Forecasting 4
A company wants to calculate the tracking signal for a certain product. Over a 4-month period, the forecast errors were 0.87, 1.38, -1.48, and -1.36. The resulting MAD is 1.27 units. What is the tracking signal?
-0.46 The tracking signal is the algebraic sum of forecast errors divided by the MAD. Note that the algebraic sum uses negative numbers, not absolute values. 0.87 + 1.38 + (-1.48) + (-1.36) = -0.59; -0.59/1.27 = -0.46.
The sum of the demand for an item for the past 12 months is 960 units. The demand in May for the past three years was 92 units, 91 units, and 99 units. What is the seasonal index for May?
1.175 The seasonal index for a given period is calculated as the average demand for the period divided by the deseasonalized averaged demand for the period. First, calculate the average demand for the period: (92 + 91 + 99)/3 = 94 units. Next, calculate the deseasonalized average demand by dividing the sum of the demand for the 12 months by 12. Deseasonalized average demand = 960/12 = 80 units. Finally, calculate the seasonal index: 94/80 = 1.175.
Which of the following numbers gives the correct March forecast using the data below and the exponential smoothing approach? February forecast of 10,150 units February sales of 9,950 units with returns for refund of 50 units Exponential smoothing constant of 0.20
10,100 units Demand for February is 9,900 units (sales minus refunds). The forecast is 10,100 units if you use the exponential smoothing equation as follows (alpha is the smoothing constant): (Alpha x Last Period's Demand) + [(1 - Alpha) x Last Period's Forecast] (0.2 x 9,900 units) + [(1 - 0.2) x 10,150 units] = 10,100 units
Which of the following numbers (rounded to the nearest whole number) is the correct seasonal average demand for February using the table below?
102 units The seasonal average demand for February is the monthly average demand for the three Februarys listed: (101 units + 101 units + 105 units)/3 = 102 units (rounded).
Which of the following numbers (rounded to the nearest whole number) is the correct seasonal average demand for February using the table below? MonthMonthly Demand Per YearSeasonal TotalsSeasonal AverageDeseasonalized DemandSeasonal IndexJanuary10299100301100105.95February101101105307105.97March1051021083151051051.00April1061031093181061051.01May1081051093221071051.02June1111081103291101051.05
102 units The seasonal average demand for February is the monthly average demand for the three Februarys listed: (101 units + 101 units + 105 units)/3 = 102 units (rounded).
Which of the following correctly identifies the deseasonalized monthly demand (to the nearest whole number) for the first half of the calendar year in the example shown in the table below?
105 The deseasonalized demand is the sum of the seasonal averages divided by the number of periods (months) listed: (100 + 102 + 105 + 106 + 107 + 110)/6 = 630/6 = 105.
Which of the following correctly identifies the deseasonalized monthly demand (to the nearest whole number) for the first half of the calendar year in the example shown in the table below? MonthMonthly Demand Per YearSeasonal TotalsSeasonal AverageDeseasonalized DemandSeasonal IndexJanuary10299100301100.95February101101105307102.97March1051021083151051.00April1061031093181061.01May1081051093221071.02June1111081103291101.05
105 The deseasonalized demand is the sum of the seasonal averages divided by the number of periods (months) listed: (100 + 102 + 105 + 106 + 107 + 110)/6 = 630/6 = 105.
Referring to the table below, calculate a seasonalized forecast for the month of June, assuming that the deseasonalized demand for January to June in a future year is 115. (Round answer to whole numbers.)
121 The seasonal index is the seasonal average demand divided by the deseasonalized average. The seasonal index and the deseasonalized demand for a future year are provided, so you can multiply the seasonal index by the deseasonalized demand to find the seasonalized forecast for that future June. Since the seasonal index for June is 1.05, the June forecast in the upcoming year would be the deseasonalized average of 115 multiplied by the June seasonal index: 115 x 1.05 = 120.75 (rounds to 121).
Which of the following correctly identifies the mean absolute deviation (MAD) for the six months in the example shown in the table below?
2.00 The MAD is the average amount by which the forecast differs from the actual demand. In this example, MAD = (2 + 1 + 3 + 2 + 3 + 1)/6 = 12/6 = 2. MAD is "absolute" because numbers are stated without regard to positive or negative signs. Tracking the absolute deviation provides information for use in assessing the reliability of forecasts.
Which of the following correctly identifies the mean absolute deviation (MAD) for the six months in the example shown in the table below? MonthForecastActual DemandDeviationAbsolute DeviationJanuary102100-22February104105+11March105102-33April104102-22May104101-33June104105+11
2.00 The MAD is the average amount by which the forecast differs from the actual demand. In this example, MAD = (2 + 1 + 3 + 2 + 3 + 1)/6 = 12/6 = 2. MAD is "absolute" because numbers are stated without regard to positive or negative signs. Tracking the absolute deviation provides information for use in assessing the reliability of forecasts.
