Sources of Demand/Forecasting
Since input data can significantly impact forecast results, an important data preparation principle is to_____ preparation principle 1
****forecast based on demand rather than orders Forecasts of demand for production need to use demand estimates rather than ship dates or sales. Forecasts are made to estimate market or organizational demand and therefore customer requests or ship dates provide incomplete information for forecasting
preparation principle 3
****record any circumstances or events related to the data. This includes regular events, such as sales promotions or price changes, as well as unusual or one-time events, such as extreme weather or labor disruptions. A promotion should cause an increase in demand, and data on this will help marketing assess the effectiveness of the promotion. If the promotion will not be repeated, the data or the results might be adjusted to account for this. Similarly, the impact of one-time events that cause a spike or drop in demand might be adjusted or smoothed out prior to forecasting.
independent demand
The demand for an item that is unrelated to the demand for other items. Demand for finished goods, parts required for destructive testing, and service parts requirements are examples of independent demand. comes from a number of internal and external sources including: - forecasting - end customers (finished goods and service parts) - replenishment orders - interplant demand or intercompany transfers (ex: between subsidiaries) - internal use (ex: R&D, quality control, destructive testing, marketing use) *** (Everything listed except forecasting represents actual orders, and this demand is differentiated from forecasted demand by calling it actual demand)
forecasting process: step 3, select horizon and planning bucket
The horizon can be long-, medium-, or short-term. The planning bucket is the period being forecasted—for example, quarterly, monthly, weekly, daily. Horizon, aggregation, and units are interrelated and are very important to get right. A long-term horizon (years or quarters) usually uses total sales in dollars; medium-term (quarters or months) uses product families in units; and short-term (weeks or days) uses products or SKUs in units.
The specific way materials and goods move depends on many factors, including:
The type and quantity of distribution channels available for use Market characteristics Product characteristics Available modes of transportation.
bill of material
a list of all parts, materials, or ingredients that comprise a single unit
exponential smoothing forecast
A type of weighted moving average forecasting technique in which past observations are geometrically discounted according to their age. The heaviest weight is assigned to the most recent data. The smoothing is termed exponential because data points are weighted in accordance with an exponential function of their age. The technique makes use of a smoothing constant to apply to the difference between the most recent forecast and the critical sales data, thus avoiding the necessity of carrying historical sales data. The approach can be used for data that exhibits no trend or seasonal patterns. Higher order exponential smoothing models can be used for data with either (or both) trend and seasonality. equation: new forecast = (a * latest demand) + (( 1 - a) * previous forecast) It gives more weight to the most recent demand information because it is based on three things: the last period's actual demand, the last period's forecast result, and a smoothing constant called alpha (α). Alpha is a number between 0.0 and 1.0 that is basically a percentage weighting where 1 equals 100 percent. It determines how much to weight the prior actual demand versus the forecast demand. It is also sometimes called a smoothing constant
forecast interval: weekly
A weekly forecast is necessary for master scheduling and can be achieved by dividing monthly product family forecasts into weekly buckets for individual products. A weekly level of detail may not be necessary for other forecasts if this level of detail does not improve planning. A weekly forecast interval increases data management and can impart a false sense of precision in this situation.
forecasting process: step 8, forecast
After making adjustments as needed, start using the forecast. If seasonality was removed for forecasting, add it back in before preparing reports that include ranges of confidence levels.
qualitative forecasting techniques
An approach to forecasting that is based on intuitive or judgmental evaluation. It is used generally when data is scarce, not available, or no longer relevant.
quantitative forecasting techniques
An approach to forecasting where historical demand data is used to project future demand. Extrinsic and intrinsic techniques are typically used.
is there a difference between correlation and causation
yes Correlation is an observation that the change in an independent variable has a measurable effect on a dependent variable. However, just because the effect can be reliably observed over time does not mean that the one thing caused the other thing. It could be that some third force is affecting both of them, or the correlation could be a coincidence.
another principle of data collection is that_______ preparation principle 2
****the data should be collected in the format in which they will be used in forecasting. For example, if the time bucket will be weekly, then collecting monthly data will be insufficient.
