Leveraging HANA for enhanced Labor Demand Planning in Warehouse Management
The Labor Forecasting and Planning module of a Labor Management system is a tactical tool for managers to determine how many resources are required to complete the planned workload. Distribution centers tend to be a dynamic demand driven environment especially with the advent of multichannel fulfillment (shipping to stores, retailers, web customers and managing returns) all out of one DC. The shipping and processing requirements for a day can vary in volume and complexity and can change during the shift (think of ecom orders!). That creates the need for an analytical tool that can look at a specific profile of work (demand) over a shift, day or week and determine what are the right number and type of employees to complete the tasks. The system takes the planned workload and calculates the required labor using Engineered Labor Standards that are pre-determined for each job function. In a standard model the labor estimate is run prior to the start of a shift and does not change or fluctuate dynamically based on what’s happening within the shift. Accuracy in the estimate is impacted by outdated standards, lack of statistical controls for the changing nature of the work force’s productivity and other factors not directly or easily captured in a standard. How can we improve the labor planning forecast?

SAP HANA and Predictive Analytics
SAP HANA in conjunction with Predictive Analytics facilitates a new approach to labor forecasting based on trend analysis. HANA is an in-memory database technology that can provide instantaneous results from transactional data. The system can look at performance over a given time period for a given task and utilize predictive analytics to forecast how well the task will be executed over the next shift or time period. Predictive analytics is smart software that learns based on actual outcomes and then provides a prediction (in this case) of a future event like labor requirements. Your data ‘trains’ the software and the results are not aggregate estimates like a forecast but are individualized by the task. HANA provides the horsepower to crank out this analysis in real time. This enhanced look at labor will help you manage better and stay on top of your engineered standards. You can run the regular labor forecast based on engineering standards and then run the predictive analytics estimate. The two can be compared by functional area which will help diagnose what’s causing a deviation from the standard. Once the source of the deviation is understood the issue can be addressed and the standard can be modified.

Intra-shift analysis
An intriguing feature of predictive analytics with labor planning is the ability to dynamically analyze data as it is happening. For example you can be three hours into the shift and run the updated trend analysis to predict how the rest of the shift will fare against the existing workload. This is a huge help in adjusting for changes in workload and the real pace of throughput for that day. New orders may have dropped to the DC, there can be delays caused by MHE downtime or system performance or weather. The updated results will give operations the chance to re-direct resources to problem areas or advise managers on the possibility of overtime. Concurrently, you can continuously run analysis against trailer moves, picks, packing, sortation volumes, conveyor speeds (the list goes on) and have the system provide up to the minute projections on task rate of completion and other pertinent metrics which are indicators of shift performance and potential issues with flow and throughput. This ‘electronic supervisor’ is constantly monitoring and analyzing and re-shooting the forecast. Such capability adds a new layer of accuracy and credibility to the operational dashboard and will drive labor cost savings.

SAP EWM and Predictive Analytics
The SAP extended warehouse management solution has built in Predictive Analytics capabilities (as of version 9.1) and connectors to HANA. This makes for a simpler implementation as other WMS systems will have to use an external Analytics application. A further advantage of EWM is the MFS (Material Flow System) architecture within EWM. This enables you to connect any automation directly to EWM and feed that machine and equipment data real time for use with analytics. SAP’s Business Objects can then seamlessly provide the reporting and data presentation.

The advantages of using HANA with predictive analytics in labor planning are: 1. Allows managers to forecast the labor based on planned workload using engineered standards and use trend analysis as a parallel model to substantiate and sharpen the numbers. 2. Dynamic intra-shift analysis can be executed to provide a point in time update on how well the operation is performing and give managers a pro-active heads-up on bottlenecks and the possibility of overtime. This is especially helpful in Ecom environments where orders drop to the DC throughout the day. 3. Operational dashboards will become more dynamic, pro-active instruments for managing the warehouse I think Predictive Analytics in WMS represents the next significant wave of innovation. The next step in this evolution will be to use the updated predictions to automatically re-direct and re-prioritize the work within the WMS and even create replenishments and other tasks. I will cover this in a subsequent blog.