Supply chain management commonly makes use of regression analysis to analyze relationships between variables and make predictions about future outcomes. Exemplary application domains are control of production levels, inventory levels, transportation costs, delivery times, customer demand, scrap reduction and customer satisfaction. In today’s blog post I write about regression analysis, its application in SCM, and its integration into major ERP systems for SCM. I will focus on demand forecasting and inventory control. Nevertheless, regression analysis e.g. also supports transportation cost analysis and decision making, as well as e.g. identification of main drivers of customer satisfaction.

Regression models for demand and inventory control

Several types of regression models are commonly used for demand forecasting and inventory control. Every model type is strong in some areas, and weak in others. Organizations carefully consider which model represents the best fit to their needs before conducting demand forecasting or inventory control analysis.

Linear regression for demand and inventory control

Linera regression models assume the relationship between the demand for a product and one or more predictor variables to be linear. Linear regression model applications can e.g. find the main drivers to demand, and thereby aid prioritization, decision-making and planning. For exampel you can try to fit a linear regression model to data comparing prices and demand for a product. How much do promotions, i.e. price reductions, increase demand?

The underlying assumption is that a reduction in price increases demand. Linear regression, as stated, assumes this relationship to be linear. SCM departments widely and commonly use linear regression. If it is clear that a relationship is not linear but, e.g. polynomial, non-linear regression models such as polynomial regression models can be applied instead.

Time series regression for demand and inventory management

Time series regression models are used when the data includes trends or patterns that change over time. Examples are seasonal fluctuations or ongoing long-term growth. In these models, the historical demand data is used to identify patterns, which are then used to predict future demand. Time series regression will help estimate required inventory levels during summer, during holidays, and throughout the year.

Logistic regression for demand forecasting and inventory analysis

Logistic regression is a type of model that is commonly used for inventory control when the data includes binary outcomes, such as whether a product is in stock or out of stock. This type of model can be used to predict the likelihood that a product will be out of stock based on factors such as demand, lead times, and supplier performance. Logistic regression can help in the search for main drivers to stock-outs and low availability.

Poisson regression for demand and inventory control

Poisson regression models are a good option when the data includes count data, such as the number of sales of a particular product. These models can be used to predict demand based on factors such as price, promotion, and marketing activity.

Examples of commercial regression analysis

I want to provide some commercial application examples of regression analysis.

Amazon uses predictive analytics and regression analysis to forecast demand and optimize inventory levels in its warehouses. This allows the company to reduce the amount of inventory it holds while still ensuring that products are available for customers when they need them.

Walmart uses regression analysis to analyze its supply chain data and optimize its logistics operations. For example, the company uses regression analysis to model the relationship between transportation costs and delivery times. This allows Walmart to make informed decisions about how to balance these factors as part of its logistics planning process.

Procter & Gamble uses regression analysis to forecast demand and optimize its production planning. By analyzing historical sales data and other relevant factors, the company can accurately predict demand for its products and adjust its production levels accordingly.

Ford uses regression analysis and other statistical techniques to optimize its supply chain operations and reduce costs. For example, the company uses regression analysis to model the relationship between supplier lead times and inventory levels. This aids decision making with regards to when to order parts, and how much inventory to hold.

Which major ERPs have regression analysis integrated?

Most major Enterprise Resource Planning (ERP) systems do not have regression analysis integrated directly into their software. ERP systems are typically designed to handle a wide range of business functions, processes and operations. ERP systems usually support business functions such as accounting, inventory management, and supply chain management. But, they generally do not include statistical modeling capabilities. Nevertheless ERP systems usually do have built-in reporting and analysis tools that can be used to extract data for use in regression analysis.

For example, many ERP systems have reporting modules that allow users to generate customized reports. These reports can then be exported to a statistical analysis software package such as R, SAS , SPSS etc. for e.g. regression analysis.

Additionally, some ERP vendors offer add-on modules or integrations with third-party analytics software that includes regression analysis capabilities. For example, SAP offers a predictive analytics module that includes regression analysis as one of its capabilities. Oracle offers a variety of analytics products, including Oracle Business Intelligence, which can be integrated with ERP systems and includes regression analysis capabilities.

I willl focus on SAP, as I am originally from Germany and SAP in widely used at German manufacturing companies.

SAP add-ons for regression analysis

There are add-ons available for SAP that allow users to perform regression analysis directly within the SAP system. One example is the SAP Predictive Analytics software, which includes a range of predictive modeling techniques, including regression analysis.

With SAP Predictive Analytics, users can build regression models using data stored within the SAP system, and then use these models to make predictions and forecasts about future outcomes. The software includes e.g. linear regression, logistic regression, and time series regression. In addition to regression analysis, SAP Predictive Analytics includes other analytics capabilities. For example: Clustering, classification, and data mining. It also includes built-in data preparation and visualization tools, making it easier for users to work with large and complex data-sets.

SAP Predictive Analytics, however, requires specialized skills and knowledge. Companies deploying SAP Predictive Analytics for the first time will have to build these skills, and will have to deploy concepts and structures that ensure that this knowledge and these skills stay and improve in the company over time.

Regression analysis add-ons for other major ERPs

There are other add-ons and extensions available for major ERP systems that provide regression analysis capabilities. Microsoft offers several add-ons for Dynamics 365 that provide advanced analytics and predictive modeling capabilities, including regression analysis. The Microsoft Power BI suite includes built-in regression analysis tools, as well as other predictive analytics and data visualization features. Oracle offers a range of analytics products, including Oracle Business Intelligence Cloud, which provides advanced analytics and data visualization tools that can be integrated with Oracle ERP Cloud. The software includes regression analysis capabilities, as well as other predictive modeling techniques.

Infor offers a range of analytics tools that can be integrated with its ERP software, including Birst, a cloud-based analytics platform that includes regression analysis capabilities. Birst allows users to build predictive models using data from multiple sources, including ERP data, and to visualize the results in real-time dashboards.

Epicor offers several analytics add-ons that provide predictive modeling capabilities, including regression analysis. The Epicor Data Analytics tool includes built-in regression analysis and other statistical modeling features, as well as data visualization tools.

There many add-ons and extensions available for major ERP systems that provide regression analysis capabilities, as well as other advanced analytics and predictive modeling features. The specific tools and software packages available will depend on the ERP system in question, as well as the specific needs and requirements of the organization using the software.

Linnart Felkl

Data scientist focusing on simulation, optimization and modeling in R, SQL, VBA and Python