Data Collection: The Decision-EnablerHeads or tails. Rock-paper-scissors. What's the preferred decision-making method for your organization? Certainly you're not using these simplistic methods. If your organization is like most, you've amassed huge -- and ever-growing -- quantities of data.
But what do you do with that data? Are you getting real answers? Are you able to determine not only what happened in your business, but why it happened? And even more importantly, can you determine -- from your data -- what will happen in the future? Organizations that can answer yes to these questions are decision-enabled.
The decision-enabled organization relies on a systematic, repeatable process for making decisions supported by data. Moreover, the process takes place at all levels throughout the organization, allowing every decision maker to react quickly with supportable decisions in an age when the competitor is only a mouse click away.
To understand what makes an organization decision-enabled, it helps to look at the recent history of decision making. Decisions are based on experience. When businesses were small -- picture a grocery store in the 1940s -- the local grocer probably knew his customers personally and had a limited stock to maintain. When customers entered the store, he knew what they wanted and stocked the store based on his experience.
As our grocery store expanded over the county, out into the state and into national and then global markets, storing and sharing experiences became overwhelming. To help the grocer track its experiences, the concept of data was developed. Methods for quantifying and storing experiences were refined and through the use of computers, which are ideal for handling enormous amounts of information, data collection proliferated.
Now that the grocery chain could accurately record and quickly recall experiences, it needed a way to apply those experiences to daily operations. The immediate answer to that need was reporting. Through relatively one-dimensional methods like spreadsheets, organizations found they could generate analyses of past trends and simple reports of their experiences such as, "soda, toilet paper and beer were our highest margin products in the U.S. last year." This is still what most organizations do today -- beyond rock-paper-scissors, but far from decision-enabled.
With reporting, organizations gained a way, at the strategic level, to reference their experiences before making decisions about the future. They could now readily recall what happened in the past, but they still could not determine -- in an organized, systematic way -- why things happened, nor could they make accurate predictions or forecasts about future expectations.
Data analysis generally includes reporting, segmentation and forecasting. Reporting tells what segmentation and forecasting tell what and why and looks into the future. In the logical evolution of things, segmentation tools came along to help strategic-level decision makers predict behavior.
These tools take an organization's experiences of many types of business-related events -- sales transactions typically being of greatest interest -- and analyze them for significant patterns. This was a big breakthrough with implications for many functional areas and various industries, with direct marketing perched high atop the list.
Forecasting, though not a new method, soon began to receive attention as a process that could be approached in a scientific way. Organizations could now begin to determine, based on their experiences, what was likely to happen in the future.
All of these techniques -- reporting, segmentation and forecasting -- help organizations develop strategies supported by their experiences. Even in these organizations, however, a barrier remains to enabling decision-makers to make systematic and supportable decisions. You may have noticed that each of these techniques is discussed only at the strategic level. What is missing is the application of these techniques at the tactical level -- the level where things actually happen.
Reporting took a step in this direction with online analytical processing, a technology that allows certain people to build reports and deploy them to decision-makers at the tactical level. However, as I mentioned, reporting is the least sophisticated of these techniques because it offers only basic summaries of past experiences. It tells only what and not why.
Enter the decision-enabled organization. The decision-enabled organization has the ability to deploy the more sophisticated techniques of segmentation and forecasting -- the techniques that tell why something happened and what is likely to happen in the future -- to the front-line decision makers in every organization.
Imagine finding a forecasting model (built at the strategic level by an analyst) in your e-mail. (A model is basically an equation with many variables that simulates real-world events.) Part of that model includes all the data on every direct mail piece you have sent over the last five years. Among the data are how many pieces were sent, when they were sent and to whom they were sent. Additionally, the model contains response data and the respondents' demographic information. Using scenario analysis software, you open the model and see a forecast of response rates for the coming year. You ask the software to show what the response would be if you changed a few variables, such as mailing more or fewer pieces.
Before you mail anything, however, you need to know to whom you should mail. Consider this example, a model-builder (strategic level) creates a model using a decision-tree application that indicates which people are predisposed to performing a desired action. You receive the model and this time you open it in a scoring application. You now can apply the model to a list of prospective recipients and the software will automatically identify the people who match the profile of a respondent. You make your decision backed by quantifiable, supportable information.
Consider the applications of this technology across functional areas, as well. When people respond to a particular direct marketing piece, the customer service reps can collect more information from them and use a scoring application to evaluate any cross-selling opportunities. Even more advanced would be a scoring system that tracks and responds automatically to customers' Web clicks. With each click, the score for a customer becomes more tightly focused based on that person's preferences and actions, leading to truly individual, one-to-one marketing.
Organizations are driven by decisions. While it is up to each member of any organization to decide upon and manage objectives, in making your organization decision-enabled, you are providing it with half of the puzzle -- the best possible information from which to make decisions.
David Cody is a senior marketing manager for SPSS Inc., Chicago, a software company specializing in business intelligence and data mining. His e-mail address is email@example.com.