This disclosure relates generally to product market research and, more particularly, to methods and apparatus to model consumer awareness for changing products in a consumer purchase model.
Market researchers attempt to advance product acceptance, product popularity, and/or product sales through any number of activities and/or changes to the product. Activities to improve product sales volume include running one or more advertising campaigns and/or running one or more product promotional activities. Changes which may be made to a product include altering product packaging, altering product trade dress, and/or altering one or more features of the product of interest (e.g., improving diaper absorbency, improving cleaning power, etc.). Products which have been subjected to such activities and/or changes are referred to as a restage product.
In attempts to determine whether one or more activities and/or one or more changes to the product improve sales, market researchers may compare sales volumes before the changes to sales volumes after the changes. For example, in the event that sales for a particular market geography and/or demographic increase after one or more advertising campaigns are executed, the market researchers may attribute such increases to the campaigns. However, information indicative of increased or decreased sales after a change to the product of interest may not identify which restage attributes are responsible for such sales changes.
In the interest of brevity and clarity, throughout the following disclosure, references will be made to the example preference modeling and consumer purchase simulation system 100 of
Market researchers, product promoters, marketing employees, agents, and/or other people and/or organizations chartered with the responsibility of product management (hereinafter collectively referred to as “sales forecasters”) typically attempt to justify marketing decisions based on one or more techniques likely to result in increased sales of a product of interest. Often, sales forecasting is an important step in the evaluation of potential product initiatives, and a key qualification factor for the decision to launch in-market. As such, accurate forecasting models are important to facilitate these decisions. One specific type of initiative that adds an extra layer of complexity compared to a new product or line extension is a restage initiative. A restage initiative replaces an existing product or group of products with a modified form of the product. Examples of modifications include, but are not limited to new product formulation(s), new packaging, new sales messaging, etc. Simulating restage initiatives typically requires modeling both the consumer response to the intrinsic product change, and also the rate at which consumers become aware and digest the change that has occurred to the product.
In a restage situation, an original product or group of products undergoes one or more changes in its intrinsic attributes to become the new (restaged) product. As such, consumers have preferences (utilities) for the original product and/or separate preferences for the restage change(s). Simulations of consumer acceptance of restage initiatives that do not address this shift in preferences in sufficient detail run the risk of overestimating or underestimating the impact of a restage change.
The methods and apparatus described herein include, in part, rules to model brand and/or product restage introductions to a market, thereby capturing more accurate information related to the product adoption. The methods and apparatus described herein address phenomena and/or one or more patterns associated with product and restage awareness after the restage product has been introduced into the market. Each consumer awareness state, if known, allows the sales forecaster to identify whether the purchase is likely to be made in view of known restage attributes (e.g., a style change, product packaging changes, feature changes/improvements, etc.), or whether the restage product will likely be purchased for one or more alternate reasons. For example, despite the fact that a restage product is available to the consumer (e.g., on a store shelf), merely purchasing the restage product does not necessarily indicate that the consumer is reacting to the restage attribute(s). Instead, the consumer may simply be accustomed to a particular brand and/or trademark, but have no knowledge that the purchased product has undergone a restage. Distinguishing between consumer awareness states allows the sales forecaster to model consumer behavior by applying utilities associated with either the original product or the restage, which further illustrates one or more reasons (e.g., one or more attributes associated with the purchased product) the consumer would deem relevant to their purchasing decision. As such, the sales forecaster may learn which attributes to, for example, greater emphasize and/or highlight during subsequent advertising efforts and/or to identify which attributes should be included in the restage product at the time it is released in the market for purchase.
Example methods and apparatus to model consumer awareness for changing products in a consumer purchase model are disclosed. A disclosed example method includes receiving utility values associated with at least one of a product or a product attribute, and identifying an agent awareness state associated with the restage product and the original product. The example method also includes calculating a choice probability for the restage product based on the estimated utility values associated with the identified awareness state, and outputting the choice probability for use in a simulation of consumer purchase.
A disclosed example apparatus includes a utility estimator to estimate utility values associated with an original product and a restage change, an awareness manager to identify a respondent awareness state of a plurality of agents associated with the restage product and the original product, and a relative probability calculator to calculate a choice probability value for the restage product and the original product based on the calculated utility values associated with the respondent awareness state.
In operation, the example simulation framework manager 202 initiates each of the example agent manager 204, the example consumer purchase simulator 206, the example awareness manager 208, and the relative probability calculator 210. Additionally, the example simulation framework manager 202 identifies and/or applies one or more shopping rules for agents, selects one or more products to be used in one or more simulations, and/or adjusts one or more attributes of a selected product during the one or more simulations, thereby allowing the sales forecaster to gain further insight related to the consumer adoption process before a product is actually released into the public market.
