Modeling & Metrics in SAP APO Workshop

Presented by Mark Chockalingam, Ph. D., Managing Principal

March 14, 2012, 8:30 - 16:30 
at the Crowne Plaza, Manama, Bahrain

APO DP workshop only - $995 per participant Early-Bird price $895 until January 20, 2012

Highly Recommended as an add-on workshop with the Demand Planning two day workshop - $1,695 Early-Bird price for all three days. Early-Bird expires January 20, 2012

view and print seminar brochure (PDF)

Register Now!

Overview

SAP APO is a popular supply chain planning tool among many Fortune 1000 companies today. APO can help you create better demand forecasts and finished goods plans, when properly implemented together with appropriate training for the planning professionals.

The main focus of the workshop is Statistical modeling and forecasting in SAP APO. We will cover the various modeling strategies including the automatic model selection procedures. We will spend quality time on explaining data analysis and graphical review in APO.

Please bring your laptops with access to your APO DP environment so you can test a couple of models on the fly.

This is a chance to discuss your challenges in using SAP APO from technological and analytical point view.

The tool also uses a variety of Error measurements as model diagnostics to assess the quality of the forecasting model and enables exception management through reporting and alerts.  In this workshop, you will also learn the mechanics behind the Forecast Error metrics available in the System.  

SAP APO has six different error metrics. There is a lot of confusion and which ones to use and how to use them and what they really mean.

1. How many of them are you currently using in your modeling?
2. How many of them you should really be using?
3. Which metric is the best indicator of forecast quality?

APO DP defines the error metrics with its own unique formula that is different from conventional calculations. So it is critical for planners to know how they are being calculated and how to use them to diagnose forecast quality. We illustrate with examples the calculations of MAPE, RMSE and MPE and the pros and cons of using each.

Forecast Alerts are a big part of Demand Modeling by Exception. We show you how to leverage the error measures to define and use Univariate Forecast Alerts. The final tool to compare among alternative forecast models is to use the Forecast Comparison Report.

On completion of the course, you will have a better understanding of the different modeling strategies and you can take advantage of the exception management practices built into APO DP to model, forecast and manage the process by exception.  Learn how to leverage:

• Forecasting exception alerts and
• Forecast Model comparisons.

Who Should Attend?

• Demand Planners
• Forecast Analysts & Managers
• Analysts in the Supply Chain
• Director of IT
• Inventory and Operations Planners
• Product Managers
• IT Business Analysts
• IT Managers
• APO support professionals
• Director of ERP applications
• Director of Supply chain applications
• IT Business Managers

Topics

8:15am - 10:15am: Statistical Modeling in APO DP
• Developing time series models in APO
• Forecast Profiles
• Completing data analysis in APO and graphical review of data
• Testing of alternate Models – Moving Averages, Smoothing, trend and seasonal models and mimicking the history approaches to develop forecasts.

10:15am coffee break

10:30 - 12:15pm Statistical Modeling continued
• Holt Winters Models
• Automatic Model Selection 1 - Strategy 50
• Automatic Model Selection 2 - Strategy 56
• Croston's Model - when to use and when not to use
• Forecast adjustments using APO DP.
• Impact of Outlier adjustments on Model
• Monitoring model performance and re-specification for improved results

12:15pm Lunch

1:15pm - 3pm Model Diagnostics
• Testing and measuring forecasting models in APO
• Adjustments and Exceptions Management
• Model Diagnostics through Forecast Error
• How to identify a biased Forecast?
• Which metric is the best indicator of Forecast quality?

Forecast Errors
• Definition of Demand Forecast Errors
• Errors over time vs. Errors Across Products
• Demand Forecast Metrics Computations in APO
• Mechanics behind MAD, ET, MPE, MSE, MAPE and RMSE

3pm coffee break

3:15 to 4:45 Additional Considerations
• What is Tracking Signal?
• Comparison of alternative Forecast Models
• Forecast Modeling through Exception
• Univariate Forecast Alerts
• Forecast comparison Report
• Summary and Conclusion

Register Now!

An additional 10% discount applies to group registrations of three or more using one payment method.