As database services become increasingly commoditized, the value of databases depends more and more upon advanced consumer analytics. The technology exists both for aggregating data to input into analyses and for acting on results of such analyses in a multitude of ways and through a variety of channels.
Although database technologies provide great power, users are painfully aware of the boondoggles that have accompanied these advances: increased complexity, costs and inefficiencies. Therefore, the goal of analysis in this scenario is to capitalize on the opportunities inherent in databases while reducing complexity and streamlining processes.
Achieving this goal requires embracing both the science and the art of marketing. A host of terms describes the basic dichotomy that has traditionally defined the approaches: “bottom-up” vs. “top-down” processing, “data-driven” vs. “theory-driven” discovery, or “inductive” vs. “deductive” reasoning.
In a recent research note, Forrester Research analyst Eric Schmitt delineates the ascendance of “left-brain marketing,” which he defines as:“The rise to predominance of analytical marketing strategies, skills and processes that are centered on audience knowledge, not media.”Schmitt explains that as marketers turn increasingly to left-brain marketing to target harder-to-reach segments who have access to a plethora of media options (and who often use two at the same time, such as the college student who surfs the Internet while watching a college football game), traditional right-brain marketing must work in conjunction with these left-brain market processes to find inventive ways to land these elusive consumers.
The Need for Data Context
However you choose to describe it, the general premise is that the most effective approach uses left-brain and right-brain processes in a complementary fashion. The two sides interact in a sort of feedback loop to constantly revise and reinforce expert opinion with the results of data analysis: deductive or right-brained processes based on expertise to generate hypotheses regarding the impact of programs or tactics on consumer behavior, and inductive or left-brain processes that provide analyses that support or refute these hypotheses.
This complementary process allows the building and evolution of models or rules of behavior that are fundamental to the effective and efficient management of customer information, allowing businesses to adapt to exceptions to rules and to capitalize on opportunities when rules are robust. These behavioral models are best derived from a combination of a top-down approach or process that is built upon sound hypotheses or business cases, and a bottom-up or data-driven approach. Data alone is not sufficient; data needs a context.
The process of building models of consumer behavior entails implementing rounds of iteration between applied business scenarios and data collection/analysis, with the goal of formulating a methodology to generate rules designed to govern the manner in which information ultimately manages the customer experience. Practical scenarios provide a framework for applying data analysis techniques to derive rules.
To effectively integrate a bottom-up/top-down approach, keep these practices in mind:
Listen to the right brain. Know what your marketers are trying to accomplish with their programs. Drill down into what specific behaviors they are trying to elicit with a given tactic or combination of tactics. This knowledge is made visible through the generation of hypotheses or the creation of business cases, and requires a thorough understanding of the marketing program and its key drivers. Analysts and business users must work closely together to generate sound objectives and translate those objectives into the required metrics.
Data analysis needs a well-defined business context to be effective. Embedding efforts in a “real-world” or applied context helps identify and evolve the data and information needs upon which both sound data infrastructures and valid models are built for managing customer relationships — building upon results and learnings, employing a phased project methodology that allows for this acquired knowledge to shape and influence strategic direction and decision-making, while providing insights to ongoing model development.
Start small. Because of the scope and complexity of database projects — with multiple entities, constituents, data sources, delivery channels and applications to be considered in solutions — it is difficult to articulate a strategic vision while integrating the complex dynamics of tactical deployment. A challenge is to keep near-term goals and project management issues clearly in sight, while simultaneously developing the larger vision.
The business case approach meets this challenge by:
- Providing a business context that is both manageable and meaningful
- Allowing solutions to be built in an iterative fashion, continually refining processes and incorporating results along the way
- Keeping users and key stakeholders engaged and informed.
Test. The development of models of consumer behavior must be guided by a set of tenable hypotheses embedded in business contexts that are constantly tested and refined. The need for generating and evolving behavioral models based on relevant patterns in the data is fundamental to enabling personalized communications and services across channels. Effective solutions require a set of rules that facilitate the automation of processes, are flexible and update as conditions change, and that implement best business practices across systems and channels. This is enabled by systematic testing of these hypotheses, which is underutilized yet increasingly necessary in today’s complex markets.
Standardize. The need to provide meaningful intelligence for effective communications, and to anticipate user needs in advance, dictates that current efficiencies and knowledge be capitalized and expounded upon. The key to connecting present with future conditions is to define a set of standard procedures for generating and evaluating relevant and timely business cases that are consistent, replicable across situations, valid and reliable. For a methodology to be effective, it must accurately identify data, metadata and rules that are critical the business and to strategic initiatives; and do so in a manner that is consistent across time, functions and situations.
Communicate and educate. Inform and refine current key projects and initiatives by using them as tutorials or practicals. Continuously educating the user about the benefits of the testing or contextual approach through the use of current projects and strategic initiatives as tutorials will elicit their needs as well as keep them engaged and informed.
This practice also will enhance key projects and initiatives through the systematic evaluation of project objectives. A dual purpose, therefore, is being served in helping to build behavioral models and refine processes for ongoing program development, and in providing important information to business users to refine objectives.
Model development requires a continuous dialog between users, analysts and developers: business users to generate the business cases for context, analysts for data requirements and computational rules, and developers for implementation through declarative rules.
The testing or contextual approach to model development focuses on the information needs of the business user in managing customer relations and the cross-functional similarities and differences in those needs. It enables a greater understanding of the current business situation and consumer behavior, and is intended to stimulate thinking about possible directions to take to maximize opportunities and to avoid pitfalls when planning programs.
With the days of mass-media and marketing to the averages behind us, left-brain practices to make sense of all the data and media fragmentation are increasingly important. Yet right-brain thinking and cannot be disregarded. While people might debate which group is more important, the synergy of the two (the yin and yang) seems essential for meaningful and practical results.
Katie Cole is vice president of analytics and research for Quris, a customer-centric email solutions agency forFortune 1000 companies. Cole has over 20 years of in-depth knowledge and experiencein data mining, statistical modeling, primary research and softwaredevelopment in a variety of industries including telecommunications,broadband, Internet and financial services.