The Analytics Spectrum: What Era Are You In?

Higher Education

Over the past two years, eight Cal­i­for­nia State Uni­ver­si­ty cam­pus­es have been involved in a col­lab­o­ra­tive stu­dent ana­lyt­ics ini­tia­tive involv­ing Oracle’s Stu­dent Infor­ma­tion Ana­lyt­ics (SIA) sys­tem.  The intend­ed out­come of this col­lab­o­ra­tion is a fore­cast­ing dash­board for col­lege plan­ners and depart­ment lead­ers to make smarter deci­sion about inter­ven­tion and ser­vice plan­ning.

Dur­ing the plan­ning phase of this effort, a com­mon ques­tion that was asked among the col­leges par­tic­i­pat­ing was whether each insti­tu­tion was ready for ana­lyt­ics.  To answer this ques­tion, I devel­oped an Ana­lyt­ics Spec­trum to give col­lege and uni­ver­si­ty lead­ers an easy way to assess whether an insti­tu­tion is indeed ready for ana­lyt­ics.

What ana­lyt­ics is in high­er edu­ca­tion some­times depends on the ven­dor who is mak­ing a pitch.  More often than not, admin­is­tra­tors feel the Dash­board Daz­zle of charts and graphs that seem to be bring­ing out com­plex rela­tion­ships among stu­dent data that only an expen­sive pro­duct can pro­duce.  The harsh real­i­ty is this: third-par­ty ven­dors have been rak­ing in hun­dreds of mil­lions of dol­lars to give pub­lic insti­tu­tions the same visu­al­iza­tions that could be pro­duced in Excel. 

In Data Sci­ence for High­er Edu­ca­tion, I defined ana­lyt­ics in high­er edu­ca­tion as the col­lec­tion and analy­sis of data to pro­duce infor­ma­tion that reduces uncer­tain­ty about the future.  The tem­po­ral com­po­nent here is key in sep­a­rat­ing ana­lyt­ics from what Dash­board Daz­zle is real­ly doing: report­ing.  If insti­tu­tions want to tru­ly imple­ment ana­lyt­ics and begin dri­ving their fore­cast­ing and plan­ning efforts with data, a clear under­stand­ing of where they are — and where they want to be — on the Ana­lyt­ics Spec­trum is nec­es­sary.

The Analytics Spectrum

In 2009, the senior vice pres­i­dent and chief mar­ket­ing offi­cer of SAS Insti­tute Jim Davis illus­trat­ed that ana­lyt­ics has eight steps:

  1. Stan­dard reports.  What hap­pened? When did it hap­pen? Exam­ple: month­ly finan­cial reports.
  2. Ad hoc reports.  How many? How often? Where? Exam­ple: cus­tom reports.
  3. Query drill down/OLAP.  Where exact­ly is the prob­lem? How do I find the answers? Exam­ple: data dis­cov­ery about types of cell phone users and their call­ing behav­ior.
  4. Alerts.  When should I react? What actions are need­ed now? Exam­ple: CPU uti­liza­tion men­tioned ear­lier.
  5. Sta­tis­ti­cal analy­sis.  Why is this hap­pen­ing? What oppor­tu­ni­ties am I miss­ing?Exam­ple: why are more bank cus­tomers refi­nanc­ing their homes?
  6. Fore­cast­ing.  What if the­se trends con­tin­ue? How much is need­ed? When will it be need­ed? Exam­ple: retail­ers can pre­dict demand for prod­ucts from store to store.
  7. Pre­dic­tive mod­el­ing.  What will hap­pen next? How will it affect my busi­ness? Exam­ple: casi­nos pre­dict which VIP cus­tomers will be more inter­est­ed in par­tic­u­lar vaca­tion pack­ages.
  8. Opti­miza­tion.  How do we do things bet­ter? What is the best deci­sion for a com­plex prob­lem? Exam­ple: what is best way to opti­mize IT infra­struc­ture given mul­ti­ple, con­flict­ing busi­ness and resource con­straints?

The­se eight steps are account­ed for in The Ana­lyt­ics Spec­trum, which is bro­ken up into three eras: the Era of Reports, the Era of Intel­li­gence, and the Era of Opti­miza­tion.

The Era of Reports

Ques­tions Answered:

  • What hap­pened?
  • How many? How often? Where?
  • What exact­ly is the prob­lem?

Meth­ods: Stan­dard report­ing, ad hoc report­ing, query/drill-down

In this Era, the pri­ma­ry focus of all ana­lyt­ics is descrip­tive analy­sis.  It’s called the Era of Reports because the pri­ma­ry out­put is, well, reports.  This is where major­i­ty of insti­tu­tions are and will con­tin­ue to be, even after invest­ing in expen­sive third-par­ty ana­lyt­ics suites.  Dash­board Daz­zle leads peo­ple to believe that evolv­ing from this Era to the next is only a mat­ter of buy­ing the right pro­duct, when in fact you must have the right peo­ple and process­es in place before any pro­duct is worth the invest­ment and license costs.

To begin evolv­ing from the Era of Reports, com­plete the Ana­lyt­ics Readi­ness Check­list for High­er Edu­ca­tion.

The Era of Intelligence

Ques­tions Answered: 

  • What could hap­pen?
  • What will hap­pen next?
  • What if the­se trends con­tin­ue?
  • What actions are need­ed?

Meth­ods: Sim­u­la­tion, Fore­cast­ing, Pre­dic­tive Mod­el­ing, Alerts

The Era of Intel­li­gence is the sweet spot of where insti­tu­tions should be.  Here, the focus is on cre­at­ing action­able intel­li­gence, pro­ject­ing out­comes based on cer­tain con­straints, and fore­cast­ing pos­si­bil­i­ties given vari­ables that you can and can­not con­trol.  I argue that this is the ide­al place for insti­tu­tions to be, since get­ting any more rig­or­ous involves the type of opti­miza­tion tech­niques that might not be well suit­ed for pre­dict­ing the behav­ior of peo­ple.

THE ERA OF Optimization

Ques­tions Answered: 

  • What is the best that can hap­pen?
  • What is the best that can hap­pen given vari­abil­i­ty?
  • How do we do things bet­ter?
  • What is the best deci­sion for a com­plex prob­lem?

Meth­ods: Machine learn­ing, data sci­ence in high­er edu­ca­tion

The Era of Opti­miza­tion comes about almost seam­less­ly once the Era of Intel­li­gence has been ful­ly digest­ed across an insti­tu­tion.  In this Era, com­pu­ta­tion­al tools are used to auto­mate the pro­cess­ing of data into infor­ma­tion and the dis­sem­i­na­tion of infor­ma­tion so that it can be mold­ed into action­able intel­li­gence.  Every insti­tu­tion should strive to reach the Era of Opti­miza­tion because it will force plan­ners to mas­ter the Era of Intel­li­gence — which is where I think all insti­tu­tions should get to and main­tain.

What Era Are You In?

Are you still in the Era of Reports, and look­ing to evolve to the next Era?  Per­haps you’re look­ing for a way to help guide the evo­lu­tion of your insti­tu­tion toward a data-dri­ven, effi­cient Era of Intel­li­gence?  Con­sid­er the Ana­lyt­ics Readi­ness Check­list for High­er Edu­ca­tion (PDF), a free guide designed to provide lead­ers with deci­sion sup­port talk­ing points for dri­ving insti­tu­tion­al efforts in ana­lyt­ics plan­ning.  It’s short, it’s free, and it’s got all the steps nec­es­sary to start the right con­ver­sa­tions.

Leave a Reply