Why predictive analytics could be a crystal ball for businesses

With ram­pant infla­tion increas­ing their costs and reces­sion fears damp­en­ing demand for their goods and ser­vices, firms nation­wide are busy cut­ting unnec­es­sary spend­ing and seek­ing new growth oppor­tu­ni­ties. 

Deter­min­ing where, when and how much to spend is always impor­tant, of course, but it’s cru­cial in a down­turn, when such choic­es can have a huge impact on a business’s medi­um-term growth prospects once the econ­o­my recov­ers. That’s why, in search of ways to refine their invest­ment deci­sions, firms are increas­ing­ly using pre­dic­tive ana­lyt­ics to help them weigh up all the poten­tial oppor­tu­ni­ties and threats. 

“Risk man­age­ment is nev­er far from dis­cus­sions among CFOs and reg­u­la­to­ry teams,” says James Pet­ter, vice-pres­i­dent and gen­er­al man­ag­er, inter­na­tion­al, at Pure Stor­age, a data stor­age firm which helps organ­i­sa­tions grow through dig­i­tal trans­for­ma­tion. “But, in the eco­nom­ic cli­mate of 2023, every senior leader in every organ­i­sa­tion will have risk man­age­ment front of mind. They’ll be mak­ing a deep assess­ment of the eco­nom­ics with­in com­pa­nies, their finan­cial struc­tures and tech­nolo­gies. 

He con­tin­ues: “Every com­pa­ny has a wealth of data and most are try­ing to do some­thing with it. But often they’re focused on under­stand­ing the cur­rent mar­ket con­di­tions and react­ing to these. I believe that there’ll be more of a push to look ahead as part of the over­all focus on risk man­age­ment. Pre­dic­tive ana­lyt­ics will play a big part in this.”

The rise of pre­dic­tive ana­lyt­ics comes as no sur­prise to Shankar Bal­akr­ish­nan, vice-pres­i­dent for north­ern Europe at soft­ware devel­op­er Ana­plan. He likens firms that rely on his­tor­i­cal data alone to nav­i­gate in such tough con­di­tions to a dri­ver steer­ing their car accord­ing to what they can see in its rear-view mir­ror. Instead, Bal­akr­ish­nan argues, they need access to more data sources to mod­el poten­tial future out­comes and so react more smart­ly to dis­rup­tive events.  

Ana­plan recent­ly worked with the South Cen­tral Ambu­lance Ser­vice Foun­da­tion NHS Trust (SCAS), which cov­ers Berk­shire, Buck­ing­hamshire, Hamp­shire and Oxford­shire, to help it devel­op a pre­dic­tive capa­bil­i­ty. Apply­ing machine learn­ing and pre­dic­tive insights to exist­ing data, Ana­plan was able to fore­cast the num­ber of emer­gency calls the SCAS ambu­lance teams would receive at any giv­en point. This has enabled the trust to deploy its resources more effi­cient­ly. 

The implementation headache

But what’s the best way to imple­ment this pow­er­ful AI-based tech­nol­o­gy? “For finance chiefs, the chal­lenge is to under­stand where to focus,” says Simon Edwards, CFO at soft­ware devel­op­er Ser­vice­Max.

One good place to start, he sug­gests, might be automat­ing func­tions in the back office. Using tech such as robot­ic process automa­tion and AI-enabled data ana­lyt­ics not only helps to improve rou­tine process­es, cov­er skills gaps and increase effi­cien­cies; it can also pro­vide intel­li­gence that can be fed into fore­cast­ing and plan­ning. What’s more, this kind of automa­tion will also enable staff to focus on more val­ue-adding tasks, he sug­gests.

Giv­en that the com­mer­cial envi­ron­ment is awash in risk and uncer­tain­ty, few lead­ers will want to trust impor­tant resourc­ing and invest­ment deci­sions to gut instinct. Indeed, risk man­age­ment may be the top pri­or­i­ty in times of cri­sis, but what if busi­ness lead­ers could avoid the cri­sis in the first place?

Whether you’re fac­ing a pan­dem­ic, a nat­ur­al dis­as­ter or a ran­somware attack, mak­ing effec­tive choic­es under pres­sure requires accu­rate and time­ly data-dri­ven insights, notes Alan Jacob­son, chief ana­lyt­ics offi­cer at data sci­ence com­pa­ny Alteryx. 

“Suc­cess­ful risk man­age­ment requires data as the course cor­rec­tor, giv­ing you the abil­i­ty to mod­el dif­fer­ent sce­nar­ios,” he says. 

Jacob­son points to trav­el and tourism as indus­tries that are bank­ing on pre­dic­tive ana­lyt­ics to help them recov­er ful­ly from the huge­ly dam­ag­ing dis­rup­tion they have suf­fered in recent years as a result of the Covid cri­sis. Air­craft man­u­fac­tur­ers are using the tech­nol­o­gy to deter­mine the most effec­tive times to per­form var­i­ous main­te­nance tasks. Air­lines are using sim­i­lar sys­tems to pre­dict demand for par­tic­u­lar flights and plan their staffing and fuelling require­ments accord­ing­ly to improve oper­a­tional effi­cien­cy and min­imise dis­rup­tion.

“Qual­i­ty data and pre­dic­tive ana­lyt­ics are also inte­gral to risk mit­i­ga­tion across the finan­cial ser­vices indus­try,” he adds. “They are invalu­able for fraud detec­tion, audit inves­ti­ga­tions and oth­er types of advanced work.” 

Accuracy matters

Of course, the suc­cess of such efforts hinges on the stan­dard of the data fed into the sys­tem. Insights based on faulty or incom­plete inputs could mis­lead deci­sion-mak­ers and poten­tial­ly cause sig­nif­i­cant harm to a busi­ness.

“Imple­ment­ing pre­dic­tive ana­lyt­ics isn’t a case of ‘run once and for­get’. It will take time and effort to analyse the find­ings, under­stand them and then tweak the pro­gram accord­ing­ly,” Pet­ter stress­es. “The risk would be imple­ment­ing a big pro­gram that doesn’t deliv­er the insights need­ed to help the organ­i­sa­tion. It’s impor­tant to have clear goals when imple­ment­ing, adjust as need­ed and con­stant­ly refo­cus to ensure that the busi­ness is get­ting what it needs.”

Indeed, numer­ous data prob­lems can lurk beneath the sur­face, espe­cial­ly if users are inex­pe­ri­enced in han­dling the out­puts gen­er­at­ed.

“Accu­ra­cy and com­pat­i­bil­i­ty are para­mount when it comes to mea­sur­ing per­for­mance across dif­fer­ent depart­ments,” Edwards says. “It’s a com­mon prob­lem that needs address­ing before it does mate­r­i­al harm to a busi­ness.”

Bal­akr­ish­nan agrees. “If lead­ers are work­ing with inac­cu­rate data, they risk mak­ing inac­cu­rate deci­sions,” he says. “At the same time, if teams have to spend hours vet­ting and val­i­dat­ing data, that makes it impos­si­ble for deci­sion-mak­ers to react at speed.” 

Despite the effort involved in get­ting pre­dic­tive ana­lyt­ics up and run­ning prop­er­ly, the ben­e­fits are obvi­ous to Pet­ter. 

“I don’t think 2023 will be the year to leave any kind of chink in your cor­po­rate armour,” he says. “What pre­dic­tive ana­lyt­ics enables is valu­able to busi­ness lead­ers. With the insights it deliv­ers, this tech­nol­o­gy has huge poten­tial to turn data into gold.”