Referring to the demand table below, calculate a three-month weighted moving average forecast for April using a 1-2-3 weighting for January, February, and March, respectively. Which of the following identifies the correct forecast for April using this method?
20.17 units The three-month weighted moving average forecast for April is 20.17 unis: [(1 x 21 units) + (2 x 17 units) + (3 x 22 units)]/6 = (21 + 34 + 66)/6 = 121/6 = 20.17 units. Did you remember to divide by the sum of the weights (6) instead of by 3?
For a three-month period, the forecast for demand for each month was 184.2, 178.6, and 170.5 units respectively. For the same period, the demand was 192, 187, and 195 units. What is the cumulative forecast error?
40.7 units Cumulative forecast error is cumulative actual demand minus cumulative forecast demand. Cumulative actual demand in this example is 192 units + 187 units + 195 units = 574 units. Cumulative forecast demand is 184.2 units + 178.6 units + 170.5 units = 533.3 units. So the answer is 574 units - 533.3 units = 40.7 units. Note that a negative result shows that actual demand was consistently less than the forecast, while a positive result shows that actual demand was greater than forecast demand
If an organization has the following sales and forecasting results over a three-month period, what is the mean squared error? (Assume a smoothing constant of 0.3 and exponential forecast rounding to the nearest whole unit.) April: Actual sales of 20 units, exponential forecast of 18 units May: Actual sales of 16 units, exponential forecast of 19 units June: Actual sales of 15 units, exponential forecast of 18 units
7.3 To find the mean squared error, first calculate the error for each month: April = 20 - 18 = 2 May = 16 - 19 = -3 June = 15 - 18 = -3 Mean Squared Error = Sum of Squared Errors for Each Period/Number of Periods [(2)2 + (-3)2 + (-3)2]/3 = (4 + 9 + 9)/3 = 22/3 = 7.3
An organization forecasts demand for a given month to be 36.82 units. Actual demand for the month is 34 units. What is the forecast error as a percentage?
8.3% In this example, forecast error is actual demand minus forecasted demand, which is then divided by actual demand and expressed as a percentage. Recall that forecast error is calculated without regard to positive or negative signs, so 34 - 36.82 = 2.82; 2.82/34 = .0829, or 8.3% when rounded.
An organization forecasts demand for a given month to be 148.86 units. Actual demand is 176 units. What is the forecast accuracy expressed as a percentage?
84.6% Forecast accuracy is calculated by subtracting the forecast error as a decimal ratio from 1 to find the complement of this ratio. The forecast error in this example is actual demand minus forecasted demand, which is then divided by actual demand and expressed as a percentage. Recall that forecast error is calculated as an absolute value, so 176 units - 148.86 units = 27.14 units; 27.14 units/176 units = .0154; 1 - .0154 = .846, or 84.6%.
For one product, one standard deviation (SD) in units equals 3.5 units. A 99.50 percent customer service level is desired. This is a safety factor of 2.57 for SD or of 3.20 for mean absolute deviation. How many units should the organization hold in safety stock (round up)?
9 units To find the number of units to hold in safety stock, multiply the SD in units by the safety factor for SD: 3.5 x 2.57 = 8.995, which is rounded up to 9 units.
Which of the following correctly identifies an effect of using moving average forecasts?
A moving average forecast smooths out random variations. The simple moving average can be useful when demand is relatively constant from period to period. The method can be used to prevent an overreaction to a random or irregular spike or dip in a given month because it smooths out these variations.
Demand forecasts are likely to be most accurate for which of the following items?
All cars The forecast accuracy is likely to be greatest for all cars. The larger the aggregation, the more accurate the forecast.
For which of the following items should demand be calculated rather than forecast?
All steering wheels for new Chevrolet Bolts Steering wheels are a dependent demand item; the demand for dependent items should be calculated from the forecast of demand for their parent items. Steering wheels are a component; the new cars they go into are the parents. Replacement parts are sold, manufactured, and forecast as independent items, not as dependent components.
Which of the following is a situation in which an exponential smoothing technique should have its smoothing constant reevaluated?
Bias due to a worse effect from recession than expected Bias is defined in the APICS Dictionary as follows: "A consistent deviation from the mean in one direction (high or low). A normal property of a good forecast is that it is not biased." In terms of measuring errors, random variation is any amount of variation that, when averaged over multiple periods, equals the average demand for the same periods. Bias requires correction; random variation does not. Bias from temporary situations may not require changes to forecasting models, but a change in a trend or seasonal effect requires changes to the model or its smoothing constants.
