Management

1) For this first question, start by going to the following hyperlink and downloading the free book on Forecasting Principles and Practices 2nd ed. by Hyndman. Next, skim through the first chapter and then flip through the remainder of the book at your own pace. Respond in the space below by telling a little about what you learned and what you might be able to learn on your own time as you make a personal commitment to read on at a future date. (Note, you now are in the possession of an excellent book on Forecasting methods and practices using a free statistical package called R.)

2) This question focuses on Chapter 2 in your textbook; What is Demand Driven Forecasting. In the space below describe what are the important concepts that need to be remembered regarding the fundamentals of Demand Driven Forecasting. Think carefully about the following as you form your answer:
• Why forecasting is important to businesses?
• What are some of the traditional forecasting methods employed in business and why they don’t work well?
• What exactly is Supply-Driven Forecasting?
• What exactly is Demand Driven Forecasting?
• Explain the following formula and what does it have to do with forecasting accuracy?
Forecast = Trend t–1 + Seasonality t–1 + Cyclical t–1 + Causal Factor(s)t–1 + Unexplained Variance (p. 81)

3)In Chapter 3 you were introduced to an Overview of Forecasting Methods. You will have learned that there are two broad categories of forecasting. Review both of these methods and then explain:
• How these two forecasting categories differ?
• The principle forecasting methods employed in each category?
• Advantages, disadvantages, and relative effectiveness of each method?

4) In Chapter 4, you were introduced to various methods for assessing forecast accuracy. For this question, begin by downloading the attached Excel File and complete the forecasting accuracy exercise. When completed upload to the completed excel file here.

5) This question references chapter 5 in your textbook: Quantitative forecasting Methods Using Time Series Methods. Briefly explain the following Quantitative Time Series methods and give an example of when each method would be most appropriately applied:
• Moving Averages
• Exponential Smoothing
• Single Exponential Smoothing
• Holt’s Two-Parameter Method

6) Use the attached Excel Spreadsheet template to complete a regression analysis tracking Regular Gas Prices (the dependent variable) as affected by Crude Oil prices (the independent variable). Instructions for the three Tasks are included in the attached file.

 

 

 

 ANSWER

Demand-Driven Forecasting

 

 

 

 

 

 

 

 

 

 

 

• For this first question, start by going to the following hyperlink and downloading the free book on Forecasting Principles and Practices 2nd ed. by Hyndman.Next, skim through the first chapter and then flip through the remainder of the book at your own pace.  Respond in the space below by telling a little about what you learned and what you might be able to learn on your own time as you make a personal commitment to read on at a future date. (Note, you now are in the possession of an excellent book on Forecasting methods and practices using a free statistical package called R.) 

Forecasting is the ability to predict the future as accurately as possible, based on adequate historical data and patterns in the past the present and any knowledge about future events. Forecasters assume that the changing environment will continue to change in a similar way in future (Chase, 2013). Forecasting situations is affected by many factors, such as data patterns types, and other elements that affect the actual outcome. A good forecast must tart by first determining what to forecast and the forecasting horizon which helps in determining the best method to apply. For example, quantitative forecasting method should be used when there is historical numerical data and evidence shows that some aspects of the past patterns are likely to continue in future. On my own time, I will learn how to use R, and practice my skills by developing forecasts for various business cases.

• This question focuses on Chapter 2 in your textbook; what is Demand Driven Forecasting.  In the space below describe the important concepts that need to be remembered regarding the fundamentals of Demand Driven Forecasting.  Think carefully about the following as you form your answer:
• Why forecasting is important to businesses?

Forecasting is very important in business because it mitigates inefficiencies in the supply chain. Demand forecasting affects a lot of aspects on enterprise-wide basis such as, the number of employees that need to be hired, products that should be shipped, raw material required, inventories to produce, plants to build as well as office supplies needed. Demand forecasting allows companies to be ready for future increases in demand rather than waiting to respond. The production process from sourcing of raw materials to shipment is time consuming and expensive hence companies can miss opportunities if they are not ready for surges in demand.

• What are some of the traditional forecasting methods employed in business and why they don’t work well?

Traditionally forecasting experts used gut-feeling judgmental override to predictions developed from simple statistical forecasts which were based on edited sales. These methods were inclined to the company baseline objectives and goals instead of focusing on market trends. Traditional forecasting methods used recent trends and seasonality, and responses to marketing. These methods don’t work well because generating demand based on historic data on past supply responses do not produce true customer demand. Moreover, using shipment history and supply response also don’t produce accurate forecast on demand because supply could be affected by other factors such as employees performance, companies capacity to produce and other supply constraints.

• What exactly is Supply-Driven Forecasting?

Supply-driven forecasting in the process of predicting future customer demand based on historic data on company’s products shipment through analyses of past data on the response of customers to supply.

• What exactly is Demand Driven Forecasting?

Demand-driven forecasting is the predictive processes of sensing, shaping and estimating future customer demands using sophisticated methods and software to analyze available historical sales data.

• Explain the following formula and what does it have to do with forecasting accuracy?

Forecast = Trend t–1 + Seasonality t–1 + Cyclical t–1 + Causal Factor(s) t–1 + Unexplained Variance (p. 81) 

The above formula is used to identify patterns and t to forecast those patterns in future using four elements four elements which are seasonality, trend, randomness and cyclical; Randomness is the unexplained variance. This forecasting method shows that after the prediction has been made, there will be some deviation of the forecast from the actual occurrence. To increase the forecasting accuracy the above formula adds causal factors such as advertising, price, competitor activities, and sales promotion. The unexplained valiance in the formula also helps experts to identify the amount of demand that is not associated with cyclical, trend or seasonality. The randomness can then be isolated, measure and analyzed to identify the extent of variance in the forecast.

