Making Data Work Harder for You
Skip to content
Data AnalyticsStrategy Marketing Jul 6, 2015

Mak­ing Data Work Hard­er for You

A strate­gic approach to data ana­lyt­ics starts with ask­ing the right questions.

Data science 101 helps a leader understand data analytics.

Yevgenia Nayberg

Based on the research of

Eric T. Anderson

Florian Zettelmeyer

Listening: Insight in Person Data Analytics

download
0:00 Skip back button Play Skip forward button 15:29

This month, Insight In Per­son looks at how busi­ness lead­ers can incor­po­rate data ana­lyt­ics into their com­pa­nies — from what the manager’s role should be to how big data can be used in prod­uct development.

Flo­ri­an Zettelmey­er, a pro­fes­sor of mar­ket­ing at the Kel­logg School and fac­ul­ty direc­tor of the pro­gram on data ana­lyt­ics at Kel­logg, explains why every leader needs a work­ing knowl­edge of data sci­ence,” and why it is impor­tant to encour­age data lit­er­a­cy across organizations.

Eric Ander­son, a pro­fes­sor and chair of the mar­ket­ing depart­ment at the Kel­logg School, dis­cuss­es how com­pa­nies can use ana­lyt­ics to see who is buy­ing their prod­ucts — and whether they’re the kind of cus­tomer that might be a har­bin­ger of failure.”

[Music prelude]

[Fred SCHMALZ: Hel­lo, and wel­come to Insight in Per­son, Kel­logg Insight’s month­ly pod­cast, pro­duced by the Kel­logg School of Man­age­ment at North­west­ern Uni­ver­si­ty. I’m your host, Fred Schmalz.

This month we take a look at how busi­ness lead­ers can incor­po­rate data ana­lyt­ics into their com­pa­nies — from why man­agers need to be involved in the data ana­lyt­ics process to how big data can be used in prod­uct devel­op­ment. In the first half of the pod­cast, we hear about how a work­ing knowl­edge of data sci­ence can help busi­ness lead­ers get ahead and man­age with confidence.

In the sec­ond half of the pod­cast, we look at how data ana­lyt­ics can tell us who’s buy­ing our prod­ucts — and whether they’re the kind of cus­tomer that might be a har­bin­ger of fail­ure for those products.

So stay with us.

[Music inter­lude]


ACT 1: Flo­ri­an Zettelmey­er on why a work­ing knowl­edge of data ana­lyt­ics is nec­es­sary for busi­ness leaders


Flo­ri­an ZETTELMEY­ER: One of the real­ly fas­ci­nat­ing devel­op­ments in our world is that it has become eas­i­er to mea­sure things.

[Fred SCHMALZ: That’s Flo­ri­an Zettelmey­er, a pro­fes­sor of mar­ket­ing at the Kel­logg School and fac­ul­ty direc­tor of the pro­gram on data ana­lyt­ics at Kellogg.

ZETTELMEY­ER: And because it’s eas­i­er to mea­sure things and it’s eas­i­er to visu­al­ize things, it is also eas­i­er to present such infor­ma­tion direct­ly to man­agers and direct­ly to deci­sion makers.

SCHMALZ: Zettelmey­er says that in today’s data-dri­ven world, man­agers need to think of data ana­lyt­ics as part of their reper­toire, rather than some­thing that falls out­side of their domain.

After all, lead­ers are the ones who will have to make strate­gic deci­sions based on the data, so they should be the gate­keep­ers when it comes to decid­ing how data can add val­ue to their business.

ZETTELMEY­ER: They now need to have a lev­el of aware­ness of what you can and what you can­not learn on the basis of the data that you observe — which, frankly, was not a skill that has tra­di­tion­al­ly been trained for a man­age­r­i­al class but was a skill that was the domain of the data sci­en­tist or the ana­lyt­ics people.

SCHMALZ: For the data ana­lyt­ics process to be effec­tive, Zettelmey­er says, man­agers need to be involved every step of the way. In order to do this, they need to have what he calls a work­ing knowl­edge” of data science.

