The aver­age mar­keter worth their salt can scratch out their customer’s rough demo­graph­ics while sit­ting alone in their room, peck­ing data points into a spreadsheet.

Cre­at­ing a prof­itably con­vert­ing seg­ment is easy. But just because the seg­ment you lay out in Excel con­verts prof­itably, that doesn’t mean your seg­ment accu­rately defines your core cus­tomer base.

Is Your Seg­ment What’s Actu­ally Delivering?

How could that be? If your seg­mented cam­paign deliv­ers, then clearly your seg­ment cre­ated your mar­ket­ing suc­cess, right?

Well, cor­re­la­tion doesn’t always imply causation.

You see, the qual­i­ties you cre­ated for your seg­ment aren’t the only qual­i­ties cer­tain mem­bers of your cus­tomers share with each other. You may prof­itably sell an app to a seg­ment of 25–34 year old males who love Apple prod­ucts and have already bought 10 apps in the last year, but some of them will also share the fact that they’re sin­gle, upper-middle class, and liv­ing in major East Coast cities.

Guess what? It’s pos­si­ble your real cus­tomers aren’t actu­ally 25–34 year old avid iPhone users. Rather, your real cus­tomers are sin­gle, upper-middle class East Coast urbanites.


In this case, your mar­ket­ing cam­paign is prof­itable because enough mem­bers of this sec­ond demo­graphic fall within the seg­ment you cre­ated and not because your seg­ment itself was on-point.

Never assume the seg­ment you cre­ate defines your actual cus­tomer base. It’s always pos­si­ble your actual cus­tomers just hap­pen to appear in high-enough con­cen­tra­tions within the seg­ment you cre­ated to pro­duce profitability.

Too Com­pli­cated for Pen and Paper

Believ­ing your seg­ment causes prof­itabil­ity, and doesn’t just cor­re­late with prof­itabil­ity, is a com­mon and easy mis­take. Your ideal cus­tomers are con­structed by data pat­terns that are a lot more com­pli­cated and diverse than any­one could ever think up on their own because us humans tend to fix­ate on just one or two char­ac­ter­is­tics we can digest.

By con­trast, our top cus­tomers use Adobe Ana­lyt­ics to run more than 100 dimen­sions through the mill when iden­ti­fy­ing what their prof­itable seg­ment actu­ally looks like. These dimen­sions are all eval­u­ated as a whole to deter­mine their rela­tion­ship to each other and their influ­ence on your desired mar­ket­ing out­come. And when you look at over 100 vari­ables at once, you’re going to see cor­re­la­tions defined by seem­ingly unre­lated attrib­utes and dimen­sions, all con­nected in ways you would never have guessed on your own.

Even if you plug 100+ data points into your spread­sheet, you still wouldn’t be able to find these con­nec­tions, and you cer­tainly wouldn’t be able to per­form next-generation seg­men­ta­tion solu­tions, includ­ing pre­dic­tive seg­men­ta­tion, with any effi­ciency. Yes, it’s pos­si­ble to per­form sta­tis­ti­cal analy­sis in Excel, but when you use spread­sheets you’re still inputting and cal­cu­lat­ing your analy­sis man­u­ally, which nat­u­rally lim­its the size, sophis­ti­ca­tion and cre­ative capa­bil­i­ties of your analysis.

Although a lot of mar­ket­ing prob­lems can be solved with a lit­tle per­sonal inge­nu­ity, accu­rate seg­men­ta­tion just isn’t one of them. If you encounter prob­lems with your seg­ment, then you must rig­or­ously and con­tin­u­ously test your assump­tions about who you’re actu­ally sell­ing to, and you have to use a mod­ern ana­lyt­ics solu­tion to do so.

What’s Wrong with “Good Enough”?

If your seg­ment deliv­ers a profit, then why worry about test­ing its valid­ity? Do you really need to know why your cam­paign deliv­ers a profit?

Let’s ignore the fact that accu­rately know­ing your audi­ence will help you cre­ate a more prof­itable mar­ket­ing cam­paign. In many ways, the data you pull from test­ing your mar­ket­ing assump­tions have pos­i­tive ram­i­fi­ca­tions that rip­ple out far beyond sim­ply know­ing who to shout your product’s praises to.

For exam­ple, if you don’t mar­ket to your real cus­tomers, then there’s a good chance your prod­uct devel­op­ment team acts with sim­i­lar mis­in­for­ma­tion. When you fig­ure out who your cus­tomers actu­ally are, your prod­uct devel­op­ment team can then opti­mize their prod­uct to best suit those peo­ple who want it the most.

Seg­ment­ing is the mar­ket­ing equiv­a­lent of build­ing a prod­uct in the lab with­out test­ing that prod­uct on real users—you may come up with some­thing that sells, but ulti­mately you’ve only taken one step toward craft­ing suc­cess. And when it comes down to it, a will­ing­ness to keep tak­ing addi­tional steps for­ward often sep­a­rates wildly suc­cess­ful cam­paigns from cam­paigns that are merely “good enough.”