• In Chapter 3 you were introduced to an Overview of Forecasting Methods. You will have learned that there are two broad categories of forecasting. Review both of these methods and then explain:
• How these two forecasting categories differ?

There are two broad categories of forecasting methods which are: qualitative methods and quantitative methods. Qualitative methods also known as judgmental or subjective methods depend on subjective assessment of a group of people or a single individual, while the quantitative also referred to as mathematical methods depend on relationship between historical data on sales and other factors. Subjective judgments are developed form a person’s gut-feelings and domain knowledge of people who are familiar with the market and events that are likely to take place in the future. However, the mathematical methods collect, store and analyze sales information using mathematical formulas to come up with the best possible prediction of customer demand in future.

• The principle forecasting methods employed in each category?

The principle forecasting methods employed in qualitative forecasting are:

• Committee judgment
• Juries of execution option
• Independent judgment
• Estimation of sales force

On the other hand, the major principle forecasting methods used in quantitative forecasting are:

• Causal – also known as proactive or multidimensional methods.

Causal methods relay on intuitive judgment of a person or a group of experts who predict future demand subjectively based on the goals and objectives of a company.

• Time series – also known as one dimensional or reactive methods.

Time series methods use patterns such as seasonality, trends and cyclic to predict that future demand will follow past patterns of sales.

• Advantages, disadvantages, and relative effectiveness of each method?

Quantitative methods are effective because they combine historic sales data with other factors such as seasonality, trend and cyclical. These methods also take into consideration unexplained variations by analyzing causal factors such as price, sales promotion and advertising to improve the accuracy of the prediction of future customer demand.

Advantages of time series methods

• It is possible to systematize time series methods and they don’t need big data storage.
• They are very effective when prediction on future customer demand is being done on large number of products.
• They prove to be very effective on short-term forecasting.
• They give almost accurate results when the products’ sales history is stable.
• There various software packages for time series methods readily available.
• They can remove small random variations by smoothing them out.
• Time series methods are not difficult to understand.

Disadvantages of time series methods

• They need a lot of historical data on sales.
• They take time to adjust to fluctuations in sales.
• Finding the smoothing weight can be tiresome and time consuming.
• They are not suitable for long-term forecasting.
• Huge changes in the already existing data can cause significant error.

Advantages of causal methods

• They can be performed on various available software packages.
• Due to the increased teaching of causal methods in institutions, most managers are familiar with these methods.
• They are more accurate in short-term and medium-term predictions as opposed to using only time series methods.
• These methods are able to do what-if analysis.

Disadvantages of causal methods

• Their accuracy is dependent on steady connection between dependent and independent variables.
• It is crucial to accurately identify the independent variable otherwise great errors will arise.
• They consume a lot of time to develop and one must have adequate knowledge on statistics.
• They are not easily systematized.
• They are more expensive to develop and maintain.

Advantages of qualitative methods

• the people who use these methods are experienced and they understand not only the events taking place in the market but also what is about to happen in future.
• When these methods are used by experts in a particular field or business, they prove to be relatively effective.

Disadvantages of qualitative methods

• They are prone to abuse by people who have little knowledge about related business and the market.
• They don’t show enough concrete evidence to support the prediction.
• Different people have different gut-feelings about each situation hence biased.
• In Chapter 4, you were introduced to various methods for assessing forecast accuracy.  For this question, begin by downloading the attached Excel File and complete the forecasting accuracy exercise.  When completed upload to the completed excel file here. 

Completed and Uploaded as an Excel file named ‘Question 4’

• This question references chapter 5 in your textbook: Quantitative forecasting Methods Using Time Series Methods.  Briefly explain the following Quantitative Time Series methods and give an example of when each method would be most appropriately applied:
• Moving Averages

It is a method used to determine the overall trend of a set of sales data by use of a past average of known past data to predict future sales. For example this method can be used in short-term forecasting of a select time periods such as demand for a product per week.

• Exponential Smoothing

This method uses a weighted sum of past observations, whereby it uses an exponentially decreasing weight with a geometrically decreasing ratio. This method can be used when data shows no clear seasonality or trend behavior.

• Single Exponential Smoothing

Also known as SES, single exponential smoothing method is a method that uses a smoothing parameter referred to as alpha, smoothing coefficient or smoothing factor to predict on univariate data which has no seasonality or trend. Alpha is set to be a number between 0 and 1. This forecasting method can be used on historic data which does not show seasonality or trend.

• Holt’s Two-Parameter Method

Also known as linear exponential smoothing this method uses three related equations to come up with the final forecast. The first equation uses the value of the past trend (last smoothed value), the second equation updates the trend itself by calculating the difference between the last smoothed values and the last equation generates the final prediction. This method is suitable when forecasting using historic data that has a trend.

• Use the attached Excel Spreadsheet template to complete a regression analysis tracking Regular Gas Prices (the dependent variable) as affected by Crude Oil prices (the independent variable).  Instructions for the three Tasks are included in the attached file.  

Completed and Uploaded as an Excel file named ‘Question 6’

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Reference

Chase, C. (2013). Demand-driven forecasting: A structured approach to forecasting. Hoboken, New Jersey: John Wiley & Sons, Inc.