So what exact­ly is a work­ing knowl­edge of data sci­ence? Zettelmey­er says the nec­es­sary skills are not learned in an advanced econo­met­rics class. In fact, they’re not real­ly tech­ni­cal skills at all — they’re think­ing skills.

ZETTELMEY­ER: You don’t real­ly need to be great at math or at stats or com­put­er sci­ence to obtain a work­ing knowl­edge. I’m not say­ing that it is super easy to get it, but it is real­ly a way of dis­ci­plined think­ing that needs some prac­tice but that fun­da­men­tal­ly does not need a lot of tech­ni­cal back­ground. There­fore, it is fun­da­men­tal­ly acquirable by peo­ple who have suc­ceed­ed in the busi­ness world, pre­cise­ly for the fact that they are good ana­lyt­i­cal thinkers.

SCHMALZ: These think­ing skills — and this dis­ci­pline — can save man­agers from blind­ly trust­ing data sci­en­tists by assum­ing that just because they have the num­bers, the analy­sis was per­formed in a rea­son­able way.

To do ana­lyt­ics right, you need more than data — you also need to draw upon an in-depth knowl­edge of your busi­ness, which is exact­ly the kind of domain exper­tise good lead­ers bring to the table.

ZETTELMEY­ER: This is some­thing that often data sci­en­tists are not so good at find­ing. But it’s some­thing that man­agers are excel­lent at find­ing. It real­ly speaks to this idea that you get both very good data sci­en­tists as well as very good busi­ness peo­ple involved in ana­lyt­ics because that’s how you make progress and avoid mistakes.

SCHMALZ: Cre­at­ing a cul­ture where ana­lyt­ics add val­ue to a busi­ness goes well beyond just putting togeth­er the right ana­lyt­ics team. Man­ag­ing ana­lyt­ics across the orga­ni­za­tion is key.

ZETTELMEY­ER: Many of the core bar­ri­ers to mak­ing ana­lyt­ics work have actu­al­ly noth­ing to do with ana­lyt­i­cal prob­lems per se. They are not about pick­ing the right algo­rithm; they’re often not even about pick­ing the right data.

What they are about is that, when you do ana­lyt­ics, you run into orga­ni­za­tion­al and incen­tive bar­ri­ers that pre­vent you from actu­al­ly doing the ana­lyt­ics or that pre­vent you from act­ing on the ana­lyt­ics. For exam­ple, imag­ine that you want to get a 360 degree view on the cus­tomer, which is a big mantra in ana­lyt­ics. Well, the only way you are going to do that is if you can get all the busi­ness units to actu­al­ly coop­er­ate to pro­vide data so that you do get a 360 degree view of the customer.

SCHMALZ: In oth­er words, if a company’s busi­ness units aren’t in the habit of work­ing col­lab­o­ra­tive­ly, using data sci­ence can quick­ly become a source of orga­ni­za­tion­al ten­sion. Per­form­ing ana­lyt­ics, like many aspects of man­ag­ing an orga­ni­za­tion, is fun­da­men­tal­ly political.

ZETTELMEY­ER: There are win­ners and there are losers from imple­ment­ing ana­lyt­ics. There are peo­ple who have dif­fer­ent agen­das. Before long, if you are a manger, you will find your­self in a sit­u­a­tion where you have team A report­ing back to you with result A backed up by data sci­ence team A.

You have team B report­ing back to you with result B backed up by data sci­ence team B. Guess what? Result A and result B are not going to be the same. The prob­lem is that you can’t real­ly go out there and say, Let me get my data sci­en­tist in order to some­how fig­ure out which of these two teams is right.” Because there are already ten data sci­en­tists in the room and they all dis­agree with each other.

SCHMALZ: Data ana­lyt­ics can­not be treat­ed as a sep­a­rate part of one’s busi­ness — it has to be done with orga­ni­za­tion­al chal­lenges in mind, and it has to be incor­po­rat­ed into the busi­ness plan itself. Too often, Zettelmey­er says, man­agers decide to col­lect data with­out know­ing exact­ly how they will use it, and this leads to prob­lems down the road.

