WO2015060866A1

WO2015060866A1 - Product demand forecasting

Info

Publication number: WO2015060866A1
Application number: PCT/US2013/066859
Authority: WO
Inventors: Sandip MUKHERJEE; Vikash Kumar SHARMA; Anil Kumar KARTHAM; Aparna MRIDUL
Original assignee: Hewlett-Packard Development Company, L.P.
Priority date: 2013-10-25
Filing date: 2013-10-25
Publication date: 2015-04-30

Abstract

Forecasting demand for a new product for a target interval can include determining an adoption factor, a replacement factor, an imitation factor, an innovation factor, and a forecast demand for the new product. The forecast demand for the new product can be based on the adoption factor, the imitation factor, the innovation factor, and an economic variable.

Description

PRODUCT DEMAND FORECASTING

Background

[0001] Product demand generally refers to the amount of a good or service that a consumer is willing and able to buy per unit of time. Demand for a product can be influenced by a host of variab!es. Variables can have distinct and/or

disproportionate effects on the demand for different products.

[0002] Product demand can be a driver of business strategy.

Brief Description of the Drawings

[0003] Figure 1 is a diagram of an example of an environment for generating a new product demand forecast according to the present disclosure.

[0004] Figure 2 illustrates a diagram of an example of a system for forecasting demand for a new product according to the present disclosure.

[0005] Figure 3 illustrates a diagram of an example of a computing device according to the present disclosure.

[0006] Figure 4 illustrates a flow chart of an example of a method for forecasting demand for a new product for a target interval according to the present disclosure.

Detailed Description

[0007] Product demand can shape resource distribution within a business. Models can be used to attempt to predict growth for particular classes of products. Some businesses may employ models to proactively shape resource distribution in an attempt to efficiently meet demand.

[0008] Businesses use estimates of future product demand to plan for activities related to the demand. Based on these estimates, businesses can adjust a host of strategic factors (e.g., pricing, promotion, channel prioritization, risk mitigation, manufacturing, partner choices, sales strategy, training, marketing, financial planning, etc.). Reliable estimates of future demand can be powerful tools in effective business planning.

[0009] Historical data points (e.g., sales, shipments, etc.) related to product offerings can be utilized within a future demand model (e.g., mathematical algorithms, demand curves, etc.) for a product. This historical data can be used as a basis upon which to infer future demand for a particular product.

[0010] New product offerings often lack historical data points as compared to established products. Without sufficient historical data points, models may not be able to construct a reliable estimate of future demand. Application of conventional econometric modeling techniques (e.g. , auto regressive models) with other independent variables to model demand may require a substantial historical database in order to product reliable estimates.

[0011] Some methodologies (e.g., the Bass model, etc.) predict growth of new generic classes of products. The methodologies assume fixed market sizes in predicting initial purchases of new product classes. The methodologies imply patterns of exponential growth of initial purchases followed by exponential decay in sales of new classes of products. The implication of exponential decay may not be applicable to new products or new models of older products (e.g., similar underlying technology of new and existing products). Further, the impact of similar products to a new product within a market place may be absent from the methodologies.

[0012] Economic indicators (e.g., Consumer Confidence Index, Consumer Price Index, Wholesale Price Index, Employee Cost Index, Producer Price Index, Gross Domestic Product, etc.) include statistics about economic activity. Economic indicators can be leading indicators (e.g., indicators that usually change before a market condition changes, such as a decrease in the University of Michigan

Consumer Sentiment Index (MSCI) indicating a softening in the personal computer (PC) and/or printer market in the succeeding fiscal quarters), lagging indicators (e.g. , indicators that usually change after a market condition changes), and coincident indicators (e.g. , indicators that usually change at approximately the same time as an economic condition changes). Indicators can be region specific (e.g., indicative of economic conditions within a specific geography) and/or industry specific (e.g. , indicative of economic conditions with a specific industry). Economic indicators can be predictors of future performance of an economy and a product within that economy. However, the methodologies associated with growth prediction models can fail to account for the effect of economic conditions and can lack the

sophistication to incorporate the predictive value of economic indicators in a reliable fashion.

[0013] In contrast, in accordance with various examples of the present disclosure, a product demand forecast (e.g., forecasted number of a products sold/purchases over an interval) can be generated for a new product (e.g. , a previously non-existent product, a nascent product, a product with short exposure to the market place, a new variant of an existing product, a product with a short existence relative to other products, a new class of products, etc.). The demand forecast can be derived from a number of components (e.g., an adoption component, a replacement component, an innovation component, an imitation component, and an economic variable component as will be discussed further herein). The number of components can include characterizations (e.g., mathematical representations) of the influence of new product adoption, its replacement cycle, behavior of buyers fitting specific classifications (e.g., innovators, imitators, etc.) purchasing this product, and the impact of economic indicators on demand.