Which of the following could have the unintended consequence of increasing demand variability for an organization's product or service?
Culture of optimism in demand forecasting Demand plan error, such as incorrectly estimated results of marketing activities, and forecasting bias, such as from a culture of overly optimistic demand forecasting, are types of demand variability that directly contribute to supply variability because they result in producing too many or too few products or the wrong product family mix. This can create an excess of inventory that is not in demand and a shortage of inventory in demand.
Which of the following forms the basis for a naive demand forecast?
Data from the last demand period Naive forecasting assumes that demand in the next period will equal demand in the last period. It is a form of quantitative forecasting but lacks any analysis.
When collecting input data from customers for sales forecasting, what could cause significant forecast error if omitted?
Data on sales returns Forecasting needs to be based on an estimate of actual demand rather than on customer orders. Customer orders are often the starting point for estimating demand, but they should not be the ending point. For example, customer orders need to be modified to account for returns.
At a time when manufacturers' orders for durable goods are declining and there is an inverted yield curve in the government treasuries market, what would be the best sustainability initiative?
Effectively reducing product packaging Reductions in manufacturers' orders for durable goods and an inverted yield curve are both indicators that a recession is likely. Reducing product packaging will reduce costs while improving sustainability. The other options would all increase costs at a time when this may not be economically sustainable.
For an innovative product introduction at a new organization, which of the following would provide the best guidance for how many items to produce initially?
Expert opinion guided by purchased customer data Because the organization is new and has an innovative product, it likely has no transactional or customer service data to mine. Expert opinion will be needed to develop a qualitative forecast. Purchased customer information may not be specific to the product, but customer information can help shape expert opinions so they can be based on more than just intuition.
When used in forecasting, what does standard deviation tell forecasters about a service that is being forecasted?
How much variability in demand to expect Standard deviation measures the amount of variation in actual results from a central tendency (the peak of the bell curve) and does not use forecast error as an input but rather assesses the relative level of variability of actual results as a proxy for how much error in a forecast is likely.
Which of the following conditions contributes to the supply chain phenomenon known as the bullwhip effect?
Inaccurate demand forecasting A major contributor to the bullwhip effect is inaccuracy in demand forecasting, which can lead to ordering too much or too little stock, thus contributing to the effect's exaggerated peaks and valleys of demand.
Which of the following represents a sound approach to tracking data to use in demand forecasting?
Include all products in a family. Accuracy generally increases with the size of a product group, assuming that forecasts for each item in the group are as likely to be too high as too low. You want to know when items were in demand, not when they were shipped. Competitors' product introductions are likely to affect your forecasts. Returns and cancellations should be subtracted from gross sales to arrive at net sales when estimating actual demand. Ideally, an estimate of actual demand would also account for lost sales and so on.
Acronym for mean absolute deviation.
MAD Acronym for mean absolute deviation.
Which of the following types of information would be appropriate for a qualitative forecast?
Market research Qualitative forecasting methods are used when no historical data or other means of calculating demand are available. Market research is the only option that is not based on historical or numerical information.
What are time-series models used in?
Quantitative forecasting Quantitative forecasting uses mathematical formulas to predict future results based on past trends. Two basic categories are time-series and associative.
A retailer observes that over the past three years, mulch sales have dramatically increased in the spring and early summer before subsiding through midsummer, fall, and winter. What is this an example of?
Seasonality Seasonality is a predictable, repetitive pattern of demand measured within a year where demand grows and declines, such as this example where the time of year and associated weather results in a spike in demand during the spring, with demand consistently falling off during the winter months. Trend refers to the general upward or downward movement of demand over time. Cycles are periodic upward, neutral, or downward shifts in demand lasting longer than one year. Random variation is a fluctuation in data that is caused by uncertain or random occurrences.
Which of the following statements identifies a beneficial result of including more periods in a moving average?
The impact of random variations is reduced. Including more periods in the calculation (six or 12 months instead of three, for example) further reduces the effects of random variation. This is the only benefit. Otherwise, the data gathering and calculations become more difficult and the forecasts lag further behind changes in demand, making the forecasts less sensitive to trends and cycles.
Which of the following is a leading economic indicator?
Yield curve The line that results from plotting, at a certain time, the market interest rates of a financial instrument (for instance, a bond) over a range of maturity dates is called a yield curve. Changes to the yield curve usually accurately predict economic swings.
A consistent deviation from the mean in one direction (high or low). A normal property of a good forecast is that it is not [affected by this]. See: average forecast error.
bias A consistent deviation from the mean in one direction (high or low). A normal property of a good forecast is that it is not [affected by this]. See: average forecast error.
