Predictive analytics is increasingly used by companies as diverse as finance, marketing and healthcare. It is credited with helping to reduce fraud, identifying the right offers to keep good customers and to predict the likelihood of hospital readmissions. For traditional industries in manufacturing, predictive analytics is synonymous with forecasting: sales and inventory. However, investing in such tools has been possible only for large companies, while many smaller and mid-sized manufacturers have to devise makeshift tools using spreadsheets and often suffer from lack of specialist resources. The US Midwest, considered the home of manufacturing has thousands of supplier companies and smaller manufacturers who can greatly benefit from these tools but cannot afford them.
John is the Operations Manager of one such small company. They make several types of anti-corrosion products for use in the pipeline industry and a challenge for him is to optimize the use of his limited production capacity to meet fluctuating demand. John believes that product planning and forecasting using predictive analytics to help answer questions like “Do we have enough production capacity for supplying our 4 main categories of product?” is critical. According to him, while they generate good quality sales and production data, they lack the resources to leverage this data to help improve operations. What would help are custom analytics apps to help them address questions such as the one above.
What should be the outcome of such a process? While the process, in principle, is essentially simple: collect sales data to build a predictive (forecast) model, the critical step is to leverage these results properly so that experienced managers like John can utilize the trend and forecast results to appropriately build-to-stock and make other production adjustments. For example, we show below a custom app that gets fed forecast results from the predictive models on a regular basis so that someone like John can monitor and understand which products need to built with urgency and which products can be managed with existing stock.
Looking at the dashboard with a little more scrutiny will reveal that for example, based on a slight seasonal dip for product 1 and a slight seasonal bump for product 2 in October, one can effectively swap resources between these two products early in August or September. So you start to stockpile product 2 with some extra shifts while slowly ramping down on product 1 towards the end of September. This would allow John to effectively balance his limited staff and machinery.
This type of insight is only the proverbial tip of the iceberg when it comes to value from analytics and its effective deployment through dashboards.
Most small manufacturing companies already have decent infrastructure for data collection. All it takes is a simple accounting system connected to an Access database, for example.
The National Center for Manufacturing Sciences (NCMS) estimates that there are nearly 300,000 small and medium sized manufacturing companies in the US who lack access to affordable computational tools and apps for their day to day operations. NCMS calls these companies the “missing middle”, manufacturers who typically employ fewer than 500 people, but are nonetheless responsible for more than twice the global employment of larger organizations.
The basic problem is the same across all small manufacturers: utilize their unique data to deliver operational and marketing advantages. While larger companies can purchase general purpose analytics software and hire experts to develop solutions, SMEs clearly lack this ability. But there are two emerging trends in business analytics that can help SMEs: open source software and affordable dashboard tools. The missing piece of the puzzle is the expertise to combine these disparate technologies affordably. Combining process knowledge in these traditional industries with advanced analytics to develop a cost effective solution for businesses is a critical need, but one that has reliable solutions.
One final, but very important questions pertains to the return on investment in analytics. This has obviously very great signficance and we will reserve it for another article!
How can you affordably implement custom analytics in your business?