New terminologies surrounding the focus on data management in manufacturing seem to spring up more and more often. In the past few years we’ve seen terms like Big Data, Internet of Things, and Cloud Computing come to the forefront of everyday manufacturing discussions. Now there’s a new term to add to your lexicon—Data Smart.
Beyond its various uses in company names, common use of the term “data smart” to convey the smart use of collected data in business seems to be of recent vintage. 2014 examples include a Data Smart book by John W. Foreman and “Data-Smart City Solutions”, an initiative by the Ash Center at Harvard Kennedy School supported by Bloomberg Philanthropies.
On the manufacturing front, Microsoft is positioning itself as key resource on data smart manufacturing by launching a site focused on helping companies take their first steps with smart data management.
To better understand how Microsoft is approaching the data smart manufacturing concept, I spoke with Sanjay Ravi, worldwide managing director for discrete manufacturing and industrial industries at Microsoft.
AW: How does Microsoft define a “data smart” manufacturer?
SR: Being data smart is ultimately about using data to drive business transformation and results in a connected world. For manufacturers, being data smart is not so much about the volume of data they can collect, but how they can foster a data-driven culture across the entire organization and use data to make smart and rapid decisions at the pace today’s business environment demands. A data smart manufacturing organization looks at not just how to improve business processes through the use of data, but how to drive new business models and services using insights gleaned from data.
AW: What’s needed in terms of technology and associated processes to become a data smart manufacturer?
SR: From a technology perspective, manufacturing companies need to evaluate platforms that offer them the most comprehensive capabilities across the data spectrum—from the ability to capture and store data to robust data analytics and rich visualization. They also need to look at the tools they are reviewing for purchase to ensure that they are enterprise ready, support hybrid cloud models, and offer ease-of-use and familiarity.
AW: More specifically, what kinds of analytics tools are needed?
SR: There are four key capabilities that are defining the new and improved set of big data tools and technologies:
• In-Memory analytics greatly improve the speed of data processing by performing analytics actions in-memory.
• The ability to combine structured and unstructured data streams to provide comprehensive insights.
• Cloud machine learning—building machine-learning capabilities into data tools that can help with predictive modeling and analytics.
• Streaming data—the ability to process and analyze streams of data from devices, such as sensors, in real-time.
Microsoft’s data platform supports all the capabilities outlined above. For example, HDInsights is a cloud-based data platform that manages data of any type, whether structured or unstructured, and of any size. The data types HDInsights can handle range from OLTP (Online Transaction Processing) in-memory technologies in SQL Server 2014 to StreamInsight and machine learning capabilities that allow you to monitor your data from multiple sources and visualize meaningful patterns, trends, exceptions, and opportunities, as well as offering predictive analytics capabilities.
AW: Once a manufacturer determines they have the right technologies and processes in place to become a “data smart” manufacturer, how should they go about setting initial goals to get a measurable return on the data?
SR: The three most critical business imperatives for manufacturers today, based on what we hear from our manufacturing customers, are to innovate, perform, and grow. Data can be a huge asset in addressing all three, but before diving into any big data initiatives, manufacturers need to clearly define the business outcomes they’re looking to drive, making sure they don’t limit their thinking to existing business processes only. Once objectives are clear, look for opportunities to apply big data principles to existing data assets. Going through that exercise will very quickly reveal where your data gaps are and what additional data sources you need to tap to get to the final insight that you’re looking for. Lessons learned from applying data smart approaches to solve one specific business problem or pilot project can then be carried forward to other initiatives as you expand use of data-driven insights across your entire organization.