Understanding the basics of predictive analytics is as easy as saying your ABC’s:
A: What is it?
As any fortune-teller will assure you, predicting the future is an art as well as a science. And that truism applies as much to business decisions as to any other aspect of life. So on the one hand we can say that predictive analytics is a branch of statistics in which information extrapolated from historical data is applied to the projection of future conditions. That’s the science side.
On the other hand, we can say that predictive analytics is using information you do have to compensate for information you don’t have (yet), in order to make better business decisions. That’s the artful part, and it can depend as much on intuition and imagination as on algorithms.
Bringing the two sides together successfully is “what’s new” in the practice of predictive analytics. In the past, processes utilized for extracting and organizing the information contained in business data were so complicated that business analysts had to work with limited amounts of rigidly structured material. And that left not much room for creative exploration, innovative ideas, and intuitive leaps.
Now, however, fast and flexible tools (utilizing in-memory processing) have finally unleashed the power of What-If.
B: Why does it matter?
For example: You can’t know exactly how many widgets people will buy next year—but you do have a lot of data that could be used to improve projections about next year’s widget sales, and thereby optimize planning for widget production. Somewhere in your organization, there is data about how many widgets were sold to what kinds of customers in each of the last five years. And there is data about manufacturing costs in relation to production volume. And there is data about purchasing patterns that might reveal emerging trends in widget usage. And . . . well, you see the point.
Although some amount of predictive modeling can be done by means of pre-formed queries, pivot tables, summarized reports, and so on—that’s just rudimentary. The ultimate value of all that data will only become evident if the people who know how to think about the data can access it easily, explore different views, test various hypotheses, and share their findings effectively.
So—with the right tools, our widget analyst might be able to perform ad hoc data searches, summarize results into visual displays, repeat the process using various scenarios, add data from external sources, and create a range of projections for group review. All in nearly real time, with no need for IT intervention.
Another advantage: With fast, user-friendly tools, analysis can be done by (or at least with) the people who understand the data, rather than the people who understand the database. And that’s a huge plus, because in-depth business knowledge is a major key to success in predictive analytics.
C: What next?
While predictive analytics is not actually “all things to all people”—it’s such a universally useful tool that there may be as many definitions and methods as there are users. And the range of business applications includes decision support, customer relations management, planning, and risk assessment, just to name some of the most obvious.
For a nice overview of the topic, plus analysis of the vendor landscape, get a free download of the Forrester Wave report “Predictive Analytics And Data Mining Solutions – Q1 2010.“ Then, for a high-level look at the diverse interests and viewpoints connecting around PA, browse the Predictive Analytics World website.
Subscribe to our blog to stay informed on predictive analytics and other data analytics topics.
Image courtesy of Skeptic.com