Last week, Supply Chain Matters shared our initial impressions of PTC’s LiveWorx 2015 conference, which had a sole focus on the emergence of Internet of Things technology. For us, one of the more interesting notions of PTC’s IoT strategic portfolio was the announced acquisition of ColdLight. According to PTC’s briefing with press and analysts, the ColdLight Neuron platform will serve as a platform for automated predictive analytics applied to IoT focused data, and PTC was willing to invest in excess of $100 million to secure this capability.

We had the opportunity to sit in on a session conducted by Ryan Kaplan, the CEO of ColdLight, where he overviewed the functionality elements of the Neuron platform. The technology provides various pillars of support in analyzing rather large data streams which include aspects of why did an event occur, and what will happen next based on various automated simulations. We and others often refer to these capabilities as ranges of descriptive and prescriptive analytics. Rather than software following programmed instructions, this type of technology capability is termed machine learning technology where algorithms actually “learn” from data and information patterns to make predictions based on such patterns. Once more, the data is analyzed in a much faster manner, sometimes as minutes or seconds.

The ability in applying such capabilities to various future IoT applications has important potential in service lifecycle management (SLM), manufacturing and other product related support needs. Once more, Kaplan pointed out that many of these capabilities can be applied without the need for rather expensive data scientists, which hinders scalability and wider scale deployment.

The aspects of machine learning being applied to support supply chain management focused business processes is not new. More than a year ago, Supply Chain Matters called attention to supply chain planning technology provider ToolsGroup application of machine learning technology to product demand forecasting. In the ToolsGroup SO99 planning application, the technology is providing support to difficult demand planning scenarios involving concentrated incidents of trade promotion, high frequencies of new product introduction, product seasonality or cannibalization. This form of product demand environment is becoming much more prevalent for supply chains supporting today’s online fulfillment and Omni-channel commerce environments.

The application of machine learning is in the analysis of the many relevant variables and interactions related to product demand, and by capturing and modeling all the relevant attributes that shape demand, while filtering out the “noise” or random demand fluctuations. For a more detailed perspective, ToolsGroup’s Jeff Bodenstab penned a recent blog describing how the Rulex machine learning technology is applied at Danone. CEO Joe Shamir penned the published article, Machine learning: A new tool for better forecasting, in the Quarter 4, 2014 edition of Supply Chain Quarterly (CSCMP Membership required to access) . Besides providing a detailed tutorial on the aspects of machine learning that can be applied to supply chain demand planning and forecasting, CEO Shamir observes that companies can now take advantage of valuable data signals that are being generated closer to the consumer, such as point-of-sale or social media channel indicators. That alone is important competitive differentiator in being able to respond quicker and more intelligently to various market signals.

The important takeaway for readers is that proven predictive analytics techniques supported by machine learning technology is no longer just a vision, and will be even more available for targeted use by supply chain and product management teams in the coming months and years. They will pave the way for more innovative processes and more proactive decision-making in many supply chain, manufacturing and product management areas. The added benefit is that many of these tools have been designed to be utilized by current users who have broad business and data knowledge without the need to staff data scientists. Targeted application such as that of ToolsGroup addresses specific business planning and process needs.

Finally, the application of predictive analytics based solutions are once again being brought to market by smaller, best-of-breed innovators who often tend to position themselves on the first mover advantage of market needs, while understanding the realities of having to augment existing systems. That is a win-win situation.

Bob Ferrari

Disclosure: ToolsGroup is a current client of the Ferrari Consulting and Research Group LLC.