It’s clear to see why attri­bu­tion mod­els are impor­tant to a busi­ness. I mean, who doesn’t want to know why sales are hap­pen­ing or, how their mar­ket­ing efforts are dri­ving crit­i­cal busi­ness value. With dig­i­tal mar­ket­ing and tra­di­tional mar­ket­ing, attri­bu­tion is king.

It’s great to know that there is only one true attri­bu­tion model, since cus­tomers only ever inter­act with one tac­tic before con­vert­ing. Right? Isn’t it that simple?

Well unfor­tu­nately, no. When talk­ing about attri­bu­tion there’s myr­iad approaches that one could take. A pop­u­lar choice in dig­i­tal mar­ket­ing is ‘Last’. What was the most recent tac­tic a cus­tomer inter­acted with before con­vert­ing? That model is a great one to start with but it’s easy to see why it doesn’t lend itself well to the com­plete picture.

That sim­ple rea­son is, peo­ple aren’t that sim­ple. In one case that’s prob­a­bly a good thing but it does make the con­cept of attribut­ing a con­ver­sion to a tac­tic more involved.

I’m going to rec­om­mend tak­ing a look at Michael Halbrook’s recent post on the core attri­bu­tion solu­tion offered with the Adobe Insight plat­form. When read­ing it, keep in mind that’s just the entry point into the wild world of attri­bu­tion mod­el­ing and analy­sis. The Adobe Insight tech­nol­ogy eas­ily allows you to expand and adapt the mod­el­ing as your busi­ness changes.

Let me take a few moments more of your time to explain some ben­e­fits of using Adobe Insight. With most analy­sis tools you can access pre-canned aggre­gated reports. Insight is dif­fer­ent. It’s an advanced data analy­sis plat­form that builds rela­tion­ships between event based data to allow for unlim­ited analy­sis possibilities.

How does Insight build these relationships?

It all starts with the root of the data model. When a record is processed it’s grouped with other records gen­er­ated by the same cus­tomer (or vis­i­tor cookie) and then ordered by time. So instead of tak­ing the infor­ma­tion and increas­ing the daily page view counts or mar­ket­ing touch counts, that infor­ma­tion is actu­ally orga­nized and cor­re­lated together.

Fig­ure 1.0 – the generic Insight data model example

It’s like hav­ing a file cab­i­net filled with files based on each cus­tomer and their inter­ac­tions with your busi­ness. You can pull out the file on “Doug” and see how he is cur­rently inter­act­ing, what he did a week ago, or even how he engaged before mak­ing that recent purchase.

Fig­ure 1.1 – the “Doug” Insight data model example

The real ben­e­fit of orga­niz­ing the data in this man­ner is that it allows you to look at the infor­ma­tion from the cus­tomer per­spec­tive. In the exam­ple above, I illus­trate the point based on a sin­gle cus­tomer but Insight pro­vides access to that level of detail as it relates to all your cus­tomers and/or seg­mented groups.

Now that we have all this really well orga­nized infor­ma­tion what next?

Insight pos­sesses a rich trans­for­ma­tion lan­guage that allows one to effi­ciently mod­ify, adapt and cor­re­late val­ues within the data model. These trans­for­ma­tions are what we use to build the attri­bu­tion mod­els. Here’s a list of a few of the com­monly used and more pow­er­ful transformations.


The Copy trans­for­ma­tion is the sim­plest but most com­monly used. It allows one to copy an input value, field or sta­tic string, into a defined out­put field. The typ­i­cal appli­ca­tion of this trans­for­ma­tion is for recast­ing a field value, correcting/restating a field value, or cre­at­ing a flag based on a defined con­di­tional like iden­ti­fy­ing a mar­ket­ing touch.


The Cross­Rows trans­for­ma­tion is one of the most pow­er­ful and com­plex with the Insight tool­box. It allows one to copy an input from another row within the cus­tomer and out­put it to the cur­rent row. That includes prior and future records. The most straight for­ward exam­ple for a use of this trans­for­ma­tion would be cap­tur­ing the channel/name of the last mar­ket­ing touch prior to a con­ver­sion (or micro-conversion).


The Flat­FileLookup trans­for­ma­tion is very straight-forward and allows the map­ping of an input field value to N out­put fields. The typ­i­cal appli­ca­tion is to pro­vide that friendly name for all the dif­fer­ent cam­paign codes your com­pany is exe­cut­ing on.


The Math trans­for­ma­tion is another very straight-forward trans­for­ma­tion and enables the use of arith­metic oper­a­tions on fields. In com­bi­na­tion with Cross­Rows and other trans­for­ma­tions, the Math trans­for­ma­tion has allowed us to build out some of the more com­plex attri­bu­tion mod­els. Such as, Even and Starter/Player/Closer. Where only a frac­tion of the total con­ver­sion value is allo­cated to the involved mar­ket­ing touches.

In the def­i­n­i­tions above you prob­a­bly noted the Copy trans­for­ma­tion can be used to correct/restate a field value and Cross­Rows can mag­i­cally work on future records. What these refer to is how Insight builds its data model for analy­sis. Insight will read all orig­i­nal raw source data and at that time allow you to apply mod­i­fi­ca­tions to field val­ues as they are processed. Sub­se­quently when a new set of data is incor­po­rated within an exist­ing data model, Insight will com­pletely restate how that cus­tomer exists within the defined dimen­sions. So not only do the dimen­sions fully cor­re­late based on the data model exam­ple above they also always reflect the most recent view, or per­spec­tive, of each customer.

The real ben­e­fits for build­ing an attri­bu­tion solu­tion within Adobe Insight are the abil­ity to con­fig­ure the model and get access to the results today, no need to setup and wait, and the inher­ent con­fig­ura­bil­ity of the rich trans­for­ma­tion lan­guage to extend your under­stand­ing of the cus­tomer base and how they inter­act with your business.