A Tipping Point Prediction: More Predictive Analytics
From every direction, predictive analytics will prevail and pervade.
Business intelligence and decision support have been utilized by organizations for several decades. Their deployment continues to spread in both breadth and depth within many of these organizations while gaining new converts in others. However, data mining, and its application as predictive analytics, has often been characterized as a technology that could only be utilized by highly skilled technical practitioners with strong statistical backgrounds.
This may have been true in the past, but is arguably no longer the situation today. While skilled technicians will continue to be needed for developing and evolving new data and text mining algorithms; and while the services of these top-notch professionals will certainly be required for the creation and validation of useful models; the bottom line today is that non-technical business users can now apply these algorithms and models in their jobs, too.
There are several reasons for this development. Over the past several years, enterprise applications vendors have taken steps to embed business intelligence functionality within their applications. While this certainly benefits the companies that license these applications, it also allows these vendors to better participate in the growing analytics market. After all, operations and analysis are two sides of the business coin, and major enterprise applications vendors including Oracle and SAP want to increase the revenue that derive from analytics, perhaps at the expense of their BI partners.
When BI functionality is embedded within operational systems, that empowers non-technical business users to deploy the technology, especially if they can run pre-defined analyses and reports as part of their normal workday, without needing to understand the underlying theory. As predictive analytics functionality is added to these applications, more users will become familiar with its benefits; they’ll almost assuredly discover how PA can help their organizations. Furthermore, vendors of analytic applications are increasingly incorporating predictive analytical capabilities, in addition to the traditional OLAP analysis and reporting mainstays.
Microsoft’s recent announcement about the availability of Microsoft Dynamics CRM Analytics Foundation, a set of business intelligence tools that Microsoft users and partners can utilize with Microsoft Dynamics CRM, is one sign of this larger trend. Among the BI functionality included in the Analytics Foundation are predictive capabilities that leverage the data mining functionality of Microsoft SQL Server Analysis Services. Microsoft has stated that it intends “to make business intelligence an integral part of a company’s everyday activities, and to enable better decision-making for everyone, from the executive to the information worker,” and this broad mission statement includes predictive analytics. Since Microsoft has the market pull to popularize technology, its actions could serve as a catalyst to help broaden the use of data mining and thus expand the overall market itself, thus achieving the proverbial tipping point.
This will certainly benefit data mining specialists such SAS and SPSS, while placing pressure on other business intelligence specialists to broaden their BI platforms to include additional data mining functionality. It will also encourage application vendors that have not done so already to expand their included business intelligence functionality and incorporate predictive analytics as well.
Many companies that utilize predictive analytics are reluctant to publicize their successes, likely because they believe that such solutions provide a significant competitive advantage. Odds are, many published estimates of data mining usage are on the conservative side. However, as applied data mining or predictive analytics moves into the hands of business users willing to tout their successes, the usage estimates will likely increase. Thus, as the overall market for business intelligence continues to grow, the utilization of predictive analytics will grow at a slightly higher rate.
One word of caution is that while business users deploying embedded predictive analytics within enterprise applications may not necessarily need to understand the underlying data mining theory, they should still confirm that the results make sense. As the use of predictive analytics spreads to non-technical business users, these folks may not have the necessary background to understand that a strong correlation is not always evidence of causality and can sometimes result from random coincidence. However, business users do tend to have the business knowledge, experience, and domain expertise to appreciate whether or not the predictions are reasonable. They should not be afraid to question results that seem out of place, while embracing those that provide meaningful insight.
About the Author
Michael A. Schiff is a principal consultant for MAS Strategies.