ZETTELMEY­ER: You have to think about the gen­er­a­tion of data as a strate­gic imper­a­tive. You can’t just go out in a com­pa­ny and some­how hope that the data that is avail­able to you and the data that gets inci­den­tal­ly cre­at­ed in the course if busi­ness is the kind of data that’s going to lead to break­throughs or the kind of data from which you can learn some­thing about your business.

Ana­lyt­ics has to fun­da­men­tal­ly start with busi­ness prob­lems. I can’t tell you how often I’ve heard the sen­tence by an exec­u­tive who comes and wants to talk to me and says, You know, we have this enor­mous amount of data. We are over­whelmed by it. We don’t know what to do with it. But we are sure there’s some­thing incred­i­bly valu­able in there.”

SCHMALZ: Whether such data can add val­ue depends on what kinds of ques­tions are asked, how the data are gen­er­at­ed, and how the lead­er­ship makes deci­sions based on its analy­sis. When ana­lyt­ics goes bad, it usu­al­ly does so because the insights are non­sci­en­tif­ic, how­ev­er much they appear to offer.

For Zettelmey­er, the way to do ana­lyt­ics well is to start with the prob­lem and incor­po­rate the ana­lyt­ics into the busi­ness plan itself. It also means empow­er­ing every­one in the orga­ni­za­tion to use their ana­lyt­ic skills to inter­pret and ques­tion the data

ZETTELMEY­ER: If we want big data and ana­lyt­ics to suc­ceed, every­body needs to feel that they have this right to ques­tion. There has to be a cul­ture where you can’t get away with think­ing as opposed to knowing.

SCHMALZ: For Man­agers, hav­ing a work­ing knowl­edge of data sci­ence has three main ben­e­fits. First, it helps you judge what good data look like. Sec­ond, it helps you iden­ti­fy pre­cise­ly where ana­lyt­ics can help the busi­ness. And third, it helps you lead with con­fi­dence. Increas­ing­ly, hav­ing this knowl­edge will become the norm for all good busi­ness leaders.

ZETTELMEY­ER: Can you imag­ine a CFO going to the CEO and say­ing, You know what? I don’t real­ly know how to read a bal­anced sheet, but I have some­body on my team who is real­ly good at it.” Okay, we would laugh that per­son out of the room today. Now, I hate to tell you but I know a whole bunch of peo­ple in mar­ket­ing and oth­er areas at the C-suite who with­out blink­ing an eye would go to the CEO and say, You know what? This ana­lyt­ics stuff is com­pli­cat­ed. I don’t have a full grasp on it, but I have assem­bled a crack­er­jack ana­lyt­ics team that is going to push us to the next level.”

I think this is an answer that is no longer acceptable.

[Music inter­lude]


ACT 2: Eric Ander­son on Har­bin­gers of Failure


[Fred SCHMALZ: No mat­ter how often you go to the super­mar­ket, you are like­ly to see new prod­ucts — from break­fast cere­als to tasty yogurt fla­vors to house­hold cleansers — com­pet­ing for shelf space with estab­lished brands.

Yet new prod­ucts tend to fail at very high rates.

Which begs the ques­tion, how can so many prod­ucts make it through the devel­op­ment process and then flop once they land in stores?

In man­ag­ing the intro­duc­tion of new prod­ucts, it turns out that pre­dict­ing their suc­cess or fail­ure doesn’t just depend upon whether lead users buy those prod­ucts, but which lead users buy them.

Pro­fes­sor Eric Ander­son, chair of the mar­ket­ing depart­ment at the Kel­logg School, want­ed to know more about why prod­ucts failed and how to iden­ti­fy those prod­ucts ear­ly in the process.

In a recent study, Ander­son and his coau­thors found that the pres­ence of cer­tain buy­ers of a prod­uct may indi­cate that the prod­uct is like­ly to fail. And they found that these har­bin­gers of fail­ure” tend to buy a lot of these failed products.

Here’s Eric Anderson.

