently, hu­man in­ter­pre­ta­tion can com­pletely change the out­come and the de­ci­sion it drives. Be­low are four key in­sights into how hu­man in­tel­li­gence paired with ef­fec­tive data in­sight and man­age­ment can un­lock your per­for­mance.

1. Scale doesn’t al­ways mean depth:

In today’s econ­omy, data has be­come the most im­por­tant cur­rency and we’ve built an un­spo­ken ‘data hi­er­ar­chy’, favour­ing data with scal­a­bil­ity as the most ac­tion­able and re­li­able op­tion. Par­tic­u­larly in times of dis­rup­tion, there are cer­tain fac­tors that play out pub­licly and out­weigh our nor­mal ra­tio­nal mind. Take Bit­coin, for ex­am­ple. It has gained no­to­ri­ety sim­ply be­cause we have let it. Driven by sen­ti­ment and spec­u­la­tion, trad­ing tech­nolo­gies have even be­gun to pre­dict Bit­coin fluc­tu­a­tion us­ing Twit­ter sen­ti­ment as a bench­mark. In this case, it mar­ried the best of both worlds: depth with an in­her­ently hu­man ac­count of sen­ti­ment, and scale by the sheer amount of data points these mod­els used.

2. All data is miss­ing some­thing:

The full pic­ture is never ev­i­dent from the data alone, which is why we must ac­knowl­edge its short­com­ings – all three of them to be ex­act. The first looks at data bias and how we must ex­er­cise cau­tion with the al­go­rithms we pro­gram and the data they in­gest; noth­ing is in­fal­li­ble. Con­sider the pre­dic­tive polic­ing tech­nol­ogy we’ve seen tested by the US gov­ern­ment in re­cent years. It has courted its fair share of con­tro­versy, and for good rea­son, as it has in­creased racial pro­fil­ing since its im­ple­men­ta­tion. The sec­ond con­sid­ers data blind spots, such as mis­at­tri­bu­tion and the trap of ad­dress­abil­ity, an over­sight that can hit a brand’s bot­tom line. Fi­nally, we need to be aware of data blam­ing. When some­thing fails, we blame the in­ac­cu­racy of the data rather than the in­ter­pre­ta­tion or ac­tual frame­work. We’re con­di­tioned to op­ti­mise and tweak a cam­paign’s per­for­mance un­til it aligns with our pro­jec­tions, and if it doesn’t work we’re at a loss as to why. Herein lies the prob­lem; we haven’t asked this ques­tion from the out­set, or con­sid­ered other fac­tors that could have ren­dered it a no-go from the start. We have wasted time, en­ergy and re­sources, all be­cause we have fo­cused on de­scrip­tive an­a­lyt­ics first.

3. Data should be rooted in hu­man truth:

Data is mean­ing­less on its own. How­ever gran­u­lar we go, there still needs to be a hu­man el­e­ment present to in­ter­pret the best course of ac­tion for a busi­ness. Net­flix changed the con­tent game with its (scar­ily ac­cu­rate) in­sight into hu­man be­hav­iour, iden­ti­fy­ing our in­nate com­pul­sion for binge watch­ing and turn­ing it into a bil­lion-dol­lar busi­ness model. Uber, in a sim­i­lar vein, has been able to use data to cap­i­talise on hu­man emo­tions, us­ing this in­sight to al­ter the pric­ing struc­ture dur­ing key pe­ri­ods.

4. ‘Cul­ture eats strat­egy for lunch’:

The cul­ture we need to work on is a cul­ture of in­te­gra­tion. The truth is you could have the most in­cred­i­ble strat­egy in the world, but it means noth­ing if you don’t have the right peo­ple in place to im­ple­ment it. To tackle this – and it’s rad­i­cal, I grant you – we need to kill all agen­das. In­stead we should build mean­ing­ful and sin­gu­lar ob­jec­tives and work more closely with clients to find a more co­he­sive way for­ward in fu­ture.

Newspapers in English

Newspapers from UAE

© PressReader. All rights reserved.