One of my favorite chap­ters from renowned math­e­mati­cian Jacob Bronowski’s mag­num opus The Ascent of Man is titled “Knowl­edge or Cer­tainty.” In the chap­ter, the author uses a vari­ety of sci­en­tific devices and meth­ods to attempt to define the human face. After lengthy mea­sure­ments, tests, and results, the face is then put into a human con­text, which alters any sci­en­tific result that had been objec­tively mea­sured pre­vi­ously. You learn that the human being mea­sured is a pris­oner of war, which places all pre­vi­ous mea­sure­ments in a dif­fer­ent con­text from that pre­vi­ously per­ceived, cre­at­ing a dif­fer­ent per­son­al­ized per­spec­tive and mean­ing from what any objec­tive device is capa­ble of record­ing. Bronowski then begs that math and sci­ence, that mea­sure­ment itself, should never be devoid of human­ity or con­ducted out­side of a human con­text, which helps to inter­pret and accu­rately report its results. He quotes Oliver Cromwell in rela­tion to the sci­en­tific method, “I beseech you, in the bow­els of Christ, think it pos­si­ble you may be mistaken”

This exam­ple, and the pos­si­bil­ity of being “mis­taken,” is the fun­da­men­tal rea­son for iter­a­tive test­ing. No longer can we rely solely on pure data or inter­pre­tive hypothe­ses and hunches alone; accu­rate analy­sis requires a com­bi­na­tion of the two when deter­min­ing the right expe­ri­ence to deliver to a valu­able audi­ence seg­ment of our dig­i­tal vis­i­tors. As we’ve cov­ered in pre­vi­ous blogs and recent announce­ments at Adobe Sum­mit, hav­ing uni­fied Adobe Analytics/Target/Marketing Cloud data in a real-time aggre­gate pro­file view of your cus­tomer is essen­tial. This mul­ti­fac­eted under­stand­ing of the cus­tomer allows for greater clar­ity and accu­racy in mea­sur­ing and uncov­er­ing the most pre­dic­tive pro­file vari­ables and oppor­tu­ni­ties for con­tent per­son­al­iza­tion and max­i­miz­ing return on invest­ment (ROI) rel­a­tive to your pri­mary busi­ness goals.

This is why rely­ing on the basic out-of-the-box seg­men­ta­tion in a point opti­miza­tion solu­tion for an over­all view of your vis­i­tor pop­u­la­tion is not enough. You need to uncover the most prof­itable gran­u­lar oppor­tu­ni­ties, through man­ual or auto­mated means, based on all of the avail­able his­tor­i­cal and real-time data on your vis­i­tors. This allows you to gain an accu­rate view of where valu­able oppor­tu­ni­ties for per­son­al­iza­tion exist, and where they do not (equally impor­tant), to effec­tively define your strat­egy. Not only does this affect your imme­di­ate returns but it also improves growth and scal­a­bil­ity within your test­ing and tar­get­ing to increase effec­tive inter­pre­ta­tion of your results.

This means dig­ging deeper than the basic seg­men­ta­tion of new vs. return vis­i­tors, or behav­ioral tar­get­ing, to deter­mine if things like time of day, recency or fre­quency of vis­its, or geog­ra­phy play a valu­able role in deter­min­ing if one expe­ri­ence is more rel­e­vant and prof­itable than another. Often going a layer deeper and cus­tomiz­ing com­pound seg­ments with the most pre­dic­tive vari­ables can pro­mote larger increases in engage­ment and expo­nen­tial returns when tar­get­ing your vis­i­tors across your dig­i­tal touch points.

Gran­u­lar data and cus­tom audi­ence seg­men­ta­tion capa­bil­i­ties are not the only fac­tors that impact accu­rate, effec­tive test­ing and opti­miza­tion. Hav­ing detailed suc­cess met­rics in your test reports is also extremely valu­able. For instance, let’s say I’m only track­ing click-throughs from an email cam­paign test to the land­ing page, and I don’t look at addi­tional key per­for­mance indi­ca­tors (KPIs) down the fun­nel (e.g., inter­ac­tion with land­ing page offers, rec­om­men­da­tions, or even page depth/consumption/time on site). By using sin­gle met­rics or sim­plis­tic out-of-the-box tools I might inter­pret the data incor­rectly, or mis­in­ter­pret the human vis­i­tor impact of my test down the fun­nel. The email test vari­a­tion might have been suc­cess­ful in terms of gen­er­at­ing basic click-through, but caused unfore­seen fric­tion later in the cus­tomer expe­ri­ence if these met­rics were over­looked, or if the solu­tion was unable to cap­ture them because of a lack of customization.