preparation principle 4
**Separating demand by customer segment Demand from distribution centers, wholesalers, retailers, and individuals might be in very different quantities and have very different timing. Summing the demand from these segments prior to forecasting might produce misleading results
3 different types of exponential smoothing
- Adaptive smoothing: A form of exponential smoothing in which the smoothing constant is automatically adjusted as a function of forecast error measurement. - First-order smoothing or single exponential smoothing: A single exponential smoothing; a weighted moving average approach that is applied to forecasting problems where the data does not exhibit significant trend or seasonal patterns. - Second-order smoothing or double smoothing: A method of exponential smoothing for trend situations that employs two previously computed averages, the singly and doubly smoothed values, to extrapolate into the future.
what are the steps in deseasonalization using a seasonal index and forecasting
1. calculate the period average demand: by summing all like periods in the series and dividing by the number of them. 2. calculate the average demand for all periods: Sum the 12 month averages and divide by 12 3. calculate the season index: (period average demand) / (average demand for all periods) 4. calculate deseasonalized demand: (period raw demand) / (seasonal index) 5. forecast using the deseasonalized data 6. apply the proper seasonal index to the deseasonalized forecast results: forecasted demand = seasonal index * deseasonalized period forecast
distributor
A business that does not manufacture its own products but instead purchases and resells these products. Such a business usually maintains a finished goods inventory.
intrinsic forecasting method
A forecast based on internal factors, such as an average of past sales. One thing to note about intrinsic methods in general is that none of them really work well for longer-term forecasting, because you rapidly run out of actual data.
extrinsic forecasting method
A forecast method using a correlated leading indicator; for example, estimating furniture sales based on housing starts. Extrinsic forecasts tend to be more useful for large aggregations, such as total company sales, than for individual product sales. Extrinsic forecasting is also known as causal forecasting, associative correlation, or explanatory forecasting. These techniques are best for long-term forecasting at the aggregate level. Extrinsic forecasting uses cause-and-effect associations to predict and explain relationships—or correlation—between variables. That is, extrinsic forecasting works to find a link between information that is available and demand. The variables used in extrinsic forecasting techniques can come from internal or external data sources. The predictor is called the independent variable; the element being predicted is called the dependent variable. Simple and multiple regression analysis are common examples of extrinsic forecasting, with the difference being whether one or several indicators are used The best information that can be used in an extrinsic forecast is leading indicators
time series forecasting
A forecasting method that projects historical data patterns into the future. [It] involves the assumption that the near-term future will be like the recent past The methods used in time series forecasting are less complex mathematically and thus easier to explain to decision makers than extrinsic forecasts. Time series methods assume that the factors that have influenced the past will continue on into the future. When that trend is unlikely to be stable, causal/associative forecasting may be needed. There are a number of types of time series forecasting, ranging from the very simple to the relatively complex. Naive forecasting, life cycle analysis, moving averages, exponential smoothing, and decomposition are examples
decomposition
A method of forecasting where time series data is separated into up to three components—trend, seasonal, and cyclical—where trend includes the general horizontal upward or downward movement over time; seasonal includes a recurring demand pattern such as day of the week, weekly, monthly, or quarterly; and cyclical includes any repeating, nonseasonal pattern. A fourth component is random—that is, data with no pattern. The new forecast is made by projecting the patterns individually determined and then combining them. The decomposition of time series data seeks to understand the patterns of demand in a given sequence of historical data.
forecast interval: monthly
A monthly forecast is not too detailed, and it gives an adequate level of precision. This forecast interval is the most common choice of forecasters, as it allows detection of seasonal patterns that are hidden in quarterly forecast intervals.
time bucket
A number of days of data summarized into a columnar or row-wise display. For example, a weekly time bucket contains all the relevant data for an entire week. Weekly time buckets are considered to be the largest possible (at least in the near and medium term) to permit effective MRP [material requirements planning].
