Big data tells fortunes

A crys­tal ball would be use­ful to any busi­ness want­i­ng to rev­o­lu­tionise sup­ply man­age­ment. See­ing the future could tell man­agers what cus­tomers will want next month, allow­ing sup­pli­ers to pre­pare. A lit­tle clair­voy­ance could also show how sup­pli­ers will per­form, help­ing avoid dis­rup­tion.

Tarot cards aside, there are tech­nolo­gies avail­able to busi­ness­es want­i­ng to pre­dict pat­terns in sup­ply and demand. Increas­ing­ly pow­er­ful ana­lyt­ics soft­ware can exploit a grow­ing num­ber of inter­nal and exter­nal data sources to cre­ate pre­dic­tive mod­els which can help slash sup­ply chain waste and avoid risk.

The trend is attract­ing increas­ing inter­est from sup­ply chain organ­i­sa­tions, says Ray Eit­el-Porter, man­ag­ing direc­tor, ana­lyt­ics lead, Accen­ture. “We see a lot of inter­est in big data in the sup­ply chain area,” he says. “There is a grow­ing aware­ness of the val­ue of data. You hear peo­ple talk­ing about data as the new oil and data as a strate­gic asset. This is com­ing from not just the con­sul­tants, but senior lead­ers with­in busi­ness.”

Although sup­ply chain and pro­cure­ment pro­fes­sion­als have long been accus­tomed to accu­mu­lat­ing and act­ing on data, the appli­ca­tion of ana­lyt­ics to a greater range of inter­nal and exter­nal data can cre­ate dra­mat­ic improve­ment in sup­ply chain effi­cien­cy, Mr Eit­el-Porter says.

One client sell­ing mobile phones achieved a 30 per cent improve­ment in accu­ra­cy of demand fore­casts, he says. “That has a dra­mat­ic impact on the busi­ness. It trans­lates into a reduc­tion in stock of 21 per cent. The com­pa­ny can reduce work­ing cap­i­tal with­out reduc­ing ful­fill­ment to cus­tomers.”

Using ana­lyt­ics to fore­cast demand can also help low­er sup­pli­er costs, so they can offer buy­ers bet­ter prices. Euro­pean online retail and cat­a­logue giant Otto, which owns Free­mans cat­a­logues, uses neur­al net­work tech­nol­o­gy, orig­i­nal­ly devel­oped for nuclear research at the CERN par­ti­cle accel­er­a­tor, from soft­ware firm Blue Yon­der. Vast amounts of his­tor­i­cal data detail­ing sales, prod­uct fea­tures, web­site posi­tion­ing and cat­a­logue posi­tion­ing, and exter­nal data such as weath­er, are com­bined to pro­duce pre­dic­tive mod­els that improve fore­casts.

The retail­er esti­mates it has the poten­tial to save at least €10 mil­lion each year with more accu­rate demand fore­cast­ing. It avoids under and over-stock­ing, and also allows buy­ers to cre­ate big­ger orders, cut­ting the cost of trans­port­ing sec­ondary orders and allow­ing sup­pli­ers to low­er cost. Otto gets a bet­ter price with­out dam­ag­ing sup­pli­er mar­gins, says Michael Sinn, the company’s vice pres­i­dent, cat­e­go­ry man­age­ment sup­port.

You hear peo­ple talk­ing about data as the new oil and a strate­gic asset

Tesco, the UK’s largest retail­er, used a big data approach to analyse how sales of dif­fer­ent food items respond­ed to weath­er con­di­tions. It built a com­put­er mod­el which took account of store loca­tion, prod­uct char­ac­ter­is­tics, recent weath­er his­to­ry and weath­er fore­casts to bet­ter pre­dict demand, sav­ing around £6 mil­lion by avoid­ing under or over-order­ing. Apply­ing sim­i­lar ana­lyt­ics to dis­tri­b­u­tion depot oper­a­tions, shelf-life dis­count­ing and spe­cial offers has cre­at­ed about £100 mil­lion annu­al improve­ment in sup­ply chain effi­cien­cy, says Dun­can Apthorp, Tesco’s pro­gramme man­ag­er, sup­ply chain devel­op­ment.

While Marks & Spencer is also reap­ing the ben­e­fits of big data ana­lyt­ics to fore­cast demand over the com­ing days and weeks, the retail­er is pre­dict­ing sup­ply pat­terns in the long term. Using a mass of glob­al data sources, such as fore­casts for labour costs in Chi­na, it is mod­el­ling the cost-effi­cien­cy of its sup­ply chain years into the future.

“The sys­tem is look­ing at where our sup­pli­ers are and where are they like­ly to be in the next few years, look­ing at demo­graph­ic trends in places like Chi­na. We have always done sce­nario mod­el­ling, but now we are run­ning mod­els more fre­quent­ly and using more exter­nal data,” says Emile Naus, head of logis­tics strat­e­gy at M&S.

Mean­while, busi­ness­es are using ana­lyt­ics to improve strate­gic pro­cure­ment. Over the past year, glob­al machin­ery man­u­fac­tur­er Cater­pil­lar has been bring­ing togeth­er data from across enter­prise appli­ca­tions to do “deep-dive ana­lyt­ics” on sup­pli­er per­for­mance, says chief pro­cure­ment offi­cer (CPO) Frank Cre­spo. Cit­i­group, one of the world’s biggest banks, has been increas­ing its invest­ment in pro­cure­ment ana­lyt­ics to “shift the infor­ma­tion advan­tage from sup­ply chain to us”, says Scott Whar­ton, the bank’s chief pro­cure­ment offi­cer.

While no pre­dic­tion can be per­fect­ly accu­rate, pro­cure­ment and sup­ply chain pro­fes­sion­als are exploit­ing the bur­geon­ing growth in data sources and per­for­mance of ana­lyt­ics tools to help pre­pare for what lies ahead.