What is the buzz around data sci­ence? The data sci­en­tist role is in demand across cor­po­ra­tions, con­sul­tan­cies, and mar­ket­ing ser­vice providers. Accord­ing to research con­duct­ed by Accen­ture, 80 per­cent of data sci­en­tist posi­tions cre­at­ed aren’t filled. There’s a hint of des­per­a­tion — “If I hire a data sci­en­tist, it will solve every­thing.” But you need more. You need a lab.

With­out a data sci­en­tist, a mar­keter can’t know whether suc­cess in a sin­gle cam­paign leads to increased com­pa­ny return on invest­ment. Data sci­en­tists dis­cov­er pat­terns in the cus­tomer jour­ney that lead to impact­ful and effi­cient mar­ket­ing.

But how?

Through data explo­ration using knowl­edge dis­cov­ery tech­niques such as data min­ing, sta­tis­tics, and pre­dic­tive ana­lyt­ics, new con­sumer behav­ior pat­terns and data rela­tion­ships are test­ed in mar­ket­ing “exper­i­ments,” also known as cam­paigns.

Discoveries In The Marketing Lab — Why This Approach?

Data sci­en­tists need the free­dom afford­ed in the “lab exper­i­ment” sce­nario to devel­op mar­ket­ing hypothe­ses, test, refine and val­i­date, using results from a pre­vi­ous cam­paign. Only after val­i­da­tion can the data sci­en­tist con­fi­dent­ly say that the method­ol­o­gy is ready to “go to mar­ket.”

Why does this mat­ter?

Because in the fast-paced dig­i­tal-on-demand world, it’s easy for a com­pa­ny to skip the dis­ci­plined lab set­ting and allow untest­ed hypothe­ses to dri­ve data deci­sions.

The 2016 Customer Journey? It’s Complicated

What you’d real­ly like to know are the details: Did your cam­paign meet your goal for leads gen­er­at­ed? If so, did you mea­sure how many of those leads remained loy­al cus­tomers two years lat­er?

Sim­i­lar­ly, one would expect sci­en­tif­ic rig­or for an exper­i­men­tal can­cer drug: Did it yield a pos­i­tive out­come? Or, did it show promise and require more research? How did patients fare two years after treat­ment?

There is a struc­ture and dis­ci­pline required to build trust in data. That’s where the lab­o­ra­to­ry is essen­tial.

In the mar­ket­ing lab we can make sense of data in areas like mar­ket­ing automa­tion, machine learn­ing, and pro­gram­mat­ic adver­tis­ing, to name a few. In a lab envi­ron­ment, data sci­en­tists have the free­dom to pull nuggets from all of that unstruc­tured data that’s so chal­leng­ing — and fol­low the trail of the cus­tomer jour­ney from end to end.

Accord­ing to For­rester, the future inno­va­tion lab will be an ecosys­tem built around the needs of a spe­cif­ic con­sumer seg­ment. We must encour­age cross-fer­til­iza­tion between dif­fer­ent types of spe­cial­ists assem­bled for their exper­tise.

The rock stars of the new lab are the rare birds who “get” con­sumers, data, dig­i­tal, con­tent, dis­tri­b­u­tion, and new prod­uct mod­els and who can cajole diverse geeks to speak the same lan­guage and work togeth­er con­struc­tive­ly.

So, can the data sci­en­tist be the rock star? The answer is yes, if she has knowl­edge of your cus­tomer jour­ney, sta­tis­ti­cal knowl­edge, prac­ti­cal indus­try exper­tise and the abil­i­ty to dis­cern cau­sa­tion and cor­re­la­tion. This is a lot to ask of one per­son, so it will like­ly take a team of experts to make a great lab.

Speed Is Essential!

Ven­ture­Beat points out that mar­ket­ing agen­cies are well-posi­tioned to fit enter­pris­es’ loom­ing out­source needs for data-dri­ven mar­ket­ing, if they can build the omnichan­nel con­tent mar­ket­ing machines pow­ered by robust data ana­lyt­ics that busi­ness will require.

Data sci­en­tists oper­at­ing in a lab envi­ron­ment will find insights.

For exam­ple, in the finan­cial ser­vice cus­tomer jour­ney, we know that indi­vid­u­als who respond to mar­ket­ing efforts for cred­it do not nec­es­sar­i­ly become cus­tomers, or if they do, they may not be prof­itable. Why?

Data sci­ence shows that find­ing respon­sive indi­vid­u­als who also can pass rig­or­ous lend­ing under­writ­ing rules can be chal­leng­ing. Con­sumers with the high­est like­li­hood to respond typ­i­cal­ly need cred­it, but may have prob­lems mak­ing their month­ly pay­ments.

So, we build a bal­anced bridge. That is, we con­nect response, under­writ­ing and con­ver­sion in a fine-tuned algo­rithm that fits finan­cial prod­ucts to cus­tomer needs while meet­ing cred­it risk rules.

This is just the begin­ning of data sci­ence and the lab approach. Once insights are derived and the pre­dic­tive ana­lyt­ics are in place, the machine will do the work — scor­ing your cus­tomers and prospects, rank­ing them, and fit­ting a prod­uct solu­tion to the indi­vid­ual using the tools and infra­struc­ture you’ve built.

What’s Next?

Get start­ed by build­ing an inven­to­ry of skills and tools need­ed to launch your lab. Recruit with­in your orga­ni­za­tion. Who has knowl­edge of the indus­try, the cus­tomer jour­ney, and busi­ness chal­lenges?

Who are your data sci­ence and BI experts?

Once you’ve searched inter­nal­ly, fill gaps by going out­side your orga­ni­za­tion. Inter­view con­sul­tants, agen­cies, and sub­ject-mat­ter experts that will design, deliv­er and hand off. Set time aside for your new lab team to devote qual­i­ty time — ide­al­ly as their main focus.