The example simulation framework manager 202 invokes the example agent manager 204 to retrieve and/or otherwise receive estimated utility values from the example utility estimator 106, which are derived from the example discrete choice exercise engine 104. A utility (a relative preference) may be estimated for one or more attributes. Attributes include, but are not limited to price, size, product feature, quantity, etc. Each attribute may further have one or more ranges (e.g., a price between $1.25 and $3.25). To estimate the one or more utility values (also referred to herein as “utilities”), the example utility estimator 106 employs a classification model, such as an example hierarchical Bayes estimation. The example hierarchical Bayes estimation estimates at a level of resolution related to the respondent rather than a more generalized population level, but any other technique to estimate utilities may be employed. As such, respondent-level estimation provides insight to heterogeneity of preferences among the population.
In the event that such utilities are estimated based on observed panelist behavior, the estimated utilities may be projected to a larger audience in a manner that comports with statistical confidence. The example agent manager initializes one or more groups of agents, which are projected from the respondents in the example human respondent pool 102, to represent simulated consumers so that each agent is associated with at least one set of utility values. For example, if the example utility estimator 106 includes utility values from 500 human respondents, then the example agent manager 204 may project a set of 50,000 agents to participate in one or more consumer purchase simulation(s), in which each agent carries one of the utility sets associated with one of the human respondents.
One or more product consideration sets, which may include original products and/or restage products, are selected by the example consumer purchase simulator 206. Available products capable of purchase by an agent during a simulated consumer purchase are arranged in one or more sets. While any product utility value calculated by the utility estimator 106 with a pattern model, (e.g., the hierarchical Bayes estimation pattern model) may identify an absolute utility value, such utility values provide insight regarding a likelihood of preference to an agent only when compared with other available products in a set to calculate a choice probability. Additionally, and as discussed in further detail below, a multinomial logit model may be used to produce one or more probabilities based on utility input(s). For example, given a set of products A, B, and C (each having its own utility value), a corresponding choice probability can be calculated in a manner that directly considers the other products within the set. In the set of A, B, and C, product A may be preferred 2:1 over product C. However, an alternate set of products A, D, and C will each have different choice probabilities by virtue of the makeup of other products available in the set. For example, while product A may be the favorite product in the set of A, B, C with a choice probability of 70%, within the context of the A, D, C set, it may instead have a choice probability of 5%. Additionally, while the aforementioned example utilities are described in view of the product and/or restage product as a whole, one or more utility values may be employed that are specific to a specific attribute of the original product and/or restage product.
In the illustrated example of
In the illustrated example of
In the illustrated example of
For circumstances in which the agent has not used the original product, but is aware of the original product 306, and has not used the restage product, and is not aware of the restage product 310, then the respondent is deemed to have a state of pre-use-original 318, in which only the utility values associated with the original product pre-use are used when analyzing and/or determining which attributes may be relevant during the purchasing decision(s) of the respondent. Such circumstances may occur when a consumer is brand-loyal and/or responsive to a familiar trademark, product packaging design, and/or trade-dress and purchases the restage product without knowledge of one or more new and/or alternate product attributes. However, for circumstances in which the agent has not used, but is aware of the original product 306, and has not used, but is aware of the restage product 312, then that respondent is deemed to have a state of pre-use-aware-original and pre-use-aware-restage 320 because utility values associated with both the original product pre-use and the restage product pre-use may be relevant to the purchasing decisions made by the consumer.
For circumstances in which the respondent is aware of the original product from prior use 308, but not aware of the restage 310, the respondent is deemed to have a state of post-use-original 322, in which only the utility values associated with the original product post-use are deemed to contribute to the respondent's purchasing decision. However, if the respondent becomes aware of the restage 312 from, for example, advertising activity, then the respondent is deemed to have a state of post-use-original and pre-use-restage 324, in which utility values from both the original product post-use and the restage product pre-use may be relevant to the purchasing decisions made by the consumer. Finally, if the respondent has used both the original product 308 and the restage product 314, then the respondent is deemed to have a state of post-use-original and post-use-restage 326, in which both the original product post-use and restage product post-use utility values may be considered as an influence to the consumer's purchasing decision.
To calculate the choice probabilities, which illustrates relational information of sets of available products, the example relative probability calculator 210 employs a probability model, such as a multinomial logit model. Any number of consideration sets may be computed by the example relative probability calculator 210 to generate choice probabilities within each set. Without limitation, choice probability values calculated by the example relative probability calculator 210 may also be affected by distribution metrics (actual and/or simulated distribution values) and/or awareness states of agents.
Product and restage utility values are used as inputs for the example simulation framework manager 202 to identify, in part, emergent patterns from one or more interactions of the agents during the consumer purchase simulation. The sales forecaster may select any number of attributes believed to be relevant to the product or restage including, but not limited to price, product features, and/or a date on which to introduce the restage into the market. The selected attributes and/or product sets are received by the example simulation framework manager 202 to simulate and model actions and interactions of the agents and observe the results of the virtual shopping trips that have occurred. For example, the simulation framework manager 202 may employ an agent based model (ABM) in which each agent is modeled as an autonomous decision-making entity. Taken together, the ABM identifies emergent patterns of the agents based on their individual choices.