Eric ANDER­SON: What spawned out of this was a fair­ly cre­ative idea about how to think about suc­cess and fail­ure of new prod­ucts. We start­ed think­ing about con­sumers who might sys­tem­at­i­cal­ly buy either win­ners or losers. Prod­ucts that are win­ners are ones that sur­vive on the shelf for a long time, things that you see all the time, like Hellman’s may­on­naise. Losers are prod­ucts that you don’t know that well because they dis­ap­pear very quick­ly — prod­ucts like water­mel­on Ore­os that are on the shelf for a short peri­od of time and then quick­ly disappear.

SCHMALZ: When research­ing win­ning and los­ing prod­ucts, Ander­son found that there are indeed cus­tomers who have a sys­tem­at­ic pref­er­ence for prod­ucts that tend to fail.

ANDER­SON: It turns out that the more you buy of these failed prod­ucts, the more like­ly you are to buy anoth­er failed prod­uct. That real­ly rein­forced the notion that these cus­tomers were behav­ing systematically.

We can talk about cus­tomers buy­ing a soda prod­uct and sys­tem­at­i­cal­ly buy­ing failed soda. Then we can show that this helps us pre­dict in oth­er cat­e­gories, like house­hold cleaners.

SCHMALZ: So who are these cus­tomers? Are they peo­ple far out on the taste spec­trum, whose gro­cery carts are loaded down with rasp­ber­ry beer and kumquat cus­tard? Are they young peo­ple with adven­tur­ous, if curi­ous, dietary habits? Mil­len­ni­als with lit­tle of the brand loy­al­ty of their pre­de­ces­sors? Are they cus­tomers grav­i­tat­ing toward the cheap­est prod­ucts on the shelf?

The answer to this ques­tion sur­prised Ander­son and his coauthors.

ANDER­SON: What we’ve dis­cov­ered thus far is in the pack­aged-goods world, these cus­tomers tend to be fair­ly edu­cat­ed, fair­ly wealthy, and have large fam­i­ly sizes.

SCHMALZ: What’s curi­ous is that one would assume afflu­ent peo­ple with larg­er fam­i­lies would tend to spend more at the super­mar­ket, so intu­itive­ly, if cus­tomers buy more of some­thing, that’s a good thing, right? Not necessarily.

ANDER­SON: The twist is that the more you sell to cus­tomers who are these har­bin­gers of fail­ure, the more like­ly you are to fail. That’s what is the sur­prise. The more you sell, the more like­ly you are to fail when those sales are going to cus­tomers who buy Diet Crys­tal Pepsi.

That’s what peo­ple find inter­est­ing and shock­ing, that you can actu­al­ly pre­dict new prod­uct fail­ure. That insight is what opens up all sorts of new pos­si­bil­i­ties for future research.

SCHMALZ: From a data-ana­lyt­ics stand­point, the con­sumer goods indus­try is an excel­lent mar­ket to test for har­bin­gers, because, through super­mar­kets’ cus­tomer-loy­al­ty pro­grams, gran­u­lar infor­ma­tion is col­lect­ed on who buys what, which helps deter­mine suc­cess or fail­ure on a micro lev­el. But as we heard ear­li­er, hav­ing more infor­ma­tion is not auto­mat­i­cal­ly going to lead to insights.

ANDER­SON: A mis­con­cep­tion is that what you need to cre­ate val­ue is lots of data or per­haps tremen­dous horse­pow­er with respect to ana­lyt­ics and com­ing up with the next best algo­rithm for pro­cess­ing the data. Here nei­ther of those actu­al­ly leads to value.

What dri­ves val­ue is what I’ll think of as cre­ative insight. It’s the spark of an idea that thinks about the world in a com­plete­ly dif­fer­ent way and doesn’t take the sta­tus quo as giv­en. It paus­es for a sec­ond and says, hey, can we think about things dif­fer­ent­ly. Here, the ques­tion we asked, where we spun every­thing upside down, is, look, the new prod­uct-devel­op­ment process always takes pos­i­tive feed­back as a mea­sure of success.