[0014] Forecasting product demand, as used herein, can include generating product demand forecasts (e.g., forecasted number of a products sold/purchases over an interval) for a new product. The demand forecasts can be generated from a model (e.g., a description of a system using mathematical concepts and language). Depending on the particular implementation, the model can predict and/or allow prediction of a metric related to demand (e.g., sales, shipments, orders, etc.) over a particular interval (e.g., day, month, calendar year, fiscal year, fiscal quarter, etc.). The prediction can be scalable (e.g. , can estimate demand in various

geographies/sub-regions, estimate demand in various consumer segments, estimate demand in various commercial segments, and/or estimate demand in various product categories). [0015] The methodology of the model can include inputs (e.g., electronic representations of data) to generate an output (e.g., product demand forecast). The inputs can include data points collected from a number of sources. For example, in some embodiments described herein, the data points can be collected internally by an entity utilizing the model (e.g., data points captured by a business regarding its products and/or operations, data points captured by a business regarding the products and/or operations of another business, data points captured by a business regarding economic conditions, etc.) and/or data points collect by a third party (e.g. , data points captured by a third party regarding a number of conditions for a business, market, segment, industry, economy, etc. , third party vendors of data, etc.). The inputs can be adjusted and/or weighted to account for seasonal patterns (e.g., holidays) or trends. For example, an input including a number of shipments of a product in the fourth fiscal quarter could be given less weight than those in the first through third quarters if the fourth quarter included a sales spike attributable to holidays occurring during the quarter.

[0016] The inputs can be mathematically arranged and/or manipulated in a manner that produces representations of an adoption component (e.g. , an adoption factor), a replacement component (e.g., a replacement factor), an innovation component (e.g., an innovation factor), an imitation component (e.g., an imitation factor), and an economic variable component (e.g., an economic variable factor), as will be discussed further herein.

[0017] Figure 1 is a diagram of an example of an environment 00 for generating a product demand forecast 102 according to the present disclosure. The environment 100 can include a number of inputs 104-1 , 104-2, 104-N. The inputs 104-1 , 104-2, ... , 104-N can include any data. The inputs 104-1 , 104-2, 104-N can include electronic representations of data. The inputs 104-1 , 104-2, 104-N can be made available (e.g., as digitally stored electronic representations) for utilization as raw data points (e.g., data represented in its native/unaltered format) and/or as formatted data points (e.g., figures modified and/or enriched from their native format for use in generating the demand forecast 102).

[0018] The inputs 104-1 , 104-2, 104-N can have a number of sources. For example, the inputs 104-1 , 104-2, 104-N can be collected, compiled, stored, and/or made available by a number of entities, including, for example, businesses associated with a particular product and/or third parties. Additionally, the inputs 104- 1 , 104-2, 04-N can respectively be collected, compiled, stored, and/or made available by the same source (e.g. , the same entity, a division of the same entity, an affiliated entity, etc.), a third party source (e.g., an unaffiliated entity, a research firm, etc.), and/or a combination of sources (e.g., the same source and the third party source).

[0019] The inputs 104-1 , 104-2, 104-N can include historical data 104-1. The historical data 104-1 can include information (e.g., records) associated with historical (e.g., having occurred in prior intervals) shipments and/or sales of a product. The information associated with historical shipments of a product can include information associated with shipments of a new product (e.g., a previously non-existent product, a nascent product, a product with short exposure to the market place, a new variant of an existing product, a product with a short existence relative to other products, a new class of products, etc.). The new product can be the product for which a product demand forecast 102 is sought. The new product can additionally or alternatively be a disruptive product (e.g. , a product including disruptive technology capable of disrupting a market for an existing or planned product). A new product can be a product that is new (e.g. , in existence, in market exposure, in utilization, etc.) relative to a similar (e.g., technologically related, functionally related, shared characteristics, etc.) existing product. Historical shipment data 104-1 can include shipment amounts. Shipment amounts can be numbers of units shipped for a particular product and/or products (e.g.., all products) of a particular class (e.g., technical grouping, functional grouping, a general type, shared characteristics, etc.).

[0020] Historical data 04-1 can include interval data including providing time frames associated with historical records. For instance, the number of units shipped can additionally include and/or be associated with the interval data (e.g., a period of time over which the shipments occurred such as a fiscal quarter). Therefore, historical data 104-1 can be a source of information regarding the sales and/or shipments of a new product over an interval of a duration related to an interval of a demand forecast 102,

[0021] Historical data 104-1 can include data indicative of consumers and consumer behaviors with regard to the new product. This data can include a number of consumers in a given market for the new product. Additionally the historical data 104-1 can include how segments of those consumers behave (e.g., timing of purchase, reasons for purchase, time until replacement, etc.) with regard to purchasing the new product. The data can also be consumer specific, with a particular piece of data being associated with a specific consumer (e.g., when a particular customer purchased a particular product). Additionally, the historical data 104-1 can include indications of consumer behaviors (e.g., what drove a consumer and/or group of consumers to make a purchase of a particular product).

