In recent years, the evolution of technology and simultaneous growth in data availability have enabled a revolution in customer-focused decision-making. Decisions have become increasingly data-driven and the time frame for these decisions continues to accelerate. Marketing Analytics is the science that connects the vast amount of data with intelligent marketing decisions. Marketing Analytics compiles inputs from all parts of the business, processes all relevant linkages and uses this new information to drive optimal marketing decisions. A company's Analytic Maturity is a measure of the extent that the key business components (Data Integration, Decision Making and Operationalization) are fully integrated across the enterprise. An analytically mature company is much better positioned to conduct effective marketing analytics.

Successful customer acquisition and retention depend upon customer understanding. Customers' needs, wants, sensitivity to offer details, channel preferences, price sensitivities, risk characteristics and demographics are all important factors in the creation or maintenance of a profitable relationship. Finding all the relevant data is one of the biggest ongoing challenges for marketing analytics.

There are several categories of data that can lead to an understanding of customers or prospects, and how they might respond to new products and offers:

• Past marketing efforts - How have they been reached in the past? What were the results, including whether they responded, how much they spent, how long they remained a customer, exactly what they bought, whether they met risk criteria, etc.

• Existing customer relationships - Depending upon the business, individual customers may have many interactions that provide a lot of useful information: transaction history, levels of spending, payment history, etc.

• Third party data on customers and prospects (whether or not they have already been contacted) - Demographics (both neighborhood and household), credit information, shared information from other marketers, and additional information that can be inferred (for example, a recently married couple might soon be in the market for baby products or a a prospect with little or no debt and a good income might be in the market for investment products.).

• New data generation - Using focus groups, surveys or test marketing (this is a form of scientific experimentation and can be very valuable because it is controlled). It can be used to test the effectiveness of product design, message, channel and price; compare different settings of these parameters; identify overall optima or customize offerings to specific segments (also identified through the same experimental process).

Whatever the data source, marketing analytics describes the approach to appropriate analysis and interpretation that drives effective marketing decisions. This is a complex multifaceted process whose components include:

•  Structuring the data - including integrating data from multiple sources

•  Understanding the data - including profiling and exploration of relationships among all variables

•  Preparation for analysis - including systematic approaches to missing imputation, outlier treatment and derived variable creation

•  Statistical data analysis - including selection of techniques that enable reliable conclusions in the specific context

•  Validation - including confirmation of initial modeling results through testing on independent samples

•  Interpretation - including extrapolation from analytic results to broader business applications

•  Implementation - transforming insights into business decisions

Marketing Analytics impact information management and decision making throughout any organization. But a one-size-fits-all solution is not appropriate. The key dimensions of an analytically mature solution to marketing analytics for companies in different industries are quite different with respect to the optimal degree of integration of data across departments and processes, the need for decisions to be centralized vs. decentralized and the extent and time scale in which operational issues need to be closely linked to one another.