The amount of time senior level executives spend discussing developing their database marketing infrastructure, at the expense of input, is stunning. While they will focus on building elaborate systems that combine the latest hardware and software, they often give limited consideration to the data contained within it.
Sure, lots of time is spent on data quality. But typically little attention is given to finding the optimal mix of data and variables needed to create data content, i.e. raw data that has been optimized to improve campaign performances.
Now, raw input is not a competitive advantage. Marketers need to understand demographic, lifestyle, behavioral, credit, summarized credit and automotive, life event, transactional, response data, segmentation tools and so on.
The key is to determine the appropriate mix of data that will make the most of marketers’ efforts. This leads us to the following data myth:
Data Myth: You have to conduct a campaign to determine the optimal data mix that will drive results.
This statement couldn’t be farther from the truth. With the appropriate infrastructure, access to campaign history data and analytical processes, a marketer can simulate the impact of new or different input on marketing communications.
Determining the optimal mix of content is not easy, and having the right access to data and the analytics to drive decisioning is critical. But by using a three-step data optimization process and following some best practices guidelines – in addition to analytics and predictive testing procedures — marketers can determine different combinations of data’s impact on their campaigns.
An Art and Science
Data optimization is both an art and a science. The “art” is knowing what data is available that can drive marketing performance. Admittedly, the landscape is cluttered: There are thousands of list brokers, hundreds of resellers, and numerous aggregators and compilers. In addition, determining what customer and partner data can impact marketing communications adds another layer of complexity.
Understanding the data available to marketers and knowing which items can be valuable can save a tremendous amount of time and money. Rather than try to become an expert overnight, a best practice is hiring an employee or partner who knows the data market and can help navigate through the clutter.
Once marketers have the necessary data, they can then use analytics to determine the optimal mix of sources and variables.
The “science” is having quantitative processes in place to create the optimal mix of data content. Data optimization can be looked at as a three-step process:
1. Source Optimization: In this step, sources are analyzed and indexed based on their ability to improve the models or mix of data that are driving DM campaigns. After this is done, marketers will determine which sources are likely to provide the highest lift and can limit their focus to “impact” data.
2. Audience Optimization: This process takes sources that survived optimization and augments the individual consumers or businesses with deeper data. This greater knowledge enables analytics that will determine the optimal prospect universe.
3. Campaign Level Optimization: This process determines the prospects that will be selected for an individual campaign. Campaign-level analytics are improved because data optimization has insured that the most impactful marketing content is being analyzed and used.
Best Practice: Ongoing Optimization
To make matters even more complex, creating an optimal mix isn’t a one-time decision; it’s something that needs constant work and attention. Data optimization has to be a continuous process. Too many times marketers wait until their campaign performance begins to degrade before reviewing the mix of data content used in their latest campaign or program.
Instead, a best practice is to instead focus on performance improvement by executing data optimization each time new campaign history data becomes available.
Craig Dempster is corporate vice president of content solutions at database marketing agency Merkle Inc.