Technology has transformed information—how it is collected, where it is housed, who has access to it, and how it is used. We have become so accustomed to instant credit card approvals, real-time Internet, and light-speed email that technology’s extraordinary role has become so routine that it goes almost unnoticed by many of us.

But it’s also frightening to some that as we go about our daily lives data about our behavior, buying patterns, travel, likes, and dislikes is being collected, stored, and increasingly put to use by marketing companies, credit card issuers, credit bureaus, and thieves. Here is a look at what is happening to all that data.

Database research (aka data mining)

Mechanisms for in-depth behavior modeling and outcome prediction have long been in place in fields such as operations research, finance, and engineering. Now, these powerful analytical tools are being increasingly used in marketing.

Traditional market research offers valuable insights into business-to-consumer likes and dislikes, but it does not study the interactions of the variables that may explain the success or failure of a product or service. Database research—or data mining—does.

Data mining takes raw facts that lack much meaning by themselves and turns them into business intelligence by grouping, categorizing, and segmenting them in ways that uncover patterns and cause-and-effect relationships. As technology transforms information uses, the need for such vigorous research in the marketing field becomes more apparent and more widespread.

Data mining and direct marketing

Database research is a particularly powerful tool in direct marketing, a discipline that produces large volumes of response data that can be analyzed. Direct-marketing practitioners know precisely how many pieces mailed, to whom, and what lists, offers, and creative were used. Database research combines that data with knowledge of who and how many responded to derive segmentations in the market.

Every addition of appropriate data can further define and segment markets making possible more efficient and effective campaigns. But technology has facilitated the advance from market segmentation to response modeling and consumer profiling.

Multivariate Response Analysis

With today’s enormous computer capacities and ever faster processing speeds, deeper and deeper marketing evaluation has become the norm. Today’s sophisticated data-mining tools enable performance examinations from multiple perspectives. Called Multivariate Response Analysis, this technique involves the simultaneous examination of several factors and determines the degree of impact of each element on one another and on the overall outcome.

There is, however, an important caveat to these increasingly complex research methods: most anything can be modeled, but not everything should be.

As analytical models become increasingly complicated, if the input variables used to predict outcome are irrelevant, the models will generate misleading results. If a model suggests a conclusion that defies logic and common sense, it has probably been corrupted by variables that have no real cause-and-effect relationship.

Taking analysis a step further – Response Likelihood Models (RLMs)

This advanced analysis can drill down into data to discover which elements optimize returns. Still more elaborate models can also reveal the likelihood of responses to individual elements.

Response Likelihood Modeling ranks input data by its degree of influence on response, which is valuable for limited budgets because there is a minimal sacrifice of response rates in exchange for much larger reduction in costs.

RLMs also predict response for even the smallest contributors to outcomes so they are particularly effective analyzing highly targeted campaigns.

RLMs can be valuable in examining multiyear renewals or premium member acquisitions. RLMs can help identify which of the usual suspects—net income, education level, dollar value per net order, and occupation—have the greatest influence on acceptance of a new and different membership offer.

Analysis over the long term

Of course, the process of acquiring data and analyzing it is hardly new. What is interesting today is how robust the process has become and how advanced analysis takes advantage of the powerful tools that new technologies provide for more and better understanding.

Raw data is largely useless, but when properly analyzed, it becomes knowledge. Knowledge, in turn, becomes business intelligence, which drives more informed decisions that should improve returns. The trends are clear—as technology continues to amaze so will we all be amazed—or intimidated—by the increasing quality and quantity of data mining and how it turns facts into knowledge to make better decisions and marketing ever more effective and efficient.