Have you ever been to a restau­rant or cafe where they offer you a “bot­tom­less” cup of cof­fee? This is the same con­cept as “free refills”, but I like the term bot­tom­less because it’s more clever.

In either case, n-dimensional seg­men­ta­tion is like a bot­tom­less cup of cof­fee. There is vir­tu­ally no limit to how many ways you can seg­ment your data and vis­i­tor activ­ity. As a mar­keter and web ana­lyt­ics pro­fes­sional, I love n-dimensional analy­sis and seg­men­ta­tion — and in this blog post, I wanted to share with you why it’s so valu­able and some ways you can profit from it.

What is “N-Dimensional”?
To make sure we’re all on the same page, let me start with a brief overview of the term “n-dimensional”. As I sug­gested above, n-dimensional refers to the con­cept of lim­it­less dimen­sions. When cou­pled with analy­sis and seg­men­ta­tion, the basic idea is that you can drill into your data in lim­it­less ways. While this may sound like an aca­d­e­mic exer­cise, n-dimensional analy­sis and seg­men­ta­tion is incred­i­bly pow­er­ful and can quickly deliver sig­nif­i­cant profit to your business.

Let’s start with an exam­ple
For instance, you might start with a key­word analy­sis for the term “web ana­lyt­ics”. Select your most rel­e­vant met­rics, in this case Searches and Rev­enue, to under­stand the over­all pop­u­lar­ity and con­tri­bu­tion of this keyword.

Now let’s say you want to drill deeper into this key­word to under­stand where it is most effec­tive and least effec­tive. Your first “dimen­sion” may be prod­ucts; in other words, you want to see which prod­ucts are pur­chased by vis­i­tors from this key­word. You see “Apple iPod” tops the list, so you fil­ter on this prod­uct to under­stand these cus­tomers better.

You may now won­der if these are new or repeat cus­tomers, so you pull up the “Cus­tomer Loy­alty” report and notice that most pur­chasers are New. So you add “New” cus­tomers to your fil­ter as well.

Now you’re curi­ous if these New Cus­tomers are younger or older shop­pers. So you bring up your Age Group report, and notice that most shop­pers are from 18–24. You add this to your fil­ter cri­te­ria as well.

At this point, you’d like to under­stand where all these peo­ple are vis­it­ing from. So you pull up the “Geog­ra­phy” report and see that most vis­i­tors orig­i­nate from Cal­i­for­nia and New York.

You’d like to try an email mar­ket­ing cam­paign to these folks, so you pull up another Prod­ucts report and ana­lyze which prod­ucts these vis­i­tors looked at but did not pur­chase. You see that many of these cus­tomers also viewed the iPod extended life bat­tery in the same visit, but didn’t buy it. So now you extract the cus­tomer IDs for these, export them to your email mar­ket­ing plat­form, and send out your remar­ket­ing campaign.

Now let’s take a step back and under­stand what you’ve just done. You’ve effec­tively seg­mented your entire web­site audi­ence by 6 dimen­sions (key­word, pur­chased prod­uct, cus­tomer loy­alty, age group, geog­ra­phy, and viewed prod­uct). This resulted in a highly tar­geted cus­tomer seg­ment that you can now remar­ket to based on their actual behav­ior and prod­uct affini­ties. How cool is that?

N-dimension = Bot­tom­less seg­men­ta­tion
So the crit­i­cal mes­sage is that n-dimensional analy­sis and seg­men­ta­tion allows you to slice and dice any data point by any other data point. As you do this, you’re drilling deeper and deeper into your data and cre­at­ing highly tar­geted seg­ments of your over­all traf­fic. I often call this “pro­gres­sive fil­ter­ing”, because your fil­ter cri­te­ria effec­tively nar­row the scope of your seg­ment with each incre­men­tal step. N-dimensional analy­sis is also some­times referred to as dynamic fil­ter­ing, drilling down, and for the adven­tur­ous, bot­tom­less segmentation!

How You Can Profit from N-Dimensional Analy­sis
As I said ear­lier, I love n-dimensional analy­sis because it’s so pow­er­ful, ele­gant, sim­ple, and just as there are lim­it­less dimen­sions you can seg­ment by, there are vir­tu­ally lim­it­less pos­si­bil­i­ties for prof­it­ing and busi­ness opti­miza­tion. To drive this point home, I wanted to offer another exam­ple beyond the remar­ket­ing strat­egy above. In this case, let’s focus on the web­site expe­ri­ence itself and more specif­i­cally, the con­ver­sion funnel.

