Copyright by Dr. Edmund Prater, University of Texas at Arlington. Sub-segment 1.6 Introduction to Forecasting

Transcription

1 Sub-segment 1.6 Introduction to Forecasting You will be going into the details of forecasting in another session. The goal of this segment, however, is to lay the groundwork by providing basic definitions and structuring the general forecasting methods. We will go into issues that SCM managers need to be aware of, since to survive in today s competitive markets, mastery of the forecasting process is a key competence required of a company. This work introduces the five viewpoints a forecaster can have of the future. It describes particularities in forecasting for a company and at an international level. In addition, it shows benchmarks in the forecasting process, the models used, the applied software solutions and systems. Finally, a best practice solution is presented and emergent phenomena introduced that might show the trend for forecasting in the future. Forecasting is a critical element in virtually every significant SCM decision. Accurate forecasts allow managers to choose the right combination of material and human resources to produce the goods and services needed by the company. Forecasts are the basis for long-range planning. This planning determines the number of production facilities, warehouses and distribution centers a supply chain manager has to work with. Marketing departments use forecasts to determine how many sales personnel are needed in individual territories and when to introduce new products. Operations and manufacturing use forecasting for capacity planning, facility layout, scheduling and inventory control. Forecasts are also used by the finance and accounting departments for budgeting purposes and HR uses the same data to drive hiring decisions. Thus, with poor forecasts you might not order enough raw materials or sub-components. This leads to a shortage of product, which leads to unmet demand, lost sales and reduced market share. On the other hand, you might order excessive amounts of materials, thus ensuring large inventories, increasing costs and reducing profits. In either case, your competitors are sure to capitalize on your poor decision making. In a later section on uncertainty, we will go into greater detail as to how poor forecasting can drive amplification or bullwhip effects. Cause low utilization of machinery and increase inventories. Suffice it to say that the basic effects you saw in the

2 introductory Beer Game are but the tip of the iceberg when it comes to the real-world impact of forecasting. Forecasting Methods There are four basic methodologies that Chopra and Meindl at Northwestern have defined. These are: Qualitative: this relies on the forecasting personal insight or intuition. Causal: this assumes that specific variables drive the forecast. For example, interest rate increases by the Fed may decrease commercial loan activity. Time Series: assumes that historical data will predict future trends. Simulation: combines causal and time series methods in order to try out what-if scenarios. Most companies use a combination of these techniques to forecast. You will go address these topics in detail in a later course. Key Concepts in Forecasting Forecasting is more accurate for shorter time frames. For example, when Wal-Mart had their store begin ordering every two weeks rather than each month, inventories were reduced because forecasting accuracy increased. All that was necessary was to reduce the forecast time frame by a few weeks. Different forecasting tools are applicable for different applications and time frames. The table below provides a basic comparison of the types of forecasts routinely done by organizations.

3 Application Short Term Medium Term Long Term (0-3 months) (3 months-2 years) (more than 2 years) Forecast Quantity Individual products or Total sales groups or Total Sales services families of products or services Decision Area Inventory management, Staff planning, production Facility location, capacity final assembly scheduling, planning, master production planning, process work-force scheduling, scheduling, purchasing management master production distribution scheduling Forecasting Technique Time series, causal Causal judgment Causal judgment judgment Another key issue is that regardless of the time frame and forecasting method that is used, there are five key concepts that must always be taken into account. These are: Impact of Technology: if in the 1970 s you were the CEO of Smith Corona typewriters and had not considered the impact of PC s on your business, you would have been blown out of the water (which is exactly what happened). Social Issues: microwave ovens were available in the 1950s but the market was not there since most women were not working outside the home. Political Issues: when governments offered subsidies for development of new products, this can cause accelerated product introduction. Hybrid vehicles in California are one such example. Legal Issues: the federal breakup of Bell Telephone definitely impacted the rate at which new products and services were offered by the phone company (companies) Environmental Issues. The demand for cleaner air can drive the development of new technologies and services. On the other hand, wide-open spaces such as Dallas, Texas do not pressure the population to make use of services such as mass-transit like citizens in Boston, MA with its higher population density limited by the harbor.

4 Given the importance of forecasting, it is often perplexing to see that companies have no centralized structure in place to ensure data accuracy. For without accurate data, no forecast can hope to be accurate. Often, different divisions within the same company will use different methods, software and databases to forecast. Meanwhile, overall company forecasts are difficult because the various software packages and databases are not compatible. This is a classic example of what can happen when companies focus on business functions rather than business processes. However, breakthroughs in technology make it possible for today s companies to gather more relevant information. Forecasts are produced on a corporate, as well as departmental level. Each unit of a corporation has to make predictions of its sales and has to approach the necessary resources to reach this goal. To optimize the process of forecasting, business units have to collaborate with other units inside and partners outside the company. The capacity of a company to integrate its suppliers and customers into its production and forecasting can provide it a competitive advantage towards its competitors. Programs like VMI (Vendor Managed Inventory) and CPFR (Collaborative Planning, Forecasting and Replenishment) are examples of collaboration to optimize the forecasting processes in supply chains. In the past, companies have mainly forecasted intuitively. Experience in the form of intuition has been the basis of forecasts. In recent years, however, companies favor scientific approaches that are based on hard facts and data. Each department has to forecast its business for the upcoming period to optimize the overall performance of the company. Technology enables the storage of huge amounts of data in data warehouses. Through data mining, a company can extract forecast-relevant data. Forecasting software and processes apply models that can improve precision of forecasts enormously.

