Predictive analytics exploit patterns found in historical and transactional data to help people identify risks and opportunities in their business. According to Gartner*, predictive analytics is of great interest to many organizations, but only a small percentage of organizations have made significant progress deploying it.

 

With such shining promise, why are many organizations yet to employ predictive analytics to improve their businesses? There are three main reasons for this:

  • Users are confused about what “predictive analytics” means. Generally, the term predictive analytics is used to mean predictive modeling, scoring data with predictive models, and forecasting. However, people are increasingly using the term to describe related analytical disciplines such as descriptive modeling and decision modeling or optimization. These disciplines also involve rigorous data analysis, which is widely used in business for segmentation and decision making but has different purposes and the statistical techniques underlying them vary.
  • It’s complicated. Predictive analytics stirs together statistics, advanced mathematics, and modeling, and adds a heavy dose of data management, to create a potent brew that many hesitate to drink. Many organizations do not know whether predictive analytics is a legitimate business endeavor or an ivory tower science experiment.
  • Business users are left out of the picture. In our view, this is the most common barrier to adoption. The tools and techniques for predictive analytics are relatively mature; however, business users do not know when and how to use them. Use of tools and processes for building predictive analytics and deriving insights from the data have been limited to a small number of highly trained and experienced statisticians and analysts. Business users are only end users who passively consume what others produce for them.

We believe that predictive analytics needs to be more pervasive to deliver significant value and competitive advantage to organizations. Predictive analytics should be part of a decision making process in which the predictive terminology should be familiar to the business users. The essence of predictive analytics is to predict a number, a category, or propensity. Business users should be able to use this functionality without memorizing algorithm names.

We think that the future of predictive analytics lies not only in statistical models predicting the future, but in the human aspects of prediction. A model may predict that 68% of potential buyers of a new product are college students, for example, but if 68% of your existing customers are college students, then this prediction doesn’t help the business a whole lot. Success requires that business users who have a deep understanding of the business, know the nature of the data, and know how to interpret the results own the process.

The bulk of the work in predictive analytics is in understanding the relationships in the historical data and using them to predict the future. The QlikView associative experience is a perfect fit for understanding relationships in historical data. It gives business users the flexibility to ask and answer their own questions and identify patterns and outliers in the data.

By using QlikView integrated with a third-party predictive analytics tool, users can get these same benefits with predictive analytics. This video shows a solution example that integrates QlikView with R, which is an open source language and environment for statistical computing and graphics. In this example, business users can conduct business discovery as they normally would with selections in QlikView. When they find an interesting data set they can click on a button that transfers the selected data set to R. R calculates the desired predictive analytics and the result set becomes part of the QlikView’s associative in-memory data model upon which the user can do further exploration. 

Gartner indicates that making advanced analytics available to an expanded set of users will require a new consumer-oriented approach. The analytics should be available at the point of the decision. The tools need to become more consumer-oriented, social, collaborative, and mobile. These characteristics are core to QlikView and with QlikView’s integration capabilities, business users can now do predictive business discoveries, expanding the Business Discovery horizon to the future!

 

 

*Source: Gartner Advanced Analytics: Predictive, Collaborative and Pervasive Report, February 2012