[0022] As shown in Figure 1 , inputs (e.g., 104-1 , 104-2, 104-N) can

additionally include, historical/ projection data 104-2. The historical/ projection data 04-2 can be historical/projection data 104-2 related to a number of existing similar products (e.g., technologically related, functionally related, shared characteristics etc.) to the new product. An existing similar product can be determined based on any number of factors (e.g., industry knowledge, technical relationships between the products, functional relationships between the products, shared characteristics between the products, similar customer bases between the products, etc.). In various embodiments of the present disclosure, an existing similar product can be determined by a similarity score exceeding a minimum threshold for an existing product. In such examples, the similarity score can result from a comparison of quantified key characteristics (e.g. , sales models, technology comparison, market comparison, descriptor comparison, etc.) between an existing product and a new product.

[0023] The historical/projection data 104-2 can include information (e.g., records) associated with historical (e.g. , having occurred in prior intervals) shipments and/or sales of the number of similar products. Historical/projection data 104-2 can include shipment amounts. Shipment amounts can be numbers of units shipped for a particular product and/or all products of a particular class (e.g., technical grouping, functional grouping, a general type, etc.).

[0024] Historical/projection data 104-2 can include interval data that provides time frames associated with historical records. For instance, the number of units shipped can additionally include and/or be associated with the interval data (e.g., a period of time over which the shipments occurred such as a fiscal quarter).

Therefore, historical/projection data 104-2 can be a source of information regarding the sales and/or shipments of any number of existing similar products over an interval of a duration related to an interval of a demand forecast 102. [0025] Historical/projection data 104-2 can also include data indicative of interactions (e.g., adoption patterns, etc.) between similar products. For instance, the interactions can include relationships between data points of similar products and how they influence one another (e.g., elasticity derived from historical conversion of products). In this context, the relationship between similar products may include the relationship between products that are both similar to a new product that is the subject of the demand forecast 102 and similar to one another. In various

embodiments, the relationship between similar products may include a proxy model relationship between at least one product that is similar to the existing product competing with the new product and at least another product that is similar to the new product that is the subject of the product demand forecast 102. In this manner, the historical/projection data e.g., 104-2 can include historical data demonstrative of the interaction of two products which are similarly positioned in the market as a new product and its competitors. For instance, the historical/projection data 104-2 for a product demand forecast 102 of a new tablet computer product may include historical data regarding the sales and/or shipments of mobile phones and smart phones as the mobile phone is similar to an existing personal computer (PC) products (representing a less mobile computing platform) and the smart phone is similar to a new tablet computer product (representing a more mobile computing platform).

[0026] Historical/projection data 104-2 can also include data demonstrative of projections related to any number of similar products. The projections can include projected sales, projected shipments, models derived from historical data related to the similar products which provide predictions, etc.

[0027] Historical/projection data 04-2 can include data indicative of consumers and consumer behaviors with regard to any number of similar products. This data can include numbers of consumers in a given market for existing similar products. Additionally the historical/projection data 104-2 can include how segments of those consumers behave (e.g., timing of purchase, reasons for purchase, time until replacement, etc.) with regard to purchasing any number of similar products. The data can also be consumer specific, with a particular piece of data being associated with a specific consumer (e.g., when a particular customer purchased a particular product). Additionally, the historical data 104-1 can include indications of consumer behaviors (e.g. , what drove a consumer and/or group on consumers to make a purchase of a particular product).

[0028] As shown in Figure 1 , inputs 104-1 , 04-2, 04-N can additionally include economic variables 04-N. The economic variables 104-N can include microeconomic and/or macroeconomic data. The data can be related to economic indicators (e.g., the indicators themselves, a function including an economic indicator, a projection derived from an economic indicator, etc.). The economic indicators can be a number of economic indicators derived from any number of sources.

[0029] The environment 100 of Figure 1 includes a data analysis manager 106. The data analysis manager 106 can be deployed on a computing device, (e.g., a computing device as described in connection with Figure 2B) for instance. The data analysis manager 106 can utilize the methodologies and models of the various embodiments described herein to derive a demand forecast 102 from the inputs 104- 1 , 104-2, 104-N.

[0030] The data analysis manager 106 can derive a demand forecast 102 based on a number of components 108-1 , 108-2, 108-3, 108-4, ... ,108-N. Each of the components 108-1 , 108-2, 108-3, 108-4, ... , 108-N can be an element of an equation that characterizes a factor important for generating a demand forecast 02 for a new product. The components 108-1 , 108-2, 108-3, 108-4, ... ,108-N can be functions of data from inputs 104-1 , 104-2, 1 04-N arranged in a mathematical operation providing a factor. The factor can be a value which represents a role of its respective component in a broader mathematical operation describing the demand forecast 102. That is, the components 108-1 , 108-2, 08-3, 108-4 08-N can mathematically describe the role and/or influence of a particular characteristic of the data from inputs 104-1 , 104-2, 104-N in arriving at a reliable estimate of demand for a new product A number of the components 108-1 , 108-2, 108-3, 108-4, ... ,108- N or their constituent elements can be further adjusted (e.g., weighted) by the data analysis manager 106 to account for seasonal fluctuations (e.g., a spike in sales associated with a holiday shopping season).

