What is pre­dic­tive ana­lyt­ics? What is pre­dic­tive mar­ket­ing? Do I need a sta­tis­ti­cian on my mar­ket­ing staff? These ques­tions and oth­ers related to analy­sis and dig­i­tal mar­ket­ing are increas­ing in fre­quency. Since it is rel­a­tively easy to col­lect data, espe­cially in the online space, dig­i­tal mar­keters have a wealth of data points to describe their online cus­tomers and site per­for­mance. Wouldn’t it be great if we could reduce the data to its most rel­e­vant points and use it to not only describe, but pre­scribe and per­haps pre­dict? Enter pre­dic­tive analytics.

What is Pre­dic­tive Analytics?

Pre­dic­tive ana­lyt­ics is the prac­tice of using data min­ing and sta­tis­ti­cal mod­el­ing to describe what is hap­pen­ing in your orga­ni­za­tion and esti­mate poten­tial out­comes. As a result of pre­dic­tive ana­lyt­ics, orga­ni­za­tions are able to fore­cast rev­enue, define attri­bu­tion mod­els, and iden­tify cus­tomer seg­ments and score them on their like­li­hood to com­plete a desired action.

Pre­dic­tive ana­lyt­ics is not new. Orga­ni­za­tions have been using sta­tis­ti­cal mod­el­ing and data min­ing for years. It is used in fraud detec­tion, rev­enue mod­el­ing, and even in HR hir­ing mod­els. It is also used in direct mail cam­paigns – iden­ti­fy­ing house­holds most likely to respond to the mar­ket­ing mes­sage. The dis­ci­pline is usu­ally found in the Busi­ness Intel­li­gence or IT groups within an orga­ni­za­tion, but is mak­ing itself known in the dig­i­tal mar­ket­ing space.

What is Pre­dic­tive Marketing?

Pre­dic­tive mar­ket­ing is the prac­tice of apply­ing data min­ing and sta­tis­ti­cal mod­el­ing to opti­mize mar­ket­ing efforts. For exam­ple, seg­ment­ing and scor­ing cus­tomers and online vis­i­tors based on a propen­sity to com­plete a desired action or defin­ing attri­bu­tion mod­els based on cam­paign suc­cess to assist with bud­get­ing and planning.

The chal­lenge with dig­i­tal mar­ket­ing is the need to respond quickly. Most orga­ni­za­tions have a BI or IT team churn­ing through data and cre­at­ing mod­els to describe behav­iors and out­comes, but the results may come after the oppor­tu­nity. The web moves fast and dig­i­tal mar­keters need to be able to respond just as fast. Hav­ing pre­dic­tive capa­bil­i­ties within the dig­i­tal mar­ket­ing group enables mar­keters to not only ana­lyze data quicker, but smarter. The result: improv­ing the effi­ciency to act.

For exam­ple, it is the end of the quar­ter and online prod­uct sales are low. A quick sta­tis­ti­cal analy­sis takes in pre­vi­ous sales activ­ity and behav­ioral data and iden­ti­fies vis­i­tor groups and behav­iors that are most likely to pur­chase given prod­ucts. Then a tar­get­ing cam­paign with spe­cific prod­uct and mes­sag­ing to these groups is launched to pro­vide more rel­e­vant pro­mo­tional data rather than a generic campaign.

Do I need a Statistician?

Remem­ber that high school or col­lege math class? The one where you spent hours work­ing on a home­work assign­ment, only to come in the fol­low­ing day and have the teacher say “Now we are going to learn how to do it by com­puter”.  The abil­ity to cal­cu­late the stan­dard devi­a­tion of a dis­tri­b­u­tion by hand does have its ben­e­fits. But when it is crunch time and deci­sions need to be made, I am going to turn to the com­puter. Make sure you have the right resource for the job.

Hav­ing some­one on the staff who under­stands the nature of the data and sta­tis­tics will help. It does not need to be a sta­tis­ti­cian, but some­one with the abil­ity to ana­lyze large data sets and per­form sta­tis­ti­cal and econo­met­ric analy­sis. A plus would be some­one with those skills and who under­stands how the busi­ness works, but those peo­ple are rare. I sug­gest hav­ing more than one per­son to bring var­i­ous skills together, or sup­ple­ment with consulting.

In Sum­mary

Many peo­ple think pre­dic­tive ana­lyt­ics is a black box, or crys­tal ball. Really it is just math and sci­ence. It is impor­tant to dif­fer­en­ti­ate that it is not a crys­tal ball telling you the future and what to do next. It is com­pass point­ing you in the most likely direc­tion. The account­abil­ity to act is still up to you. Unlike a crys­tal ball or black box, you can take apart a com­pass to under­stand how it works. Sim­i­larly, you can break a sta­tis­ti­cal model into smaller com­po­nents to under­stand the math and assump­tions dri­ving the out­come, which will help inform you of the deci­sions to make and the actions to take. The idea is to take the guess­work and sub­jec­tiv­ity out of the deci­sions that need to be made and become a lit­tle more edu­cated and effi­cient with action.