seasonality
A predictable repetitive pattern of demand measured within a year where demand grows and declines. These are calendar-related patterns that can appear annually, quarterly, monthly, weekly, daily and/or hourly. The key point about seasonality is that it repeats over the analysis period and thus can be isolated from other sources of variation and removed temporarily so that it will not influence forecasting.
delphi method
A qualitative forecasting technique where the opinions of experts are combined in a series of iterations. The results of each iteration are used to develop the next, so that convergence of the experts' opinions is obtained. The Delphi method is a more involved and sophisticated qualitative forecast. It involves surveying experts and collating their responses into a document that keeps the responses anonymous. Anonymity is used for two reasons. - First, it helps prevent dominant personalities from influencing the group opinion. When this "groupthink" effect is in play, otherwise independent thinkers might become emotionally committed to an unrealistic forecast. - The other problem anonymity prevents is a "stake in the ground" mentality—when a person has already publicly committed to a forecast result and doesn't want to lose face by changing his or her declared position. Since the Delphi method is anonymous, it is easier to change a position given more information. The Delphi method has had good success at arriving at reliable forecasts, but it is time-consuming and labor-intensive. It is often used only for strategic-level estimation.
leading indicators
A specific business activity index that indicates future trends. For example, housing starts is a leading indicator for the industry that supplies builders' hardware. These economic or demographic indicators tend to be among the first types of data that can show a change in a trend. Study of extrinsic data may show a statistical relationship between one or more leading indicators and demand. The relationship may be positively correlated or negatively correlated. The key point is that extrinsic forecasting is often the best way to detect changes in a trend. It is best used in long-term planning at the total company demand or product family level, because correlation is a macroeconomic observation. Determining the inputs to these methods requires statistical analysis comparing the organization's historical demand levels against data on the indicator(s) over time, which will result in the input data needed for a simple formula as well as some statistics like the relative level of correlation between the indicator and demand.
backorder
An unfilled customer order or commitment. A backorder is an immediate (or past due) demand against an item whose inventory is insufficient to satisfy the demand.
time series analysis
Analysis of any variable classified by time in which the values of the variable are functions of the time periods. Time series analysis is used in forecasting. A time series consists of seasonal, cyclical, trend, and random components.
base series
Another term that you may hear used when calculating the seasonal index is base series A standard succession of values of demand-over-time data used in forecasting seasonal items. This series of factors is usually based on the relative level of demand during the corresponding period of previous years. The average value of the base series over a seasonal cycle is 1.0. A figure higher than 1.0 indicates that demand for that period is higher than average a figure less than 1.0 indicates less-than-average demand. For forecasting purposes, the base series is superimposed upon the average demand and trend in demand for the item in question.
demand management what are the demand inputs?
Channel family-level forecasts disaggregated to the mix level and then to the end-item level at lowest level stocking points Item forecasts for lowest level stocking points (using time series analysis and other techniques) rolled up through regional to a systemwide total for use by master scheduling
forecasting process: step 4, gather and visualize the data
Collect the data and organize them using the planning bucket chosen. Then put the data in chart form to reveal trends, seasonality, or random variation. Visualizing helps in selecting the right forecasting technique.
how does one choose the best forecast method? The choice of forecasting method should consider the degree to which the available data are complete versus incomplete and the degree of stability in the data:
Complete versus incomplete data: If all the required sales data for a particular item as well as the causal variables are available, then the data are considered complete. Incomplete data would have limited—or even an absence of—sales data for a particular product or would not have the causal variables identified. Stable versus unstable data: Stable data have a distinct pattern such as seasonality or trends. There's a randomness and no distinct pattern to unstable data.
customer relationship management what are the demand inputs?