[0031] As shown in Figure 1 , a component can be an adoption component 108-1. An adoption component 108-1 can include an adoption factor that indicates the portion of potential buyers of new and/or similar existing products who will choose the new product over the similar existing products based on the data from inputs 04-1 , 04-2, ... , 104-N (e.g. , the portion of potential buyers for computing devices who will purchase the new tablet computing product over the existing similar PC product). The adoption factor can indicate a first portion of a plurality of potential buyers of a new product who will purchase the new product over a similar existing product in a targeted interval (e.g., the interval corresponding to an interval of the demand forecast 102).

[0032] The adoption factor can be based on an elasticity of the new product derived from the inputs 04-1 , 104-2, 04-N. For example, the adoption factor can be based on elasticity derived from historical conversion from a proxy model (e.g., historical conversion from similar existing products, such as the similar existing products of the proxy model, which display similar adoption patterns to the new product and a similar and/or competing existing product). The adoption factor for a PC to new tablet computer product demand forecast can, for example, be based on an elasticity coefficient derived from historical conversion from a mobile phone to smartphone adoption pattern as the mobile phone to smartphone demonstrates a similar adoption pattern and moving towards mobility.

[0033] The adoption factor can additionally be based on a quantity of a similar existing product (e.g., a competing similar product) shipped during an interval. For example, the adoption factor for a PC to new tablet computer product demand forecast 102 can, for example, be based on the quantity of PC shipments within the same and/or preceding fiscal quarter(s) as is being forecasted in the product demand forecast 102. The quantity of PC shipments during the interval can be instructive of the amount of customers exposed to similar technology and, therefore, the customer base of the new tablet computer product.

[0034] As shown in Figure 1 , a component can be a replacement component 108-2. A replacement component 108-2 can include a replacement factor that indicates a portion of shipped and/or sold new products (e.g., installed new product base) that will be replaced (e.g., repurchase of the new product) after a certain period of time based on the data from inputs 104-1 , 104-2, 104-N. The

replacement factor can, as a result, be a function of the lifetime of the new product since consumers may repurchase the same product once the original wears out and/or purchase the new product once the similar product wears out. More specifically, the replacement factor can be based on a summation of a number of historical shipments preceding a target interval (e.g., the interval associated with the demand forecast 102).

[0035] The replacement factor can additionally be based on a weight function assigning weight to respective ones of the installed base of purchasers based on a temporal proximity of a corresponding purchase of the new product to the target interval. That is, the replacement factor can include a weight function formulated by considering the fact that the customer who has purchased a product (e.g., the new product) in an earlier time interval in relation to the target interval of the demand forecast 102 has a higher probability to purchase again in the target interval of the demand forecast 102 compared to a customer who has purchased the product in a more recent time interval to the target interval of the demand forecast 02. For example, when utilized in producing a new product demand forecast 02 for the tenth interval, the replacement factor can include a weight function that assigns more weight to an existing customer who made his purchase in the first interval versus one who made her purchase in the ninth interval, in this manner, the replacement factor can include a weight function that assigns weight to purchases and/or shipments in inverse correlation to a temporal proximity of the purchase to the target interval.

[0036] As shown in Figure 1 , a component can be an imitation component 108-3. An imitation component 108-3 can include an imitation factor that indicates a portion of consumers of the new product that adopt (e.g., purchase) a new product after talking to other consumers and reading product reviews. These consumers purchase the new product based on what others are saying about it and/or the behavior and opinions of others. Imitative consumers can be identified and/or classified based on the data from inputs 104-1 , 104-2, 104-N, an understanding of the market, and/or other sources indicative of consumer behaviors.

[0037] The imitation factor can be based on a variable between zero and one derived from historical imitative buyer information. The historical imitative buyer information can include any and all data sources 104-1 , 104-2, 104-N, and/or market understanding which characterize the buying behavior of consumers, particularly data sources which characterize factors influencing the purchasing behavior of a consumer with regard to a particular product (e.g., the new product, an existing similar product, etc.).

[0038] The imitation factor can additionally be based on a cumulative total of shipments and/or purchases of the new product through the time interval preceding the particular time interval. That is, if the product demand forecast 102 is forecasting shipments for a new tablet computing product in the fourth quarter of 20 3, the imitation factor can be based on the total amount of sales of the new tablet computing product up through the third quarter of 2013. The imitation factor may be based on a function including the total number of capable buyer of the new product also. The total number of capable buyers of the new product can be estimated using historical data 04-1 related to the new product and/or historical/projection data 104- 2 related to similar existing products. Furthermore, the imitation factor can be the result of a function including the total number of capable buyers of the new product, the number of units of the new product already sold and/or shipped, and a variable between zero and one indicative of a number of consumers that are imitative consumers.