5 Five Viewpoints of the Future The viewpoints of supply chain managers determine the selection of forecasting information. Before a forecast is created one must determine how those managers view the future. It is imperative to back-up forecasts with both logical and credible data, but credibility varies depending on the audience and situation. As a consequence, some thought should be given in determining what angle to seek when choosing forecasting methods, based on decision maker s viewpoints of the future. There are five basic categories in which these viewpoints can be classified to help determine adequate forecasting methods. They are: extrapolators, pattern analysts, goal analysts, counterpunchers and intuitors. The key is to realize that neither viewpoint alone is perfect for every situation, but the right combination of these viewpoints and their techniques will produce the most valid information for supply chain managers. Extrapolators Technology Trend Analysis Fisher-Pry Analysis Gompertz Growth Limit Analysis Learning Curve Pattern Analysts Analogy Analysis Precursor Trend Analysis Morphological Matrices Feedback Models FIVE VIEWPOINTS OF THE FUTURE Goal Analysts Impact Analysts Content Analysis Stakeholder Analysis Patent Analysis Roadmaps Counter Punchers Scanning, Monitoring, Tracking Alternate Scenarios Cross Impact Analysis Intuitors Delphi Surveys Nominal Group Conferencing Structured &Unstructured Interviews Competitor Analysis Quantitative Qualitative Forecasts

6 Extrapolators Extrapolators fall into the quantitative category when it comes to the type of data that they value in coming up with forecasts. Their basic belief is that the future is reasonably predictable, based on past trends. One weakness of this belief is that it fails to take dramatic changes into consideration that could occur in the present but are not considered in the past data. Engineers usually follow this viewpoint on the future. Common techniques and methods used are: Trend Extrapolation Fisher-Pry Substitution Analysis Gompertz Substitution Analysis Growth Limit Analysis Learning Curves Pattern Analysts Pattern analysts also value quantitative information and are very similar to extrapolators. They basically believe that history repeats itself and that one may forecast the future by identifying and analyzing situations from the past and applying the cycles to future circumstances. A good example of a pattern analyst is a pure scientist. Common techniques and methods used are: Analog Analysis Precursor Trend Analysis Morphological Analysis Feedback Models Goal Analysts Goal analysts rely less on quantitative data and more on the belief that the future is determined by the actions and beliefs of a collection of individuals, organizations and institutions. Their basis of forecasting is determining the goals of trendsetters and supply chain mangers, and figuring out how large of an impact they can have on future trends. Individuals in marketing are usually goal analysts. This thinking takes on a more real-

7 world approach but also does not emphasize the importance of other forces that impact change. Forecasting methods and techniques commonly used are: Impact Analysis Content Analysis Stakeholder Analysis Patent Analysis Counter-Punchers Counter-punchers are on the qualitative side of the spectrum when it comes to the data that they use in forecasting. Their belief is that the future is a result of random events and the best way to stay in tune is to follow trends and plan accordingly. There is a high degree of judgment used in this type of forecast. Typical counter-punchers are usually executives. Common techniques and methods used are: Scanning, Monitoring and Tracking Alternate Scenarios Cross-Impact Analysis Intuitors Finally, intuitors believe that a mixture of driving forces, random events, and the actions of key individuals and institutions shapes the future. The world is too complex to use a rational technique to project the future, so one should gather all the data available and use personal intuition to make forecasts. Intuitors are usually executives as well. Their common techniques and methods are: Delphi Surveys Nominal Group Analysis Structured and Unstructured Interviews Levels of Forecasting Forecasting is one important part of company achievement. One must know the number of products the customers want. However, world trading is changing to a global

8 supply and demand chain. Knowing local customers and doing organizational forecasting is not enough. Companies must carry on international forecasting as well. Organizational Forecasting The main factors affecting accuracy on this level are: Sales and marketing involvement Software design procedures Forecasting team structure Data input Parameter setting Sales and operation planning Top down approach This approach is common and simple because companies forecast on an aggregate level by using basic inputs such as historic data and promotion effects. It is suitable for predictable sales products. However, management sometimes biases forecasting results because they are trying to meet sales targets. Bottom up approach This approach is more complex than the first one. There is a larger number of forecasters involved. Sales representatives or local officers usually forecast at a local level. The approach is proper for specific promotional products and effective in planning for warehousing, manufacturing and transportation. The disadvantage is that lower level historical data is usually lost or not completed. International Level All factors mentioned in the organizational level still have an affect on forecasting results, but cultural differences among regions will be another important factor. For example, Latin American cultures prefer bright and vibrant colors while Asians do not. Also, Americans prefer sweeter foods than Europeans do. Therefore, when companies engage in international forecasting, culture cannot be overlooked.