In the illustrated example of
While the example preference modeling and consumer purchase simulation system 100 has been illustrated in
The example instructions 400 of
The example utility estimator 106 (block 406) receives inputs that contain details of the choice tasks presented to respondents including attribute composition of the alternatives, prices for each alternative, and any other variable being considered for inclusion in the choice model. Additionally, the example utility estimator 106 may analyze the conditions of the choice tasks combined with the respondents' choice data, and produce utility value estimates that best fit the respondent provided choice data. As described above, a best fit may be estimated via execution of the hierarchical Bayes estimation technique(s), but is not limited thereto.
The example agent manager 204 receives the estimated utility values from the utility estimator 106 as inputs and projects such utility values into one or more agent sets (block 408) to be used in one or more consumer purchase simulation(s). To account for one or more shopping circumstances that a consumer may experience in the market, the example consumer purchase simulator 206 generates one or more virtual purchase consideration sets that the agents may experience during the purchase simulation(s) (block 410). Simulation conditions may include, but are not limited to price changes, availability of the products, availability of restage products, promotional elements (e.g., the presence of coupons, in-store displays, etc.) and/or one or more alternate product sets.
When a restage is released into the market (i.e. achieves distribution and is offered for sale), awareness of the restage and its associated attributes may not be immediately known to all consumers. Disparity of awareness from one consumer to another consumer may be due to several reasons including, but not limited to, advertising activity (e.g., some geographic regions may spend more/less on advertising than other geographic regions), promotional activity (e.g., in-store displays, in-store announcements, coupons, etc.), and/or restage product presence differences due to different distribution condition(s) and/or lag-time in distribution within, for example, markets at a greater distance of a distribution center versus markets at a closer distance to the distribution center. As described above, the original product includes an associated utility value that is composed of one or more attributes unique to the product. Similarly, each restage product has an associated utility value that is the result of its unique one or more attributes. If estimations, trends, and/or predictions occur without considering whether or not the consumer is aware of the original product or the restage product, then one or more resulting estimations, trends, and/or predictions may overestimate and/or underestimate the effect of impressions on the consumer. The example awareness manager 208 identifies an awareness state of each agent in view of the original product and/or restage product (block 412). As described in further detail below, one or more characteristic models are employed to calculate choice probabilities and reveal emergent behaviors based on the awareness state that, in part, minimize and/or eliminate overestimation and/or underestimation of impressions (e.g., advertising impressions).
While the illustrated example determination of awareness state (block 412) of
If the selected agent is aware of the original product (block 504), then the example awareness manager 208 continues to determine whether the selected agent is also aware of the restage (block 514). If not, then the example awareness manager 208 associates the agent as having an awareness state related only to the original product of interest (block 516) so that only the utility values associated with the original product of interest are applied after during the shopping simulation(s). The utility values associated with the original product of interest are applied even if, during the shopping simulation(s), the agent is actually considering restage product. This models a situation in which a consumer may purchase the restage product while being in a state of ignorance that a change has occurred to the original product and/or its associated attributes (e.g., feature improvement, trade dress changes, packaging changes, etc.). For example, the agent may purchase the restage product because that agent is primarily familiar with a logo, trademark, shape, and/or trade dress of the original product from which the restage is derived.
If the agent is aware of the original product of interest (block 504), and also aware of the restage product (block 514), then the selected agent is associated with an awareness state associated with both the original product of interest and the restage product (block 518). As a result, utility values from both the original product of interest and the restage product will be applied during subsequent modeling activities. In the event that there are no additional agents for which an awareness state is to be determined (block 510), control returns to the example process 400 of
In the illustrated example process 400 of
Awareness states for each agent in view of each restage are provided to the example simulation framework manager 202 to simulate agent purchase conditions (block 416), which may include requesting consideration of the original product utility, applying a combination of both the original and restage utility values, or applying no utility value at all. As described above, the example awareness rules in connection with
The processor platform P100 of the example of
The processor P105 is in communication with the main memory (including a ROM P120 and/or the RAM P115) via a bus P125. The RAM P115 may be implemented by dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), and/or any other type of RAM device, and ROM may be implemented by flash memory and/or any other desired type of memory device. Access to the memory P115 and the memory P120 may be controlled by a memory controller (not shown).
The processor platform P100 also includes an interface circuit P130. The interface circuit P130 may be implemented by any type of interface standard, such as an external memory interface, serial port, general-purpose input/output, etc. One or more input devices P135 and one or more output devices P140 are connected to the interface circuit P130.
Although certain example methods, apparatus and articles of manufacture have been described herein, the scope of coverage of this patent is not limited thereto. On the contrary, this patent covers all methods, apparatus and articles of manufacture fairly falling within the scope of the appended claims either literally or under the doctrine of equivalents.