[0039] As shown in Figure 1 , a component can be an innovation component 108-4. An innovation component 108-4 can include an innovation factor that indicates a portion of consumers of the new product that adopt (e.g., purchase) a new product on their own, independent of the opinions of other consumers. In some instances, these may be consumers that adopt the new product with understanding of a need for the product and/or an appreciation of the technology associated with the new product. Innovative consumers can be identified and/or classified based on the data from inputs 104-1 , 104-2, 04-N, an understanding of the market, and/or other sources indicative of consumer behaviors. Alternatively or additionally, the portion of innovative consumers can be the portion of consumers that have unexplained purchasing behavior. That is, if the sum of all purchases of a product (e.g., the new product) minus the purchases attributable to the adoption component 108-1 , the purchases attributable to the replacement component 108-2, and the purchases attributable to the imitation component 108-3 leave a number of unexplained purchases (e.g., purchases not attributable to any of the three components 108-1 , 08-2, and 08-3, the unexplained purchases can represent the innovative consumers). Moreover, this unexplained portion of purchases can be an unexplained portion of purchases over a particular time interval. In various embodiments, the innovation factor can be forecasted based on a moving average model of the unexplained portion of the purchases.

[0040] As shown in Figure 1 , a component can be an economic variable component 108-N. An economic variable component 108-N can include an economic variable. The economic variable can be associated with a particular interval (e.g. , the same target interval associated with the product demand forecast 102 for the new product). The economic variable can be based on data from inputs 104-1 , 104-2, ... , 04-N. For instance, the economic variable utilized to derive a new tablet computing device product demand forecast 102 for the fourth quarter of 2013 can be an economic indicator and/or a function of an economic indicator, such as a price index measure of inflation (e.g. , Consumer Price Index (CPI), Wholesale Price Index (WPI), etc.) for the third quarter of 2013. The economic variable can be greater than one or less than one depending on whether the economic indicator is increasing or decreasing (e.g., indicating an increase or decrease in an economic condition).

[0041] As described above, the data analysis manager 06 can derive a demand forecast 102 for the new product based on a number of components 108-1 , 108-2, 108-3, 108-4, 108-N and their associated factors. For instance, the data manager 106 can generate an output including a product demand forecast 102 derived from a mathematical operation including an adoption component 108-1 , a replacement component 108-2, an imitation component 108-3, an innovation component 108-4, and an economic variable component 108-N. The new product demand forecast 02 can include forecasted shipments and/or sales for the forecasted interval. Additionally, the new product demand forecast 102 can be segmented by geographical regions, commercial segments, and/or commercial segments.

[0042] In an example, the data analysis manager 106 can derive a new product demand forecast 102 for a new product (e.g., new tablet computer product) based on the following mathematical operation:

Tablet shipment of a time interval = /(Adoption component, Replacement component, Innovation component, Imitation component) x (economic variable component).

More specifically, the data analysis manager 106 can derive a new product demand forecast 102 for a new product (e.g. , new tablet computer product) based on the following mathematical operation:

Sj = W +∑ - Nj) + U . ] X E

Where, Si is the total shipments of the new product in j^th time interval. Where, S , is the i^th interval shipment. Where, ! is the elasticity derived from historical conversion in a proxy model of products that displayed a similar adoption pattern (e.g., mobile phone to smart phone). Where, Pj is the shipments of the similar and/or competing product (e.g. , PCs) in the j^th time interval. Where, Wtj =({j-i)/j)², wherein this is a weight function formulated by considering the fact that a customer who has brought a product in an earlier time interval has more probability to purchase the product again in the current interval as opposed to a customer who has bought the product in a recent time interval. Where, q is an imitation factor which is a positive number between zero and one derived from historical information related to the purchasing behavior of consumers. Where, m is a total number of capable buyers of the product estimated by historical data related to the new roduct and/or historical/projection data related to similar products. Where, ^tnis represents cumulative

shipments until (j-1 )^th time interval. Where, U_., is an unexplained portion of the sales/shipments (e.g., the portion attributable to innovative consumers). It can be forecasted based on a moving average method factoring the portion of historical shipments and/or sales not attributable to other components (e.g., adoption component 108-1 , replacement component 108-2, and imitation component 108-3). Where, E ; is an economic variable (e.g. , CPI, WPI, etc.) in the j^th time interval.

[0043] Figures 2 and 3 illustrate examples of systems 230, 350 according to the present disclosure. Figure 2 illustrates a diagram of an example of a system 230 for forecasting demand for a new product according to the present disclosure. The system 230 can include a data store 232, a management system 234, and/or a number of engines 236, 238, 240, 242, 244. The management system 234 can be in communication with the data store 232 via a communication link, and can include the number of engines (e.g., adoption engine 236, replacement engine 238, imitation engine 240, innovation engine 242, forecast engine 244, etc.). The management system 234 can include additional or fewer engines than illustrated to perform the various functions described herein.

[0044] The number of engines can include a combination of hardware and programming that is configured to perform a number of functions described herein (e.g., determining an adoption factor). The programming can include program instructions (e.g., software, firmware, etc.) stored in a memory resource (e.g., computer readable medium, machine readable medium, etc.) as well as hard-wired program (e.g., logic).