Customer orders pending release to the supplier Changes in ordering patterns
producer storage with drop ship
Customer puts in an order request TO retailer and the retailer puts that order request through TO the producer. the producer then sends the inventory TO the customer is used for items that customers like to view in stores but that are too bulky to have in retail inventory, such as refrigerators. The difference in this model is that the demand comes from the retailer rather than the customer.
cycle
Cycles usually refer to the wavelike patterns observed in the growth and recession trends of the economy over years. Unlike seasonality, economic cycles do not repeat over a predictable period of time, so this type of forecasting is left to economists. Another example of a cycle that could influence a trend is a product's life cycle.
dependent demand
Demand that is directly related to or derived from the bill-of-material structure for other items or end products. Such demands are therefore calculated and need not and should not be forecast. A given inventory item may have both dependent and independent demand at any given time. For example, a part may simultaneously be the component of an assembly and sold as a service part.
business planning and their use for forecasts
Executive involvement in forecasting at the business planning level helps to minimize the business risk of getting the forecasts wrong at the strategic level, while executive involvement at the S&OP level helps to reconcile the functional plans of the various departments (e.g., engineering, manufacturing, human resources, finance, sales) In a technique known as pyramid forecasting, the marketing and sales forecasts can be rolled up and reconciled against the management forecast. Executives and senior management use forecasting in setting the direction for the organization, planning expansions of products, considering new product line proposals, and evaluating potential mergers, acquisitions, and international ventures Senior managers often use causal models, statistical regression tools, and correlation analysis to improve the results of their forecasting efforts
as MPC (manufacturing planning and control) moves from strategic planning to master production scheduling, forecasts need to be more detailed and should require less executive management direction The entity or function that will be using the forecast strongly impacts both the horizon and interval choices:
Executives/senior management (strategic and manufacturing business plan levels): - Horizon—two to ten years for strategic plans, about 15 months for annual budgets. - Interval—annually, sometimes less frequently Sales and operations planning (tactical level): - Horizon—15 to 24 months - Interval—standard practice is monthly; in some cases, quarterly Master scheduling and control (operational level): - Horizon—from several weeks to 18 months (minimum cumulative lead time) - Interval—ongoing, numerous updates (depending on manufacturing environment)
forecasting process: step 6, prepare the data for the technique
For example, if there is seasonality, it should be temporarily removed prior to forecasting.
forecasting process: step 1, determine purpose
For example, it could be to determine demand for production, capacity requirements, or staffing levels.
forecasting process: step 9, achieve consensus on the forecast
For example, the sales and operations planning process is used to get everyone to agree to plan and execute using the same forecast numbers. This is called a one-number system.
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, an inaccurate mix forecast can create material shortages and inventory problems.
master planning and scheduling and their use for forecasts
Forecasts are used by master planning and scheduling to determine the number of finished products or components at the individual end-item level to make and when to make them. The master production schedule also feeds into the rough-cut capacity plan. With frequent changes in production priorities as the start of production nears, inventory allocations and shipment destinations also require frequent and detailed forecasts. This function uses time series data analysis techniques such as moving averages and exponential smoothing to generate its demand forecasts.
distribution requirements planning and their use for forecasts
Forecasts used in distribution planning are for both product families (the forecast created for S&OP) and end items or stockkeeping units (from the master scheduling forecast) at distributed inventory stocking locations. One such forecast is the mix forecast. While the item forecast is what the actual replenishment timing and quantity are based on, the aggregate product family forecast is important for two reasons: 1. During S&OP, the product family forecast is used to ensure that resources are available to transport, handle, and store replenishment inventory within major distribution channels. The process that addresses channel resource needs during S&OP is logistics resource requirements planning (LRRP). 2. During DRP, which is an input to master scheduling, product family forecasts by channel are disaggregated to the item-mix level and then are allocated down through the bill of distribution to plan the replenishment of inventory stocking locations.
forecasting is needed for different purposes in different manufacturing environments. give examples of different types of uncertainty in each manufacturing environment that may require forecasting
Make-to-stock—variations in demand stated in forecasts for each inventory location Assemble-to-order—variations in quantity, customer order timing, and product mix Make-to-order—size of the backlog and the level of company resources necessary to finish the engineering and produce the products per the specifications Engineer-to-order—hiring difficult-to-find specialists such as design engineers and ordering materials with long lead times
forecasting process: step 10, continuously improve
Monitor error levels and set policies for when error levels are too high. When they are, refine communications, data, or processes and techniques.