Each one of these mea­sures tells me, yes, keep going, keep going. What we point out is, here’s a mea­sure of suc­cess. You sell more to these har­bin­gers of fail­ure, and it’s com­plete­ly neg­a­tive­ly cor­re­lat­ed with actu­al mar­ket per­for­mance. That’s what twists things and throws it upside down.

Much of the val­ue is locked into what kinds of ques­tions you ask. It’s very dif­fi­cult to get new insights with­out new questions.

SCHMALZ: Infor­ma­tion about the pres­ence and pur­chas­ing habits of har­bin­gers of fail­ure can help com­pa­nies improve the new prod­uct devel­op­ment process to give those prod­ucts that do appear a greater chance of suc­cess. The key is to devel­op met­rics that allow firms to clas­si­fy prod­uct-test­ing respon­dents into har­bin­gers and non-har­bin­gers as ear­ly in the process as possible.

ANDER­SON: By under­stand­ing that dif­fer­ence ear­ly on, you can start to make the appro­pri­ate invest­ments. You can align your invest­ment in the prod­uct and how you’re posi­tion­ing it with your retail part­ners to say, this is serv­ing a dif­fer­ent pur­pose. This prod­uct might be a niche prod­uct and is serv­ing a dif­fer­ent goal rather than being a mass-mar­ket product.

SCHMALZ: The rea­son these prod­ucts fail is not because they are inher­ent­ly bad ideas. It’s per­haps that they were incor­rect­ly posi­tioned as mass-mar­ket prod­ucts when they may have been niche prod­ucts. Data ana­lyt­ics can help deter­mine that prod­uct posi­tion­ing, both in the super­mar­ket and wher­ev­er peo­ple make purchases.

ANDER­SON: If we have that kind of data, I think we can start to explore, does this extend into home elec­tron­ics? Does it extend into appli­ances? We think it does. I think my con­jec­ture is that it will extend into oth­er indus­tries, but hope­ful­ly it will lead to new insights and new dis­cov­er­ies along the way as we start to explore oth­er categories.

[Music inter­lude]


SCHMALZ: This pro­gram was pro­duced by Drew Calvert, Fred Schmalz, Kate Pro­to, and Michael Spikes. Spe­cial thanks to Kel­logg School of Man­age­ment fac­ul­ty Flo­ri­an Zettelmey­er and Eric Anderson.

You can stream or down­load our month­ly pod­cast from iTunes, or from our web­site, where you can read more about Flo­ri­an Zettelmeyer’s and Eric Anderson’s research and get tips on data ana­lyt­ics for man­agers. Vis­it us at insight​.kel​logg​.north​west​ern​.edu. We will be back next month with anoth­er Insight In Per­son podcast.

Featured Faculty

Eric T. Anderson

Hartmarx Professor of Marketing, Professor of Marketing, Director of the Center for Global Marketing Practice

Florian Zettelmeyer

Nancy L. Ertle Professor of Marketing

Suggested For You

Most Popular

Organizations

How Are Black – White Bira­cial Peo­ple Per­ceived in Terms of Race?

Under­stand­ing the answer — and why black and white Amer­i­cans’ respons­es may dif­fer — is increas­ing­ly impor­tant in a mul­tira­cial society.

Leadership

Why Warmth Is the Under­ap­pre­ci­at­ed Skill Lead­ers Need

The case for demon­strat­ing more than just competence.

Most Popular Podcasts

Careers

Pod­cast: Our Most Pop­u­lar Advice on Improv­ing Rela­tion­ships with Colleagues

Cowork­ers can make us crazy. Here’s how to han­dle tough situations.

Social Impact

Pod­cast: How You and Your Com­pa­ny Can Lend Exper­tise to a Non­prof­it in Need

Plus: Four ques­tions to con­sid­er before becom­ing a social-impact entrepreneur.

Careers

Pod­cast: Attract Rock­star Employ­ees — or Devel­op Your Own

Find­ing and nur­tur­ing high per­form­ers isn’t easy, but it pays off.

Marketing

Pod­cast: How Music Can Change Our Mood

A Broad­way song­writer and a mar­ket­ing pro­fes­sor dis­cuss the con­nec­tion between our favorite tunes and how they make us feel.