[0045] The adoption engine 234 can include hardware and/or a combination of hardware and programming to determine an adoption factor. The adoption factor can include an electronic representation of a first portion of a plurality of potential buyers of a new product who will purchase the new product over a similar existing product. The adoption factor can include an electronic representation of a first portion of a plurality of potential buyers of a new product who will purchase the new product over a similar existing product in a target interval. The first portion of the plurality of potential buyers of a new product who will purchase the new product over a similar existing product can be derived from a historical conversion rate among a group of products displaying a similar adoption pattern (e.g., a proxy model). For example, the adoption factor can be based on an elasticity derived from historical conversion among a number of products such as a mobile phone to smart phone conversion as a proxy model for a PC to table computing device conversion. The adoption factor can be further based on a quantity of the similar existing product shipped over a particular interval. For example, the quantity of the similar existing product shipped over the target interval.

[0046] The replacement engine 238 can include hardware and/or a

combination of hardware and programming to determine a replacement factor. The replacement factor can include an electronic representation of a number of prior purchases of the new product that will generate a repurchase of the new product. That is, the replacement factor can include a portion of the new products already purchased that will be replaced with a repurchase of the new product. Furthermore, the replacement factor can be based on a summation of a number of historical shipments of the new product and/or a similar existing product preceding the target interval of the demand forecast. Additionally or alternatively, the replacement factor can be based on a weight function that assigns weight to respective ones of the number of prior purchasers of the new product and/or the similar existing product based on a temporal proximity of a corresponding purchase of the product to the target interval. The weight can be assigned in an inverse correlation with the temporal proximity to the target interval. [0047] The imitation engine 240 can include hardware and/or a combination of hardware and programming to determine an imitation factor. The imitation factor can include an electronic representation of a second portion of the plurality of potential buyers of a new product who will purchase the new product based on a behavior of a third party. Furthermore, the imitation factor can include an electronic representation of a second portion of the plurality of potential buyers of a new product who will purchase the new product based on an opinion of a third party. The plurality of potential buyers can be based on historical data sets related to the number of capable buyers for the new product and/or historical data sets related to the number of capable buyers for a similar existing product. The imitation factor can be based on a random variable between zero and one derived from historical imitative buyer information and/or a cumulative total of shipments and/or purchases of the new product through the time interval preceding the target interval.

[0048] The innovation engine 242 can include hardware and/or a combination of hardware and programming to determine an innovation factor. The innovation factor can include an electronic representation of a third portion of the plurality of potential buyers who will purchase the new product independent of the behavior of a third party. The innovation factor can also include an electronic representation of a third portion of the plurality of potential buyers who will purchase the new product independent of the opinion of a third party. The innovation factor can be based on a moving average of unexplained historical shipments of the new product and/or a similar existing product. The unexplained historical shipments can include a portion of a sum total of historical shipments preceding the target interval that are not attributable to the adoption factor, the replacement factor, and/or the imitation factor.

[0049] The forecast engine 244 can include hardware and/or a combination of hardware and programming to forecast demand for a new product based on the adoption factor, the replacement factor, the innovation factor, and an electronic representation of an economic variable. In various embodiments of the present disclosure, the forecast engine 244 can include hardware and/or a combination of hardware and programming to forecast a sales model of a new product. The sales model can include an electronic representation of an amount of shipments and/or purchases of the new product over a target interval based on a function of the adoption factor, the replacement factor, the imitation factor, the innovation factor, and a function of a number of economic variables. Economic variables can be region specific and/or product specific (e.g., uniquely adapted to an economic measure including the new product and/or its product class, a variable which the new product and/or its product class are especially sensitive to and/or for which strong correlations exist, etc.). The economic variables can include economic measures (e.g., a leading measure of an economic metric, a lagging measure of an economic metric, a coincident measure of an economic metric) over the course of the target interval. In various embodiments the economic variables can include microeconomic and macroeconomic variables. The macroeconomic variables can be based on a price index measure of inflation (e.g. , CP!, WPI, etc.). When utilizing leading economic variables, the economic variables can include economic measures over intervals preceding the target interval.

[0050] The forecast engine 244 can forecast demand by applying the inputs and resulting factors within a mathematical operation, for example one or more of those discussed in relation to Figure 1 , to generate forecasted demand as an output. In various embodiment of the present disclosure, the forecast engine 244 can forecast demand by multiplying an electronic representation of a sales forecast, wherein a sales forecast is a function of an adoption factor, a replacement factor, an imitation factor, and an innovation factor, by an electronic representation of a function of an economic variable.

[0051] The forecast engine 244 can additionally include hardware and/or a combination of hardware and programming to provide any adjustments to any input, component, factor, and/or output to account for irregularities (e.g., spikes in shipments) due to unique interval specific conditions (e.g., holiday seasons).

[0052] Figure 3 illustrates a diagram of an example of a computing device 350 according to the present disclosure. The computing device 350 can utilize software, hardware, firmware, and/or logic to perform a number of functions herein.