moving average
Once the data are deseasonalized, they are ready for use in forecasting with the moving average or weighted moving average techniques A moving average is simply the average of a certain number of past periods of demand. It is called moving because the latest set of periods is always used. An arithmetic average of a certain number (n) of the most recent observations. As each new observation is added, the oldest observation is dropped. The value of n (the number of periods to use for the average) reflects responsiveness versus stability in the same way that the choice of smoothing constant does in exponential smoothing. There are two types of moving average: simple and weighted. equation: moving average forecast = (sum of demand for most recent set of periods) / number of periods Common examples include three- and six-period moving averages. As you forecast the next period, you drop the oldest period and add the most recent period. 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. However, if there is a change in a trend, this method would be slow to respond to it. It would lag the trend, in other words.
forecasting process: step 5, choose the forecasting technique
One or a combination of methods may work best for the combination of purpose, aggregation, time horizon, data availability, trend, and where the data fall on the stable-dynamic continuum.
dynamic demand patterns
Products and services with dynamic demand patterns, such as innovative products, may have shifting trends, no seasonality or seasonality that shifts unpredictably, and/or a high degree of random variation that masks trends and seasonality. In these cases, forecasting is trying to find only the base (average) demand for use in planning; more flexible manufacturing strategies may be needed. Another option might be to find ways to make a dynamic demand pattern more stable (e.g., communicate better to reduce or eliminate the bullwhip effect).
marketing what are the demand inputs?
Promotions that will cause demand spikes
forecast interval: quarterly
Quarterly forecasts are appropriate in a few industries with long production lead times, as in engineer-to-order environments. They are also appropriate in the later years of a multi-year forecast but may hide seasonal demand patterns.
random variation
Random variation is any variation left over after seasonality and trends have been accounted for. Random variation reflects that customers vary when, where, and in what quantities they buy products; the level of variation can differ greatly. If random variation is small, forecasting will be fairly accurate. If it is large, errors will be high. To find the relative amount of random variation, sum the differences between the forecast and actual sales for a reasonable period of time for the given forecast type. Variance around the average that results in no bias (no consistent deviation from the mean in one direction) is the random variation. For this reason, random variation and bias are often reviewed together.
distribution centers and customers
Reports of special events that will cause demand spikes Anomalous purchases in recent periods
sales what are the demand inputs?
Sales force estimates for inventory storage locations Replenishment needs for vendor-managed inventory
forecasting process: step 7, test the forecast using historical data
Since periods in the past already have actual results, you can forecast using, say, June data to produce a July forecast and then compare it to what actually happened in July.
forecasting process: step 2, set level of aggregation and units of measure
Specify total sales in units or dollars, product families, products, or SKUs (stockkeeping units).
correlation
The relationship between two sets of data such that when one changes, the other is likely to make a corresponding change. If the changes are in the same direction, there is positive correlation. When changes tend to occur in opposite directions, there is negative correlation. When there is little correspondence or changes are random, there is no correlation.
sales and operations planning and their use for forecasts
Those working in sales and operations planning use forecasts to identify planned sales and product family output and to reconcile functional plans. These forecasts will be in monetary and volume terms at the product family level. Customer plans and current demand information are the key inputs into these forecasts Historical or time series data are also used as inputs into these forecasts. S&OP often uses aggregate forecasts based on forecasts for individual products in each product line and refined using current trend information and marketing data on customers' future plans that could influence demand. S&OP may use regression analysis and time series forecasts, including decomposition.
roll up forecast key steps
Units of individual products are forecasted and then denominated in dollars. Unit and dollar values are consolidated at the product family level. Product family dollars are combined in a total roll-up forecast.
distribution requirements planning (DRP) addresses the following issues:
When, where, and how much inventory is needed at the distribution centers within a manufacturer's distribution network in order to fulfill customer orders What resources are needed and when to ensure that replenishment inventory can be delivered, handled, and stored according to the replenishment plans
transaction channel
a distribution network that deals with change of ownership of goods and services including the activities of negotiation, selling, and contracting transaction channel addresses the transfer of funds and ownership between the selling organization and the consumer. ex: sales department - distribution partner - retail location - consumer has a reverse logistics role. transaction channel will need to authorize the return as well as move refunds downstream to the intermediate and/or final customer.
