How to make Preventive Maintenance a way of life in SAP APO Support Projects - Part 1
After working in software industry and SAP APO, when I look back, I feel compelled to compare the way we execute our support projects with the way we executed preventive maintenance in typical manufacturing industry. Just like any manufacturing industry, role of preventive maintenance cannot be underestimated even in software industry.
Speaking specifically of SAP APO, maintaining APO solution free of any master / transactional data issues plays very important role in ensuring that planner (who is the end user of the planning system) always gets a defect free system to begin with. Many a times, the lack of data accuracy in APO is the reason that planners tend to do manual planning using spreadsheet. They deploy complicated formulae in their excel sheets not realizing that even more advanced logic is available in APO. Ensuring data accuracy in the planning system is like maintaining basic hygiene. This will not only increase planner's belief in supply and demand plans created by APO, but it will also give him / her more time to explore advanced features in APO that will in turn help maximizing value out of solution they have deployed.
In a typical support project, we come across varieties of issues. Many of them are repetitive in nature and can be avoided by careful study of the logs even before the issues are encountered / logged by end users.
• Batch jobs / process chains are deployed in any steady state APO project for performing tasks such as loading history, calculating forecast, releasing forecast from DP to SNP, SNP engine run, etc. Each of these batch jobs generate logs / spools that warn us about the logical / technical errors and potential issues that will be faced by planners.
• History load from APO DP cubes to planning area happens through standard transaction / sapapo/rtsinput_cube. If the CVC is absent for any reason, and then historical data (such as past orders / shipments) cannot be loaded from cube to planning area and hence will appear as an error. Careful examination of the product codes can highlight if the codes are obsolete (and hence no longer required in APO) or if they are new and will need to be setup as a CVC in APO. This can then be rechecked with the Demand Planners to ensure correct CVCs are created in time to prevent any data loss.
• Forecast creation process typically uses one or the other statistical forecasting models to create uni-variate forecast. Alerts can be created to calculate the measure of error and highlight in case of exceptional errors.
• Forecast release from DP to SNP happens at a product location level. If a particular product location is not setup in APO SNP, then the corresponding forecast will not get loaded from DP to SNP and hence a portion of the total forecast for that product may get lost thus resulting in in-sufficient planning.
In the next part, we will discuss on errors that occur in master data area and also focus on important checks that can be put in place to ensure preventive maintenance.