Service Trends

Blog archive

Azure ML Leading Microsoft Partners, Customers into the Future

With the Microsoft Azure Machine Learning (Azure ML) preview now available, Microsoft partners are beginning to explore how predictive analytics can help their clients.

Supporting tactics used by retailers like Amazon.com and Target, the predictive analytics made possible through Azure ML allows partners to change the game for businesses of every size. An emerging opportunity, Azure ML can enable partners to help even the smallest clients mine data to uncover patterns and build competitive edge.

The definition of predictive analytics, according to Belinda Allen, a Microsoft MVP and business intelligence (BI) evangelist, is the use of a BI technology to produce a score for each customer, prospect or product that will help you decide what will happen with more accuracy than guessing.

Allen uses the example of Amazon.com to illustrate the practical implications of predictive analytics. As every Amazon.com shopper has experienced, suggestions for items related to the products that you have previously viewed or purchased are offered when you visit the site. Those suggestions are based on historical data and the predictive-analytic algorithms that Amazon.com has built over time.

"One of the reasons that predictive analytics are so important is that people are buying things so differently than they used to. They want products targeted directly to them," Allen explained. "Predictive analytics allows us to learn from our data, learn specifically about our customers, and how to treat each customer individually -- which gives you competitive advantage."

Most of the clients that Allen, who is also principal of Smith & Allen Consulting Inc., talks to about predictive analytics are in retail or sell physical products. In those situations, the clients are moving from making decisions based on purely historical reports to a more predictive approach. Retailers are leading the way, understanding that it costs far less to sell to a current customer than to acquire a new one.

At reImagine 2014, the Dynamics conference held recently in Fargo, N.D., Allen introduced the concepts of predictive analytics to customers and partners. Allen explained how predictive analytics can solve business problems and, even more importantly, how it develops over time.

"I wanted to show them that these are not projects with a start and finish. It's an ongoing process that builds," Allen explained. "The more data that you have, the more you can do with it."

Allen cites three fundamental requirements to support a predictive analytics project:

  • A clear definition of the kind of information that you want to predict
  • An understanding of how you will use the predictive scores that you gather from the data
  • Good, clean data

During her presentations on predictive analytics, Allen walks through a simple example on the Azure ML site using sample data available through Azure data services. A recording of one of her presentations is posted here (while the entire presentation is valuable, the Azure ML demo starts at 30:20).

As Azure ML gains visibility, Allen sees growing enthusiasm in the partner community. "They are super excited about what predictive analytics can bring," she said. "For example, there is a new feature in Microsoft Dynamics GP sales order processing that supports pop-up suggestions to encourage add-on sales. It would be ideal for an API from Azure Machine Learning -- an ever-evolving data set -- to populate that field."

The Azure ML preview provides an easy way for partners to get a head start on the emerging opportunity in helping every-size company put their data to work. Whether you work with small companies or enterprises, you will add value and cement your relationships with clients by helping them tap data to predict behaviors and compete more effectively.

How are you helping clients tap the potential of their data? Add a comment below or send me a note and let's share your story.

Posted by Barb Levisay on December 04, 2014 at 9:34 AM