Adobe Tar­get allows you to dynam­i­cally apply any and all suc­cess met­rics to ana­lyze the rel­a­tive impact of test vari­a­tions on vis­i­tor behav­ior. This is where it’s immensely valu­able to have syn­chro­nized Adobe Ana­lyt­ics suc­cess met­rics to apply to results, even retroac­tively, as enabled by Adobe Mar­ket­ing Cloud’s uni­fied def­i­n­i­tion and appli­ca­tion of audi­ence seg­ments and suc­cess met­rics between Ana­lyt­ics and Tar­get. This allows for fur­ther qual­i­fi­ca­tion of results with a broader, more detailed view of per­for­mance pro­vided by the full con­text of Ana­lyt­ics data. Tar­get also pro­vides the abil­ity to apply rev­enue and cus­tom scor­ing to suc­cess met­rics and page val­ues rel­a­tive to con­sump­tion. This pro­vides the rev­enue impact and detailed report­ing needed to empower you to imme­di­ately show the pos­i­tive impact of your test­ing and opti­miza­tion efforts rel­a­tive to the val­ues and rev­enue streams that mat­ter most to your exec­u­tives and com­pany. Tar­get is the deeply cus­tomiz­able, “plug-and-play” solu­tion that can imme­di­ately make you a rev­enue dri­ver and opti­miza­tion hero within your orga­ni­za­tion, espe­cially as you re-platform or redesign your site, launch new mar­ket­ing cam­paigns or appli­ca­tions, or look to imple­ment a cross-channel cam­paign solution.

For exam­ple, one of our clients is a large finan­cial insti­tu­tion whose main con­ver­sion goals are new accounts, new invest­ments, and cross-sells. The com­pany began with test­ing and tar­get­ing based on geog­ra­phy (what bank­ing oppor­tu­ni­ties were avail­able by region) and basic behav­ioral tar­get­ing (what cat­e­gory of invest­ments a vis­i­tor was view­ing). Nat­u­rally they saw some ini­tial impres­sive returns, which is always valu­able in fuel­ing your pro­gram. Given these ini­tial gains, many busi­nesses might have deemed this a suc­cess and stopped there. This finan­cial insti­tu­tion, how­ever, used Target’s APIs to reg­u­larly import updated online and offline branch pro­file data as well as third-party data such as credit score, to cre­ate a scor­ing table for pre­qual­i­fi­ca­tion of dif­fer­ent offers based on a customer’s level of invest­ment with the bank and other qual­i­fi­ca­tion cri­te­ria. The com­pany was then able to match the cus­tomer to the most rel­e­vant offer based on a matrix of pos­si­ble offers to dis­play. By doing this, they saw lift in hun­dreds of per­cent­age points in click-throughs from the home­page. Again, they could have deemed this a suc­cess and called it a day. How­ever, they were also able to con­struct and apply cus­tom suc­cess met­rics, such as cus­tomer life­time value, to deter­mine the larger impact of deliv­er­ing more per­son­al­ized offers in terms of their cus­tomers’ longevity and hap­pi­ness as well as the expo­nen­tial returns the bank was expe­ri­enc­ing through this rel­e­vant per­son­al­ized engagement.

A cross-channel view of all of these suc­cess met­rics, or cus­tom com­bi­na­tions therein, along with the expanded view of analytics-enhanced report­ing, gives you the best set of report­ing data on which to apply your human inter­pre­ta­tion, allow­ing you to be “less mis­taken” in your judg­ment as well as to accu­rately ascribe the busi­ness value and rev­enue lift that your detailed data-driven deci­sions have generated.