pyramid forecasting
a forecasting technique that enables management to review and adjust forecasts made at an aggregate level and to keep lower-level forecasts in balance. This technique includes a number of forecasts that may rely on quantitative forecasting methods to one degree or another: - Sales and marketing forecasts at the individual product level (sales volume and units) - product family forecasts based on a roll-up of product-level forecasts (sales volume and units) - A business-level forecast at the corporate level (sales volume) that is more strategic and could reflect plans to expand promotions, open new marketing channels, or compete in new market segments - A management forecast that reconciles the roll-up and business-level forecasts Pyramid forecasting combines both quantitative and qualitative elements. It provides a means of coordinating and integrating multiple forecast sources in an organization and ensuring consistency among the organization's strategic direction, goals or constraints, marketing and sales efforts, and use of manufacturing resources.
multiple regression models
a form of regression analysis where the model involves more than one independent variable Multiple regression is an extension of simple regression; there are multiple predictive variables rather than just one. For example, one could add marketing spend to the roofing sales analysis to determine if this increases or decreases the predictive value of the equation.
panel consensus
a judgmental forecasting technique by which a committee, sales force, or group of experts arrives at a sales estimate A more structured version of this is called the Delphi method.
seasonal index
a number used to adjust data to seasonal demand. A seasonal index provides information on how much each seasonal period's demand has varied from the average demand in the past, and the index is therefore used to estimate how much seasonal demand will vary from average demand in a future seasonal period. A seasonal index of 1.0 means that the period's demand is equal to average demand. A value less than 1.0 means that the period's demand is less than the average; a value greater than 1.0 shows that the period's demand is greater than average. A simple use for the seasonal index is to create a seasonal forecast, which applies seasonality to a forecast of average demand per period.
life cycle analysis
a quantitative forecasting technique based on applying past patterns of demand data covering introduction, growth, maturity, saturation, and decline of similar products to a new product family. This is a comparative technique that analyzes and adapts existing data and patterns and extrapolates them to create a forecast based on historical information.
econometric model
a set of equations intended to be used simultaneously to capture the way in which dependent and independent variables are interrelated. An econometric model could be used to explain the demand for housing starts by looking at the consumer base, internet interest rates, personal incomes, and land availability.
regression analysis
a statistical technique for determining the best mathematical expression describing the functional relationship between one response and one or more independent variables. The first thing to do in regression analysis is to visualize the data.
judgement/expert opinion approach
a team of experts or other stakeholders use their detailed knowledge of products, customers, forecasting methods, historical error rates, and the market to generate a forecast or, more frequently, to adjust the quantitative forecast. Stakeholders may include senior management, salespersons, marketing committees, customer service representatives, and others. Tracking these adjustments separately from any quantitative component will help in determining whether the modifications are increasing or decreasing in accuracy and in showing whether they are introducing bias by being consistently high or low. One judgmental technique used is panel consensus
historical analogy
a...technique based on identifying a sales history that is analogous to a present situation, such as the sales history of a similar product, and using that past pattern to predict future sales. Using a similar product as a proxy should be explored as a possibility prior to resorting to a pure qualitative forecast. This is called forecasting by historical analogy
smoothing constant
also known as smoothing factor or alpha factor In exponential smoothing, the weighting factor that is applied to the most recent demand, observation, or error. In this case, the error is defined as the difference between actual demand and the forecast for the most recent period. The weighting factor is represented by the symbol α. Theoretically, the range of α is 0.0 to 1.
curve fitting
an approach to forecasting based on a straight line, polynomial, or other curve that describes some historical time series data.
weighted moving average
an averaging technique in which the data to be averaged is not uniformly weighted but is given values according to its importance. The weighted moving average (or weighted average) forecasting method places weights on the periods being averaged, usually to put greater emphasis on the more recent periods and relatively less on the more distant periods. When calculating the weighted moving average, you divide by the sum of the weights rather than the number of periods. One drawback to moving averages is that to do this forecasting, you need to assemble data from multiple periods.
