Microsoft Dynamics vendors provide comparisons and opinions to professionals in the ERP/Accounting software selection process

 
 

deFacto Global

Predictive Analytics: Crossing the Chasm


Email | Print

Many technologies follow a curve in which the original products are complex and expensive and can be deployed by only the largest and most resource-rich companies, but gradually become more widely adopted as they become more affordable and easily deployable. This is happening in the realm of predictive analytics, aided by new machine learning tools and data sets available in Microsoft’s Azure cloud.

Predictive analytics is an area of data mining in which computers use machine learning and statistical analysis to find patterns and predict trends. Predictive analytics typically involves collecting and operating on Big Data.

While predictive analytics can yield real advantages, its use has been restricted by the scarcity and high cost of the three main elements required:

  • Data science experts for developing predictive models.
  • Large data processing capacity.
  • Large, clean data sets.

As Forrester analyst James Kobelielus observes, the core problem with the first generation of predictive analytics offerings is that “many of them remain power tools with a steep learning curve and a commensurately high price.” The key trend, he says, “is the move toward user-friendly, self-service, BI-integrated predictive analytics tools that encourage more pervasive adoption.”

Barriers Fall

As in many areas, the cloud is proving to be a big benefactor in making predictive analytics more widely available and affordable—including tools, applications, storage, and computing power. Microsoft’s Azure Machine Learning, which recently went live, provides a cloud-based toolset that makes predictive modeling easier, as well as providing inexpensive processing capacity and large external and subject-focused data sets.

Microsoft vice president Joseph Sirosh, the executive in charge of Azure Machine Learning, said the cloud solves the “last mile” problem. Before a service like this, he explained, you needed data scientists to identify the data set, as well as an IT team to build and support the application. The last part alone often took weeks or months to code and engineer at scale. Azure Machine Learning simplifies the process and provides a way to build the same application in hours, he said.

Opens the Door

The good news is that businesses of all sizes, and across all sectors, can now benefit from employing predictive analytics for a wide variety of uses. The availability of easier and more affordable predictive analytics tooling and data sources in the cloud also gives ISVs opportunities to work with clients to develop predictive analytics solutions for particular applications and industries.

As Barb Levisay reports, the low cost and high availability of computing power in the cloud is bringing machine learning to the masses. Microsoft Azure Machine Learning, she notes, “opens the door for any business and business partner to experiment and test without a big investment.”

Uses Will Expand

The current use of predictive analytics spans many industries and applications, such as marketing, sales, credit scoring, product pricing, insurance risk assessment, healthcare treatment, fraud detection, security, and more.  Real-time data as well as historical data can be mined to provide valuable insights that can reduce risk and inform strategy.

In sales, for example, a predictive model can be used to measure the probability of a lead converting to a sale. In marketing, predictive analytics can help businesses gauge customers' buying habits to promote the most relevant products at the right time and in the right channels.  In healthcare, predictive analysis can be used to diagnose and determine best treatments, as well as to determine which patients are at risk of developing certain conditions, such as diabetes, asthma, or heart disease.

One of the best-known applications is credit scoring, which is used to model an applicant’s credit history and other fiscal data to gauge credit-worthiness. Real-time uses also include profiling customers through their buying habits and social media activities to offer special deals and promotions as they are searching or shopping.

As predictive analytics expert Martin Hack observes in Wired, machine learning is vastly superior to traditional data processing for particular tasks, such as determining the reasons behind customer churn.

Ease Is the Key to Universal Access

Recognizing the advantages businesses can gain from predictive analytics, Gartner analysts forecast that organizations that use predictive performance metrics will increase their profitability by 20 percent by 2017. With more organizations able to perform predictive analytics, more organizations will be able to reap these benefits, and many new uses will be discovered.

But as Martin Hack observes, “Ultimately, for machine learning to impact the world around us in a truly meaningful way, we have to deliver machine learning in a smarter, more usable form.”

This is where Azure Machine Learning and deFacto's platform can serve as complementary pieces, with Azure Machine Learning providing all the resources needed to build and operate predictive models, and deFacto Business Modeler providing all the resources needed to create business models to perform advanced forecasting. Together, these two solutions give business managers new levels of forecasting and planning capability.

To learn more about the deFacto Business Modeler platform, read the whitepaper “The New World of Financial Analytics.”

By Michael Neubarth, deFacto Global

Comments are closed.