As much as I like music, I am sorry to dis­ap­point you but this blog is not about the Video Music Awards. We will not be talk­ing about Justin Timberlake’s reunion show with NSYNC or why Daft Punk didn’t win the best song of the sum­mer. Instead we will dis­cuss how to eval­u­ate your mar­ket­ing chan­nels in terms of the assets they bring to your busi­ness and how all this is pos­si­ble with Adobe Ana­lyt­ics Premium.

We will con­sider that all your chan­nels are already setup cor­rectly in Reports & Ana­lyt­ics. You have iden­ti­fied the chan­nels such as Email, Paid Search, Dis­play, Social, Nat­ural Search…. and set the proper guide­lines in the Mar­ket­ing Chan­nel Pro­cess­ing rules under the Admin Tools. If you need a refresher, check out this series. For the sake of this post, we will con­cen­trate on two chan­nels (Dis­play and Paid Search). Through Gen­e­sis Inte­gra­tion, you are able to bring in Dis­play and Paid Search impressions/clicks/cost data and put them against con­ver­sion met­rics in your report­ing. You are able to look at how well your cam­paigns are doing. How do we define “well”? This varies from one busi­ness to another but for sim­plic­ity we will con­sider that high conversion/transaction rates means “well”.

Look­ing at this in terms of an exam­ple, let’s assume that your busi­ness is a Music Store called “House Music Store (HMS)”. HMS is an online store sell­ing elec­tronic music and pro­duc­tion plu­g­ins. HMS is fully aware of the impor­tance of the dig­i­tal pres­ence and there­fore spends a con­sid­er­able amount of their yearly bud­get on dis­play and paid search cam­paigns. Through 3rd party ven­dors, HMS is able to look at impressions/clicks/spend met­rics side by side to the trans­ac­tions met­ric asso­ci­ated with these campaigns:

blog 1

You are happy to see that your dis­play cam­paigns are cost­ing less but result­ing in higher trans­ac­tions and in return mak­ing you more $$$. This prompts you to con­sider adjust­ing your spend­ing in next year’s bud­get plan­ning; you decide to spend a lit­tle more money on Dis­play cam­paigns and a lit­tle less on Paid Search. This is a valid way of look­ing at increas­ing your profit but what hap­pens if in real­ity Paid Search cam­paigns bring you high value cus­tomers spend­ing $100+ per month while Dis­play cam­paigns attract one-time buy­ers spend­ing less than $10 per year? This means shift­ing your bud­get towards Dis­play might not be the best move as you will miss out on the oppor­tu­nity of reach­ing out to high value customers.

This is all great but first, how do you iden­tify high value cus­tomers? Sec­ond, once iden­ti­fied how do you link these cus­tomers to the mar­ket­ing chan­nels to mea­sure their suc­cess? Don’t you worry, we got you cov­ered and this is where Data work­bench comes into the res­cue. All your cus­tomer infor­ma­tion exists in your CRM data­base. Every time a cus­tomer signs up online, a unique cus­tomer ID is gen­er­ated. In the back end sys­tem, you have a data­base relat­ing the cus­tomer ID to a list of trans­ac­tional attrib­utes such as the account ID, trans­ac­tional ID, total money spent and more. This will be sim­i­lar to this table:


Now we are look­ing at two sources of data, 1) The Web data col­lected by Reports & Ana­lyt­ics and 2) The CRM data stored in your back end sys­tem. As the account ID is a com­mon key between CRM and the WEB data, we can lever­age the Uni­fied Cus­tomer Process in Data Work­bench to tie these records together and roll them up to a top level key, i.e. cus­tomer id. Doing so, not only you are able to tie a customer’s web behav­ior and offline attrib­utes together but also dis­trib­ute your cus­tomers into dif­fer­ent bands. Based on these bands, you can build a visitor/customer level seg­ment to bucket them accord­ingly. In this case, we chose 4 bands:


The final step is to develop some cus­tomized met­rics to match the tiers in the Cus­tomer Bands. For example,

High Value Trans­ac­tions = Trans­ac­tions [Cus­tomer Bands = “High”]
Mid Value Trans­ac­tions = Trans­ac­tions [Cus­tomer Bands = “Mid”]

Now you are ready to look at your Mar­ket­ing Chan­nels again. This time, you can look a lit­tle closer on your trans­ac­tions in terms whether these were high value trans­ac­tions or a one timer:


This model can be eas­ily extended to other met­rics within your dataset. This can be applied to new accounts, attri­bu­tion mod­els and more. This will give you a broader view of the value of each mar­ket­ing chan­nel and the assets it brings to your business.