Correlate the Use of Causal Analysis in SAP APO with the Underlying Mathematical Model

  • by Alok Jaiswal, Consultant, Infosys Limited
  • February 23, 2016
Learn how to carry out causal analysis in SAP Advanced Planning and Optimization (SAP APO) when demand is correlated with some known and measureable factor (demand is a function of some variables). Understand how regression is used to find correlations between a single dependent variable (y) and one or more independent variables (x1, x2…xn) and how it is carried out.
Learning Objectives

By reading this article, you will learn about:

  • Forecasting and causal analysis
  • The objects in Demand Planning needed to use this function
  • Business scenarios for demonstration
  • Regression and mathematical models used in regression
  • Carrying out linear regression in SAP Advanced Planning and Optimization (SAP APO) with one independent variable
  • Carrying out linear regression in Microsoft Excel with one independent variable
  • Carrying out linear regression in SAP APO with two independent variables
  • Carrying out linear regression in Microsoft Excel with two independent variables
  • Interpreting results and evaluating model accuracy
Key Concept

Forecasting is a process involving qualitative and quantitative techniques to predict future demand values. In forecasting techniques, regression analysis is used for prediction and forecasting. It basically helps to identify the relationship between an independent variable such as price on a dependent variable such as demand. Simple linear regression is a function of a single independent variable while multiple linear regression (MLR) is function of more than one independent variable.

Forecasting is part of the Demand Planning (DP) and management process in SAP Advanced Planning and Optimization (SAP APO). The system reads historical data and calculates the corresponding values that it proposes as the future data. Causal models can be very useful when you can determine (and measure) the underlying factors that drive the demand of your product. Causal analysis can be carried out not only on a product but on any characteristic as per business scenario.

You can use causal models in scenarios in which you have independent variables, such as price, that have some influence on a dependent variable, such as sales quantity. With this information, you can set up the system to make sales react in a certain way when you change the price in the future. Multiple linear regression (MLR) is a form of causal analysis model. It enables you to analyze the relationship between a single dependent variable and several independent variables. I focus on how to use causal models in forecasting demand in SAP APO along with understanding the underlying mathematical model.

Overview of Forecasting and Causal Analysis

Forecasting is one of the components of an organization’s DP and management processes. It basically helps to answer questions such as what the expected demand is for a product in the next six to 12 months. Forecasting methodologies can be divided into two modes – subjective and objective. Subjective methodologies are more qualitative and are heavily dependent on judgment and educated guesses/surveys. Objective methodologies, on the other hand, are more quantitative and rely heavily on past data. Within objective methodologies two widely used techniques are time series and causal models.

Time series models are used where one believes by studying patterns in the past data, future demand can be predicted. Components of time series models are level, trend, seasonality, error, and cyclical. Causal models are used where demand is dependent on some known factors. For example, temperature can be one of the major factors that can influence the sales of ice cream or the rate of unemployment can be studied to correlate with number of occurrences of crime in a society.

Figure 1 shows a high-level classification of forecasting techniques and commonly used methods.


Figure 1
Classification of forecasting techniques
 

Note
A naïve forecast is a basic forecast method in which the past period’s actual value is used as the forecast for the next month without any adjustments.

To use causal analysis techniques within SAP APO, you need to complete following steps:

1. Set up SCM master data in SAP APO for product locations.

2. Set up and configure DP objects such as:

  • Storage bucket profile
  • Planning bucket profile
  • Master planning object structure (MPOS)
  • Planning area
  • Characteristic value combination (CVC)
  • Time series objects
  • Planning book and data view
3. Configure causal analysis in SAP APO by:
  • Maintaining a forecast profile
  • Loading time series data in the planning book
  • Executing MLR
  • Interpreting the results

Alok Jaiswal

Alok Jaiswal is a consultant at Infosys Limited.

He has more than six years of experience in IT and ERP consulting and in supply chain management (SCM). He has worked on various SAP Advanced Planning and Optimization (APO) modules such as Demand Planning (DP), Production Planning/Detailed Scheduling (PP/DS), Supply Network Planning (SNP), and Core Interface (CIF) at various stages of the project life cycle.

He is also an APICS-certified CSCP (Certified Supply Chain Planner) consultant, with exposure in functional areas of demand planning, lean management, value stream mapping, and inventory management across manufacturing, healthcare, and textile sectors.


See more by this author


Comments

No comments have been submitted on this article. 


Please log in to post a comment.

To learn more about subscription access to premium content, click here.