The overarching goal of testing and optimization is to effectively segment visitors and uncover profitable targeting opportunities, eliminating the guesswork involved in the personalization of the customer experience in order to maximize conversion and return on investment (ROI). Adobe Target, the optimization and personalization solution in Adobe Marketing Cloud, provides real-time marketer access, control, and insight in testing and optimization programs. Target assists marketers in not only seeing and capitalizing on real-time targeting opportunities based on best-performing content at a granular level, leveraging all available data, but also managing and growing a program efficiently based on Adobe’s proven maturity model, industry successes, and best practices.

Even in a full-service model, where Adobe Consulting manages the entire program (delivering recommendations and results from the start), the concept is to educate and provide your team members the opportunity to have ownership over their program. Initial quick wins are evangelized, further promoting a testing and optimization culture throughout the business and eventually, when team members understand the strategy, techniques, and execution of testing and optimization, they can assume a self-service model and run the program themselves.

Control is a key factor in driving the right optimization based on your specific business initiatives, as well as reacting in real time to the trends you see in visitor data. This is why you should be skeptical when a point solution pushes a full-service model as its sole option.  You should be wary of what appears to be, or is presented as, “marketer control” in a solution when it really is not. For example, true marketer control is not  provided when a point solution has you fill out a simple form field request in the test design that then requires a vendor to code and to execute document object model (DOM) manipulation behind the scenes (essentially relinquishing control and execution to an outside vendor for questionable return).  The real reason why point solutions will often push a full-service model is that their product and user interface (UI) are cumbersome and confusing. These “solutions”  often rely on custom coding and development work for anything above a basic button color change, text swap, or simple test on a static element.  Even more marketer control is lost in this scenario because most websites are not built or coded cleanly enough to enable easy test creation with solutions that provide limited DOM manipulation in their “what you see is what you get” (WYSIWYG) editors.

These providers also lack experience and are not confident in their ability to teach or to train you in the best methods for testing, optimization, and program growth.  Adobe Target is designed for growing a program, and is built on a foundation of our proven maturity model, which guides businesses from basic incidental testing to advanced personalization and an embedded optimization program.  Our Adobe Target Solution also provides guardrails within the visual onsite editor n that allow for less reliance on coding for more dynamic elements, with easy instructions on how to execute more advanced testing and targeting scenarios without necessarily requiring coding or development resources.

There is a fundamental concern for the accuracy of your tests when control is relinquished in your optimization program. However, this fear is unfairly extended to the use of automation, perhaps because the algorithms provided for automation by most vendors are delivered as a “black box.”  In this scenario, algorithms are merely selected by the marketer with little or no customization capability and run “beneath the hood,” obscuring marketer insight into how or why the solution is rendering a specific decision.

A major differentiator for the Target solution, in terms of our out-of-the-box recommendations algorithms, is our commitment to enabling the marketer to not only customize the location and data being leveraged for real-time decisioning, but to run tests to determine which custom algorithm performs the best in different scenarios. This ability will be extended with major updates to our recommendations UI and algorithms beginning in August.

Our modeling system, which was renamed automated personalization with our Adobe Target Premium June 25 launch, also provides insight into how it self-learns, within its residual variance framework, uncovering the most predictive variables to leverage for personalizing an experience to an individual based on a series of distinct possible content variations or offers. The reports automatically generated by the modeling system are also improved with the new Premium release. They are much more graphic, and provide detailed, advanced segment discovery by testing and determining what are, or just as importantly what are not, the relative influence or predictability of hundreds of distinct profile variables. In addition, we offer a detailed description of how our modeling system self-optimizes by constantly testing its effectiveness to ensure that it adapts to trends and changes in customer behavior for the most effective personalization to the individual.  This level of accuracy safeguard is not found in point solutions that claim to offer “automated behavioral targeting,” nor is there the depth of profile data and variables provided by other solutions that can be found with our unified master marketing profile (a modeling system is only as intelligent as the data fed into it). This includes data synchronization across Adobe Marketing Cloud solutions along with a series of APIs that provide ease of regular data import and updating.

Adobe Target Premium lets users measure and optimize its automated personalization offering in much the same way that they can test and optimize the out-of-the-box recommendations algorithms.  You can now choose between residual variance and random forest algorithms, which provides an alternative automated statistical technique to fine-tune the best automation for more scenarios across more of your digital properties to free you up for those opportunities where manual or rules-based targeting control is warranted. More algorithms will be introduced shortly for even further flexibility and managed algorithm and automation control.

When you can harness various methods of machine learning now available in Adobe Target Premium, you can identify opportunities for personalization and maximizing visitor engagement more quickly, increasing conversion and loyalty while not relinquishing the understanding, control, and confidence over how these powerful tactics are being used to drive even quicker success for you and your program.

0 comments