dependent demand inventory examples
assemblies and subassemblies fabricated and purchased components raw materials
stable demand patterns
demand is predictable Stable demand patterns may have fairly steady trends, predictable seasonality, and minimal random variation. Products with stable demand patterns can be forecasted with low error, so make-to-stock manufacturing environments can be used profitably.
what are the 10 steps in the forecasting process
determine purpose sat level of aggregation and units of measure select horizon and planning bucket gather and visualize the date choose the forecasting technique prepare the data for the technique test the forecasting using historical data forecast achieve consensus on the forecast continuously improve
direct/internal channel
directly from manufacturer to customer The direct/internal channel is often used for transfers to other plants or business units or for B2B sales. Demand could be for inventory or for ETO or MTO goods. B2C sales typically rely on a network of distribution centers rather than directly shipping from the manufacturer, except perhaps in the case of ETO or MTO goods. Regardless, demand in this case is directly from the customer.
item forecasts are used for....
distribution requirements planning (DRP) for inventory stocking locations such as regional DCs, satellite DCs, and stores These forecasts rely on a number of sources that provide and analyze the data used for estimates of independent demand in DRP. Forecasts are more accurate at the aggregate level, so even though it is necessary to break the forecast into individual distribution centers, accuracy will be affected.
qualitative techniques
forecast probable events based on human judgment, experience, intuition, and educated guesses rather than math. While they lack scientific precision, these techniques are often used in volatile situations or when there are no historical data available, such as for a new product or to create a strategic-level forecast where the long time horizon results in significant uncertainty and historical data would not be highly useful. Qualitative methods are also useful in modifying a quantitative forecast based on things like assumptions about the economy or competitor actions qualitative forecasts depend on the relevant experience level of the forecasters Bias is another risk. At a high level, forecast bias is basically a consistent deviation from the mean in one direction, and, with a qualitative method, this could be due to optimism or aggressive management goals. Qualitative techniques include historical analogy, judgmental/expert opinion, and the Delphi method. hard to maintain consistency in the level of accuracy in these types of forecasts over time. Organizations work to improve these types of forecasts by conducting market research or test marketing or studying similar or competitor products
tactical level plans
forecasting is conducted using aggregate information on product families for a time horizon of one to three years. Aggregate-level forecasts here are an input to the sales and operations planning process, and the output will be a single set of demand numbers that demand professionals agree to generate in terms of sales, supply professionals agree to produce, and finance professionals agree will be financially profitable. This information is used to plan budgets and labor requirements as well as to acquire items that have long lead times.
operational level plans
forecasting is conducted using item-level information over the time horizon required for master scheduling. This horizon will be determined by the lead times required for ordering, receiving, and using all materials and resources needed to produce the requested items plus a little extra time to give planners some leeway.
trend
general upward or downward movement of a variable over time (e.g., demand, process attribute). Trends can also be flat. trends can be influenced to varying degrees internally by things like promotions and externally by things outside one's control, such as an economic cycle.
what is the key difference between dependent and independent demand
independent demand originates from sources outside the control of the organization while dependent demand originates from internal sources or sources the organization can control. Often independent demand will be for items that the organization sells as individual units and dependent demand will be for the materials used to make those units. Independent demand is determined using forecasts and order management. Dependent demand is not forecasted; instead, it is calculated as part of the material requirements planning process. Items can be subject to both independent and dependent demand.
collecting data on demand involves what
involves collecting information on customer requests rather than just tracking sales or shipments. the organization should track when the customer wanted it to be shipped, since this provides information on demand for that period. Thus estimating total demand for a given period might involve adding the following: - Net sales (sales minus returns) - Backorders - Requests that could not be fulfilled due to rejected lead time or price quote - Requests that had to be filled from other plants or locations
echelon
is a category of the supply chain, such as a distribution center, a distributor, or a retailer. A given supply chain may or may not include a given echelon. While these intermediaries are not the final customer, they are the ones placing resupply orders, and so demand from these sources needs to be taken into account. These sources of demand will aggregate the demand from their downstream supply chain customers.
exclusive and select channel
manufacturer TO distributor TO retailer TO customer in the exclusive and select channel, a distributor is shown placing the orders Similar networks could have a distribution center or other intermediary placing the orders.
complex channel
manufacturer TO manufacturer* TO distributor TO retailer TO customer In the complex channel, demand is from an internally owned distribution center, which is likely directly connected to the manufacturer's master production schedule through a distribution requirements planning module.