Nearly every web­site has some form of con­ver­sion fun­nel. And with every con­ver­sion fun­nel comes attri­tion — vis­i­tors who start the process but do not fin­ish it. Take the shop­ping cart check­out process — one that most of us are prob­a­bly inti­mately famil­iar with by now. With each step of the shop­ping cart check­out process, you’ll lose vis­i­tors. It’s inevitable.

Now, pathing or click­stream reports like those offered in Omni­ture Site­Cat­a­lyst will show you were these peo­ple are leav­ing the process. But because peo­ple do not nec­es­sar­ily fol­low a lin­ear check­out process, any click­stream report may be some­what mis­lead­ing because peo­ple who diverge from the expected path may end up course cor­rect­ing and com­plet­ing the check­out anyway.

For this rea­son, Omni­ture offers the Fall­out report — which allows you to do “multi-node” fun­nel analy­sis. The Fall­out report basi­cally ignores the spe­cific paths vis­i­tors take, and instead, focuses on key pages that all vis­i­tors should touch at some point in a suc­cess­ful visit.

In the shop­ping cart exam­ple, these might include “Add to Cart”, “Billing”, “Ship­ping”, “Order Con­fir­ma­tion”, and “Thank You”. No mat­ter what path they take, your vis­i­tors would have to touch these 5 pages at some point in a suc­cess­ful visit (hypo­thet­i­cally speaking).

So the Fall­out report is great for under­stand­ing where peo­ple bail out at each major mile­stone in the process. But this still doesn’t tell you why they are bail­ing out — and that’s where n-dimensional analy­sis can be incred­i­bly valuable.

With n-dimensional analy­sis, you can pull up this same fall­out report to iden­tify these major attri­tion points. Once you’ve iden­ti­fied these points, you can add a fil­ter that Excludes all vis­its with an Order. This means your seg­ment will include all non-ordering visitors.

Now, pull up your Most Pop­u­lar Pages report and com­pare this to a suc­cess­ful visit? What are the major dif­fer­ences? Is there one page that peo­ple see more often as non-converters than they do as con­vert­ers? If you’re like most of your peers, you’ll quickly see a page or two that jumps out.

Now, keep­ing your seg­ment cri­te­ria intact, pull up another report that shows any error codes that may have resulted in these vis­its. Error codes may include form errors, page not found, redi­rects, authen­ti­ca­tion issues, etc and are eas­ily cap­tured within Site­Cat­a­lyst. Are there any error codes that are par­tic­u­larly preva­lent? If so, this rep­re­sents another great oppor­tu­nity for improve­ment. Why are these vis­i­tors get­ting the error page so fre­quently? Are they new cus­tomers try­ing to cre­ate a cus­tomer account? Are they from a par­tic­u­lar coun­try or state? Per­haps they are all from Canada and your zip code field isn’t accept­ing their postal code? Or alter­na­tively, are these vis­i­tors pre­dis­posed to visit the online help sec­tion? If so, what help arti­cles are they look­ing at in par­tic­u­lar? What help terms are they search­ing for? From which spe­cific page are they exiting?

As you might have guessed, each of these ques­tions rep­re­sents a unique dimen­sion you can fil­ter by to under­stand why these vis­i­tors are unsuc­cess­ful in the check­out process. And there are dozens more you can seg­ment by — all depend­ing on the rich­ness of your under­ly­ing data set. With n-dimensional analy­sis, the pos­si­bil­i­ties are truly limitless.

But wait, how can you do n-dimensional analy­sis?
For­tu­nately as an Omni­ture cus­tomer, n-dimensional analy­sis is just a mouse click away. Omni­ture Dis­cover offers you the abil­ity to slice and dice your data by any dimen­sion in real-time. If you aren’t already tak­ing advan­tage of Omni­ture Dis­cover , be sure to ask your Account Exec­u­tive or Account Man­ager how you can. And if you’re not an Omni­ture cus­tomer, but are inter­ested in n-dimensional analy­sis, please do not hes­i­tate to con­tact us and we’d be happy to demon­strate Omni­ture Dis­cover to you.