It is hard to believe that both of these paint­ings were cre­ated by the same artist.  At left, Picasso’s The Old Fish­er­man, dis­plays amaz­ing detail and depth of knowl­edge about the human fig­ure.  At right, Three Musi­cians, presents amaz­ing con­trast and is so sim­ple that it almost seems like any­one could have painted it with­out any train­ing at all, and yet it is the more well-known of the two.  I have often won­dered why that is.

Believe it or not, there is a strong tie in here with web ana­lyt­ics.  The rea­son that the sec­ond image is more pow­er­ful is not because of its sur­face appear­ance; it is the depth of knowl­edge of the artist behind it.

Recently one of our largest media clients asked the Adobe con­sult­ing team to help them iden­tify four to five key user mar­ket seg­ments using pre­dic­tive ana­lyt­ics and the data cap­tured by Site­Cat­a­lyst.  This was an extremely chal­leng­ing task con­sid­er­ing the client has over 80 mil­lion unique vis­i­tors to their site each month, each with unique behav­ior and con­tent affini­ties across many indi­vid­ual vis­its.  It was our task to char­ac­ter­ize each indi­vid­ual user, and then find a way to group these users into use­ful and action­able seg­ments that could be used by the client’s mar­ket­ing team to hone their efforts.

We started by iden­ti­fy­ing the Site­Cat­a­lyst met­rics that were most impor­tant to our client’s rev­enue stream.  Most of these met­rics cen­tered on video mea­sure­ment but we also incor­po­rated visit infor­ma­tion (length of visit, entry time, entry page, etc.), refer­rer infor­ma­tion (includ­ing search key­words), and geo­graphic loca­tion.  We then cre­ated an algo­rithm that could take all avail­able infor­ma­tion from each user and cre­ate a series of user pro­files to which each indi­vid­ual vis­i­tor belongs.

Once each vis­i­tor had been assigned a usage pro­file, we used advanced data min­ing algo­rithms to crawl through this mas­sive list and iden­tify com­mon usage behav­iors across the entire group to cre­ate four dis­tinct mar­ket seg­ments.  Each seg­ment rep­re­sented unique chal­lenges and oppor­tu­ni­ties to our client.  Some of these chal­lenges and oppor­tu­ni­ties they already knew about, but they were also sur­prised at how many oppor­tu­ni­ties they had not per­ceived.  For exam­ple, our client had a very large pro­por­tion of users who were com­ing to their site for a cer­tain video through social ref­er­ences and were not engag­ing with the site.  After look­ing into this video more closely, our client real­ized that they were not pre­sent­ing any addi­tional con­tent to these users.  Because of this, they were miss­ing out on a huge oppor­tu­nity to cross sell addi­tional con­tent.  They were also able to iden­tify which user types seem to con­gre­gate around cer­tain types of con­tent and thus, sig­nif­i­cantly edu­cate their mar­ket­ing efforts.

Many of the rec­om­men­da­tions that we made to our client might have seemed ret­ro­spec­tively obvi­ous to a sea­soned web ana­lyst.  How­ever, like the paint­ings above, it took an extremely deep dive into our client’s data cou­pled with an under­stand­ing of their mar­ket and busi­ness model to bring this “mas­ter­piece” to the sur­face.  Even more sat­is­fy­ing to our client was the knowl­edge that they were no longer rely­ing on heuris­tics, but were instead cre­at­ing data-driven strate­gies around solid analysis.

If you seem to be stuck on the sur­face of your data and can­not go deeper, con­tact the Adobe Con­sult­ing team for help uncov­er­ing the mas­ter­pieces below that pre­dic­tive ana­lyt­ics can bring to light. Speak with your Adobe account or sales rep­re­sen­ta­tive to learn more.