People often ask me what exactly can you do with SAP Predictive Analytics. I tell them you can do just about any type of analytic analysis you want with your internal data and any external data sources you can get your hands on. As an example, I just completed a Proof of Concept (POC) using Predictive Analytics at a large food distributor focusing on three improvement areas:
- Inventory Slot Location Optimization: Able to reduce order pick time by 2%, resulting in annual savings of more than $450K using SAP Predictive Analytics (Expert) with the output in SAP Lumira using the Market Basket / Apriori algorithm. The gain was achieved by placing materials that have a high “confidence” of being picked together, near each other on the floor, thus reducing travel time in the warehouse.
- Customer Churn: Identified 4% of customers that are “at risk” of leaving the business. I used R for data cleanup and analysis combined with SAP Lumira for presentation. I used the recency, frequency, and monetary binning technique with five categories. The POC identified additional fields to support the analysis, such as customer, territory, customer type, and profitability percent. In the future, I will use the PAMK algorithm, which identifies the optimum number of categories for analysis.
- Material Rationalization: Found 4% of materials that should be considered for rationalization (obsolete) based on recency, frequency and monetary criteria. The POC identified additional criteria for analysis, such as product life cycle, phase, and profitability percent. A POC recommendation was to combine material segmentation with the analysis to get a more in depth view within the material segments. In the future, I will consider using the Cluster Algorithms within SAP Predictive Analytics instead of the binning technique for Customer Churn and Material Rationalization and continue to use SAP Lumira when appropriate.
The possibilities are endless, but these are three concrete, customer examples of how predictive analytics can reduce costs in your supply chain. To learn more about predictive analytics, view our on demand predictive analytics webinar to see a demo.