[0053] The computing device 350 can be any combination of hardware and program instructions configured to share information. The hardware for example can include a processing resource 352 and/or a memory resource 356 (e.g., computer- readable medium (CRM), machine readable medium (MRM), database, etc.) A processing resource 352, as used herein, can include any number of processors capable of executing instructions stored by a memory resource 356. Processing resource 352 may be integrated in a single device or distributed across multiple devices. The program instructions (e.g., computer-readable instructions (CRI)) can include instructions stored on the memory resource 356 and executable by the processing resource 352 to implement a desired function (e.g., determine an adoption factor).

[0054] The memory resource 356 can be in communication with a processing resource 352. A memory resource 356, as used herein, can include any number of memory components capable of storing instructions that can be executed by processing resource 352. Such memory resource 356 can be a non-transitory CRM or MRM. Memory resource 356 may be integrated in a single device or distributed across multiple devices. Further, memory resource 356 may be fully or partially integrated in the same device as processing resource 352 or it may be separate but accessible to that device and processing resource 352. Thus, it is noted that the computing device 350 may be implemented on a participant device, on a server device, on a collection of server devices, and/or a combination of the user device and the server device.

[0055] The memory resource 356 can be in communication with the

processing resource 352 via a communication link (e.g., a path) 354. The

communication link 354 can be local or remote to a machine (e.g., a computing device) associated with the processing resource 352. Examples of a local communication link 354 can include an electronic bus internal to a machine (e.g. , a computing device) where the memory resource 356 is one of volatile, non-volatile, fixed, and/or removable storage medium in communication with the processing resource 352 via the electronic bus.

[0056] A number of modules 358, 360, 362, 364, 366 can include CRI that when executed by the processing resource 352 can perform a number of functions. The number of modules 358, 360, 362, 364, 366 can be sub-modules of other modules. For example, the provisioning module 358 and the request module 360 can be sub-modules and/or contained within the same computing device. In another example, the number of modules 358, 360, 362, 364, 366 can comprise individual modules at separate and distinct locations (e.g., CRM, etc.).

[0057] Each of the number of modules 358, 360, 362, 364, 366 can include instructions that when executed by the processing resource 352 can function as a corresponding engine as described herein. For example, the adoption module 358 can include instructions that when executed by the processing resource 352 can function as the adoption engine 336. In another example, economic module 366 can 6859

18

include instructions that when executed by the processing resource 354 can function as the forecast engine 344.

[0058] Figure 4 illustrates a flow chart of an example of a method 470 for forecasting demand for a new product for a target interval. At 472, the method 470 can include a computing device developing a sales forecast for the new product for the target interval. In some examples, the sales forecast is a function of an adoption factor 476, a replacement factor 478, an imitation factor 480, and an innovation factor 482. The adoption factor 476 can include an electronic representation of a first proportion of a plurality of potential buyers of a new product who will purchase the new product over a similar existing product. The replacement factor 478 can include an electronic representation of a proportion of a number of the new product already purchased that will be replaced with a repurchase of the new product. The imitation factor 480 can include an electronic representation of a second proportion of the plurality of potential buyers comprising imitative buyers who base a purchase of the new product on an opinion of a third party. The innovation factor 482 can include an electronic representation of a third proportion of the plurality of potential buyers comprising innovative buyers who will purchase the new product independent of the opinion of the third party. In various embodiment of the present disclosure, the innovation factor 482 can be based on a moving average of unexplained historical shipments, wherein the unexplained historical shipments include a proportion of a sum total of historical shipments preceding the target interval not attributable to the adoption factor 476, the replacement factor 478, and the imitation factor 480.

[0059] At 474, the method 470 can include determining an impact of a macroeconomic indicator on the sales forecast for the target interval. The impact can be derived by, for example, multiplying an electronic representation of the sales forecast by an electronic representation of a function of a macroeconomic variable for the target interval. The macroeconomic variable can, for example, be based on a price index measure of inflation. Furthermore, all of the factors, the resulting sales forecast, the economic variables, and the demand forecast can be adjusted to season irregularities (e.g., adjusting data from particular fiscal quarters to factor for spikes in shipments/sales due to a holiday season falling within the fiscal quarter). For example, the target sales forecast can be adjusted to account for historical seasonality corresponding to the target interval such as by revising the product demand forecast for the fourth fiscal quarter upward to account for historical spikes in sales of the new product and/or similar existing products during a holiday in the fourth fiscal quarter.

[0060] In the detailed description of the present disclosure, reference is made to the accompanying drawings that form a part hereof, and in which is shown by way of illustration how examples of the disclosure may be practiced. These examples are described in sufficient detail to enable those of ordinary skill in the art to practice the examples of this disclosure, and it is to be understood that other examples may be used and the process, electrical, and/or structural changes may be made without departing from the scope of the present disclosure.