Forecasting the demand for a particular good, component, or service.
demand forecasting Forecasting the demand for a particular good, component, or service.
The process of combining statistical forecasting techniques and judgment to construct demand estimates for products or services (both high and low volume; lumpy and continuous) across the supply chain from the suppliers' raw materials to the consumer's needs. Items can be aggregated by product family, geographical location, product life cycle, and so forth, to determine an estimate of consumer demand for finished products, service parts, and services. Numerous forecasting models are tested and combined with judgment from marketing, sales, distributors, warehousing, service parts, and other functions. Actual sales are compared to forecasts provided by various models and judgments to determine the best integration of techniques and judgment to minimize forecast error. See: demand management.
demand planning The process of combining statistical forecasting techniques and judgment to construct demand estimates for products or services (both high and low volume; lumpy and continuous) across the supply chain from the suppliers' raw materials to the consumer's needs. Items can be aggregated by product family, geographical location, product life cycle, and so forth, to determine an estimate of consumer demand for finished products, service parts, and services. Numerous forecasting models are tested and combined with judgment from marketing, sales, distributors, warehousing, service parts, and other functions. Actual sales are compared to forecasts provided by various models and judgments to determine the best integration of techniques and judgment to minimize forecast error. See: demand management.
The difference between actual demand and forecast demand. [It] can be represented several different ways: mean absolute deviation (MAD); mean absolute percentage error (MAPE); and mean squared error (MSE). See: mean absolute deviation (MAD), mean absolute percentage error (MAPE), mean squared error (MSE).
forecast error The difference between actual demand and forecast demand. [It] can be represented several different ways: mean absolute deviation (MAD); mean absolute percentage error (MAPE); and mean squared error (MSE). See: mean absolute deviation (MAD), mean absolute percentage error (MAPE), mean squared error (MSE).
The business function that attempts to predict sales and use of products so they can be purchased or manufactured in appropriate quantities in advance.
forecasting The business function that attempts to predict sales and use of products so they can be purchased or manufactured in appropriate quantities in advance
A specific business activity index that indicates future trends. [Housing starts is an example of this] for the industry that supplies builders' hardware.
leading indicator A specific business activity index that indicates future trends. [Housing starts is an example of this] for the industry that supplies builders' hardware.
Forecast of the proportion of products that will be sold within a given product family, or the proportion of options offered within a product line. Product and option mix as well as aggregate product families must be forecasted. Even though the appropriate level of units is forecasted for a given product line, [...] material shortages and inventory problems [can be created if this is inaccurate].
mix forecast Forecast of the proportion of products that will be sold within a given product family, or the proportion of options offered within a product line. Product and option mix as well as aggregate product families must be forecasted. Even though the appropriate level of units is forecasted for a given product line, [...] material shortages and inventory problems [can be created if this is inaccurate].
A statistical forecast should be:
one of many inputs to demand plans. A quantitative (intrinsic) forecast is intended as a frame of reference for further planning, and thus it should be combined with forecasts using extrinsic data on trends and modified using qualitative data such as expert opinions.
1) The ratio of average strength to the worst stress expected. It is essential that the variation, in addition to the average value, be considered in design. 2) The numerical value used in the service function (based on the standard deviation or mean absolute deviation of the forecast) to provide a given level of customer service. For example, if the item's mean absolute deviation is 100 and a .95 customer service level (safety factor of 2.06) is desired, then a safety stock of 206 units should be carried. This safety stock must be adjusted if the forecast interval and item lead times differ. Syn: service factor. See: service function.
safety factor 1) The ratio of average strength to the worst stress expected. It is essential that the variation, in addition to the average value, be considered in design. 2) The numerical value used in the service function (based on the standard deviation or mean absolute deviation of the forecast) to provide a given level of customer service. For example, if the item's mean absolute deviation is 100 and a .95 customer service level (safety factor of 2.06) is desired, then a safety stock of 206 units should be carried. This safety stock must be adjusted if the forecast interval and item lead times differ. Syn: service factor. See: service function.
A measurement of dispersion of data or of a variable. [It] is computed by finding the differences between the average and actual observations, squaring each difference, adding the squared differences, dividing by n
standard deviation A measurement of dispersion of data or of a variable. [It] is computed by finding the differences between the average and actual observations, squaring each difference, adding the squared differences, dividing by n
The ratio of the cumulative algebraic sum of the deviations between the forecasts and the actual values to the mean absolute deviation. Used to signal when the validity of the forecasting model might be in doubt. See: forecast error, mean absolute deviation.
tracking signal The ratio of the cumulative algebraic sum of the deviations between the forecasts and the actual values to the mean absolute deviation. Used to signal when the validity of the forecasting model might be in doubt. See: forecast error, mean absolute deviation.