Manufacturing Planning and Control for Supply Chain Management notes that it is critical that
material and capacity planners identify all sources of demand for items. Because they are managed in different ways, it is important to distinguish between items with independent demand and dependent demand.
least-squares method
method of curve fitting that selects a line of best fit through a plot of data to minimize the sum of squares of the deviations of the given points from the line. Simple regression (also called linear regression) uses the least-squares method The least-squares method is used to make an association between the dependent variable y (the thing you are trying to predict) and the independent variable x (the predictor) y=mx+b
business to business commerce (B2B)
one of the two basic categories of demand requires a distribution channel and a transaction channel Business conducted over the internet between businesses. The implication is that this connectivity will cause businesses to transform themselves via supply chain management to become virtual organizations—reducing costs, improving quality, reducing delivery lead time, and improving due-date performance.
2 different forecasting techniques
qualitative: uses judgment and expertise quantitative forecasting: which uses mathematics, intrinsic (time series) and extrinsic (casual) These methods can be used individually or sometimes in combination,
independent demand inventory examples
retail, wholesale, and manufactured finished goods service or replacement parts MRO (maintenance/repair/operations) supplies
naive forecasting
simply assumes that the last period's demand will be this period's forecast for example, that last June's results will be this June's results It can be cost-effective but does not account for trends, and any random spike or trough in demand would be carried forward. These basic models will be off by a substantial amount when random variation has occurred in the prior period, since they will repeat that spike or dip.
distribution channels
the distribution route, from raw materials through consumption, along which products travel distribution channel handles the disposition of goods or services ex: production facility - road transportation - warehouse - road transportation - retail location - delivery partner - consumer has a reverse logistics role. The distribution channel will need to be able to move returned items upstream Distribution channels describe the various echelons in the supply chain that generate demand for products and services
forecast horizon
the period of time into the future for which a forecast is prepared
business to consumer sales (B2C)
the second category of demand also requires a distribution channel and a transaction channel Business being conducted between businesses and final consumers, largely over the internet. It includes traditional brick and mortar businesses that also offer products online and businesses that trade exclusively on the internet.
forecast interval
the time unit for which forecasts are prepared, such as week, month, or quarter there are advantages and disadvantages to each interval
strategic and manufacturing business plan levels
time horizons are very long. The purpose of these forecasts is to determine total sales volume for the organization in dollars so the organization can acquire or divest capital assets and prepare to enter new markets or collaborate in new supply chains. A key result will be long-term financial goals that the organization's functions will work to achieve.
As part of demand management's forecasting activities, all manufacturing environments require the forecast and other critical information to be shared among....
various functions, including sales and operations planning (S&OP), master scheduling (MS), and their customers Each environment type has different information needs for those entities: - Make-to-stock—demand forecasts for S&OP; actual demand for the master production schedule (MPS); next inventory replenishment for customers - Assemble-to-order—demand forecasts, product family mix for S&OP; mix forecasts and actual demand for the MPS; configuration issues, delivery date for customers - Make-to-order—demand forecasts, engineering detail for S&OP; final configuration for the MPS; design status, delivery date for customers - Engineer-to-order—demand forecasts based on similar products previously created, with the focus on engineering hours, for S&OP and the MPS; delivery lead time for customers
what are the two steps in time series forecasting
visualizing and deseasonalizing Visualizing the data is an important part of forecasting. You can often spot seasonality or some other trend or cycle quickly and decide how best to conduct the forecast. Seasonality can be observed for the months in a year or the hours in a day and so on. The first thing to do when preparing to use a time series forecast is to determine if there is seasonality