[0061] In addition, the proportion and the relative scale of the elements provided in the figures are intended to illustrate the examples of the present disclosure, and should not be taken in a limiting sense. As used herein, the designator "N", particularly with respect to reference numerals in the drawings, indicate that a number of the particular feature so designated can be included with a number of examples of the present disclosure. As used herein, "a" or "a number of something can refer to one or more such things.

Claims

What is claimed:

1. A non-transitory computer readable medium storing instructions executable by a processing resource to:

determine an adoption factor including an electronic representation of a first portion of a plurality of potential buyers of a new product who will purchase the new product over a similar existing product;

determine a replacement factor including an electronic representation of a number of prior purchases of the new product that will generate a repurchase of the new product;

determine an imitation factor including an electronic representation of a second proportion of the plurality of potential buyers who will purchase the new product based on a behavior of a third party;

determine an innovation factor including an electronic representation of a third proportion of the plurality of potential buyers who will purchase the new product independent of the behavior of the third party; and

forecast demand for the new product based on the adoption factor, the replacement factor, the imitation factor, the innovation factor, and an electronic representation of an economic variable.

2. The medium of claim 1 , wherein the forecast includes a quantity of the new product forecasted to be shipped over a target interval.

3. The medium of claim 1 , wherein the instructions executable to determine the adoption factor include instructions executable to determine:

an elasticity derived from a historical conversion among a number of products within a proxy model, and

a quantity of the similar existing product shipped over the target interval.

4. The medium of claim 3, wherein the proxy model includes a model of a number of products demonstrating a similar adoption pattern to an adoption pattern of the similar existing product to the new product.

5. The medium of claim 4, wherein the instructions executable to determine the replacement factor include instructions executable to determine: a summation of a number of historical shipments preceding the target interval; and

a weight function assigning weight to respective ones of the number of prior purchasers based on a temporal proximity of a corresponding purchase of the new product to the target interval, wherein the weight assigned is inversely correlated with the temporal proximity to the target interval.

6. The medium of claim 1 , wherein the economic variable includes an economic measure indicative of an economic condition during the target interval.

7. The medium of claim 6, wherein the economic variable includes a region specific economic measure.

8. A system for forecasting product demand, comprising:

an adoption engine to determine an adoption factor, wherein the adoption factor includes an electronic representation of a first proportion of a plurality of potential buyers of a new product who will purchase the new product over a similar existing product in the target interval based on a historical conversion rate among a group of products displaying a similar adoption pattern;

a replacement engine to determine a replacement factor, wherein the replacement factor includes an electronic representation of a proportion of the new products already purchased that will be replaced with a repurchase of the new product, wherein the replacement factor is weighted based on a purchase time interval and a lifetime of the new product;

an imitation engine to determine an imitation factor, wherein the imitation factor includes an electronic representation of a second proportion of the plurality of potential buyers comprising imitative buyers who base a purchase of the new product on an opinion of a third party;

an innovation engine to determine an innovation factor, wherein the

innovation factor includes an electronic representation of a third proportion of the plurality of potential buyers comprising innovative buyers who will purchase the new product independent of the opinion of the third party; and

a forecast engine to forecast a sales model of the new product, wherein the sales model includes an electronic representation of an amount of shipments of the new product over the target interval based on a function of the adoption factor, the replacement factor, the imitation factor, the innovation factor, and a function of a number of economic variables.

9. The system of claim 8, wherein the imitation factor is based on:

a variable between zero and one derived from historical imitative buyer information; and

a cumulative total of shipments of the new product through the time interval preceding the target interval.

10. The system of claim 9, wherein the plurality of potential buyers are based on a historical capable buyer data set for the new product.

11 . The system of claim 9, wherein the plurality of potential buyers are based on a similar product capable buyer data set.

12. A method for forecasting product demand comprising:

a computing device developing a sales forecast for a new product for a target interval, wherein the sales forecast is a function of:

an adoption factor including an electronic representation of a first proportion of a plurality of potential buyers of a new product who will purchase the new product over a similar existing product;

a replacement factor including an electronic representation of a proportion of a number of the new product already purchased that will be replaced with a repurchase of the new product;

an imitation factor including an electronic representation of a second proportion of the plurality of potential buyers comprising imitative buyers who base a purchase of the new product on an opinion of a third party; and

an innovation factor including an electronic representation of a third proportion of the plurality of potential buyers comprising innovative buyers who will purchase the new product independent of the opinion of the third party; and

determining an impact of a macroeconomic indicator on the product sales forecast for the target interval.

13. The method of claim 12, wherein the innovation factor is based on a moving average of unexplained historical shipments, wherein the unexplained historical shipments include a proportion of a sum total of historical shipments preceding the target interval not attributable to the adoption factor, the replacement factor, and the imitation factor.

14. The method of claim 12, wherein the macroeconomic indicator is based on a price index measure of inflation.

15. The method of claim 12, wherein the sales forecast is the sum of the adoption factor, the replacement factor, the imitation factor, and the innovation factor, and wherein determining the impact of a macroeconomic indicator on the product sales forecast for the target interval is achieved by multiplying the sales forecast by the macroeconomic indicator.