9 The Forecasting Process Among the forecasting functions, people, process, technology and resources, the process is the most important factor. It connects people to technology and resources. Because all departments of a company have to be involved in the forecasting process it is essential to coordinate and synchronize actions among them. The lead-time determines how far ahead a forecast has to be conducted. The shorter the lead-time, the higher the probability of setting accurate forecasts. To take full advantage a company has to synchronize forecast-period and lead-time. To make the forecast as precise as possible, forecasted and actual results have to be monitored and compared carefully. In an ever-changing business environment, companies have to make sure that their forecasting process still meets market realities. Best Practices in Forecasting Forecasting is essential in any fast moving sector where demand or product mix changes. Experience over the last few decades of commercial forecasting has identified good practices, which improve the likelihood of providing effective forecasting. A good forecast should be timely, accurate, reliable, made in meaningful units for the users, simple to understand, and documented in a way that enables later review. For a good forecast Focus on short range forecasts as these are the drivers of immediate sales and replenishment activities. Forecast at a product family level wherever possible as this will be more accurate than the individual SKU level. The forecaster should understand the business well to review and finalize the forecast. The original forecasts should be retained and regularly measured for errors using various techniques.

10 A good forecasting software should be used that will provide an easy to use and visual interface as well as good forecasting tools. Minicase: Why Supply Chain Forecasting is a best practice at HP: At HP merged with Compaq, managers needed simple and consistent supply chain models for the combined businesses. They knew they had to identify the best practices in each organization and apply them to the merged operations. The management used a wide range of communication techniques, supported by computing infrastructure, in order to find large savings for the combined operations. The parties agreed on common business tools and published their choice so that it was available for other members of the organization. At every node of the supply chain, a consensus forecast confronts net demand and supply. As a consequence, a detailed communication flow was provided for all members of the supply chain. The forecast of demand is updated on a daily basis. If a plan changes, the reason has to be explained to the supply chain partners. HP uses SAP s Enterprise Resource Planning software and uses the Internet to distribute a daily action plan, based on the demand forecast to all the impacted members of the supply chain. HP s Ink Jet division cut its forecasting time for the supply cycle by three days. The Enterprise Server Group business unit has shortened the change order process from three weeks to just 24 hours. Through all this measures, HP continuously increased the efficiency of its operations and enhanced its customer satisfaction. Summary In a dynamic and global market place, rules of the game change rapidly due to factors such as uncertainty in demand, changes in technology, and changes in competition. Business forecasters attempt to control the demand uncertainties by using various forecasting techniques and also by using a fair amount of insights, judgment and intuition. As this suggests, business forecasting may involve both qualitative and quantitative data.

11 While forecasting, supply chain managers subjectively assess the alternatives based on quantitative data. The subjectivity of supply chain managers can be summed up in five viewpoints: extrapolators, pattern analysts, goal analysts, counter-punchers, and intuitors. The forecasting function involves people, process, technology and resources. Forecasting may be done at the level of individual, organizational and international. At individual level, customer tastes and preferences are considered. The organizational level will include the aggregated data on consumers, advertising effects and economic trends. Analysis at the international level may focus at cultural differences between countries in terms of powerdistance, uncertainty avoidance, individualism / collectivism, and masculinity / femininity. What forecasting approach works well may depend on individual companies and their strategic goals. Any business problem or opportunity requires valid forecasts, which should start with defining the objectives of the forecast. Various techniques may be selected from different views since the validity of the forecast will depend on the breadth of techniques used. Essentially, the forecast that is best able to capture the future will be the one that will be the most successful. References Brichi, Audrey Goins, "The Inclusive Supply Chain- A Competitive Advantage: Strategies for Minority and Women Business Sourcing," Institute for Supply Management 8gth Annual International Supply Management Conference, April 2004 Hamel, Gary, "Revolution vs. Evolution: You Need Both," Harvard Business Review, May 2001, pp Hannon, David, "Radar View! Purchasing Works with Logistics to Improve Cost, Quality And Supply," Purchasing, July 15, 2004, pp. 44 Fisher, Marshall L, "What is the Right Supply Chain for Your Product," Harvard Business Review, March-April 1997, pp Morgan, James P, "Supply Chain Management Faces Tough Scrutiny," Purchasing, August 17, 2004, pp Shmidt, Michael and Steven Aschkenase, "The Building Blocks of Service," Supply Chain Management Review, July/August 2004, pp

12 SGV & Co, "Supply Chain Management," SGV Bulletin, March-April 2001, Issue No.2 Wetlaufer, Suzy, "The Business Case against Revolution: An Interview with Nestle's Peter Brabeck," Harvard Business Review, February 2001, pp Zwolinski, Kevin, "Changing Supplies: Altering an International Supply Chain Can Involve Significant Organizational Change," Manufacturing Engineer, February 2003, pp