WO2022140384A1 - Value exchange model for customer goals-to-business growth analysis - Google Patents

Abstract

System and methods for quantifying brand value of products to customers in a market leveraging customer-centric parameters, such as Customer Performance Indicators (CPIs) to input into a model that predicts Future Customer Value (FCV). The quantified FCV is a composite score that translates to evaluating key performance indicators of commercial entities selling the products. The FCV can be customized by adjusting one or more CPIs in the model to develop the ideal single or combination of CPIs that most affect brand value of the products or brand. Customized FCV can be used to determine the value delivered to customers, recommend adjustment to KPIs, and predict future customer behavior that drive the optimal lift of quantifiable brand revenue, or performance. Finding the optimal lift, based on quantifiable customer-centric parameters, produces the best fit outcome for both brands selling the products and for consumers purchasing the products.

Description

VALUE EXCHANGE MODEL FOR CUSTOMER GOALS-TO-BUSINESS GROWTH

ANALYSIS

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority and benefit from the U.S. Provisional Patent Application 63/128737, filed December 21, 2020 and titled, “CPIs + THE VALUE EXCHANGE MODEL, which is incorporated herein by reference in its entirety for all purposes.

BACKGROUND

[0002] Nearly all companies rely on Key Performance Indicators (KPIs) to manage and steer their businesses. Typical KPIs revolve around important business outcomes, such as revenue, revenue growth, profitability, and the drivers thereof from an organizational point of view. Sometimes, these drivers may often include "customer-focused" KPIs such as the new customer acquisition rate, customer retention rate, measures of customer spend, and a myriad of customer sentiments. However, the vast majority of KPIs measured and tracked by most companies represent aspects of their business that the average customer would likely care very little about.

[0003] For decades, organizational teams have been established - e.g., consumer insights, customer experience, customer service - and millions of dollars have been spent on improving business operations through the lens of the customer. However, most teams lack mechanisms to operationalize customer centricity like they do KPIs - essentially, most teams rely on consumer or customer parameters or indicators in an abstract form based on company-assumed consumer perceptions or company predictions based on past company performance about the offered products or services. Even with this gap in available customer or consumer data, most organizations claim to be customer centric — their vision statements evidence this as such. Organizations still exist in a business climate in which no entity has quantifiably linked customer goals to business outcomes. This lack of quantitative link of business performance to customer goals with the offered products and services results in sub-optimal customer experiences and loyalty and sub-optimal business results.

[0004] A select group of companies evaluate their financial outcomes or brand value against customer metrics such as consumer satisfaction (CSAT), a net promoter score (NPS), and customer effort score (CES). These customer metrics evaluate the customer experience from a company- side perspective of whether the product or service satisfied the customer when the customer interacted with the company. These conventionally used customer metrics evaluate customer satisfaction from a company perspective, similar to gaining feedback on the quality of a product in a focus group. These customer satisfaction metrics consider the commercial relationship between the customer and the products or services from the company perspective and fail to evaluate the commercial relationship from the customer perspective.

[0005] These drawbacks limit the use and effectiveness of existing models to evaluate commercial relationships between customers and companies. Therefore, the industry could benefit from improved systems and methods to evaluate the commercial relationship between businesses and their customers that identify, evaluate, and predict customer behaviors and experiences, as well as future revenue growth potential.

SUMMARY

[0006] Disclosed methods and systems quantifying customer-centric metrics of the value delivered to customers by companies as a result of commercial transactions. These customercentric parameters or Customer Performance Indicators (CPIs) each relate to evaluating the commercial transaction between a customer and a brand from the perspective of the customer, then quantifying that metric to extract the value received by the customer in the transaction. Because the selected CPIs are customer-centric - quantifiable data relating to the customer perspective of the brand - these CPIs can be used to model Future Customer Value (FCV). Quantifiable data means data that can have an absolute or relative value associated with it. That quantifiable value can be represented in any suitable value and can include qualitative data as well or can itself be a value that reflect qualitative data around a consumer-centric metric. FCV is a metric that predicts how much value brands can expect to extract from customers into the future. The quantified FCV is a composite score that can then be translated or correlated to various key performance indicators of the commercial entities selling the products. The predicted effects on FCV can be customized by adjusting one or more selected CPIs in the model to develop the ideal single or combination of CPIs that most affect brand. Those custom FCVs are used to determine the value brands extract from the commercial relationship with the customers, recommend adjustments to KPIs based on the customer-extracted value, and predict future customer behavior that drives the optimal lift of quantifiable brand performance. Finding the optimal lift, based on quantifiable customer-centric parameters, produces the best fit outcome for both brands selling the products and for consumers purchasing the products.

BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Non-limiting and non-exhaustive embodiments of the invention are described with reference to the following drawings. In the drawings, like reference numerals refer to like parts throughout the various figures, unless otherwise specified, wherein:

[0008] FIG. 1 is an example process flow diagram showing value exchange of a customer’s commercial experience with a brand.

[0009] FIG. 2 is an example model showing a functional relationship between Customer Performance Indicators (CPIs), Future Customer Value (FCV), and key performance indicators (KPIs).

[00010] FIG. 3 is an example process flow diagram showing how CPIs and FCV are used to provide predictive, prescriptive, and diagnostic data.

[00011] FIG. 4 is an example plot of a value exchange curve.

DETAILED DESCRIPTION

[00012] The subject matter of embodiments disclosed herein is described with specificity to meet statutory requirements and to convey the scope of the subject matter to those skilled in the art, but this description does not limit the scope of the claims. The disclosed subject matter may be embodied in other ways, may include different elements or steps in the same or a different order, and may be used in conjunction with other existing or future technologies.

[00013] The disclosed methods and systems are able to quantifiably model the value exchange between customers and corporations based on the premise that when corporations deliver products and services based on CPIs, revenue growth occurs and both parties benefit. The disclosed methods and systems are different from conventional approaches that base company growth or change on past events, single transaction feedback on product performance, and customer satisfaction with the product by instead focusing on quantifiable customer-, or more human-centric, data relating to the overall importance of the product or brand to the customer and the performance of the product or brand to the customer. [00014] The overall importance of the product to the customer is a brand-agnostic, quantifiable evaluation of the value the customer receives from the product industry, category, or type. For example, the disclosed systems and methods ingest quantifiable data on value, or delivery on universal goals (or CPIs) a customer extracts from the purchase of the product from a brand.. In this ingestion of data, customers may provide quantifiable input related to the product category in general, a subset of high performers in the product category, or the like. The brand-specific importance of the product to the customer is a brand-specific, quantifiable evaluation of the value the customer receives, or outcomes it derives in association with CPIs, from the specific product offered by the target brand. For example, the disclosed systems and methods ingest quantifiable data on the value a customer extracts from the purchase from the target company, previous experience data of a customer with a competitor, and the like.

[00015] The customer-centric overall importance and brand-specific importance data are ingested as multiple CPIs. CPIs are the functional, emotional, and social goals that matter most to customers in terms of their own human-centered goals. A CPI or a combination of CPIs is selected to operationalize an analytical model of value exchange between the CPI(s) and company outcomes. The CPIs that extract the most value for the customer are selected and evaluated to determine an optimal revenue lift or “target lift” when the company delivers on them. Here, “target lift” is defined as a quantifiable improvement or progress towards a commercial outcome or goal. Target lift can be defined as measurable metrics that indicate revenue improvement for the brand and its business. Target lift can be a specific amount or value of improvement (e.g. % improvement in customer conversion rate) that a company wishes to adjust a particular outcome or goal. Target lift can also be simply a stated outcome or goal improvement, such as improving an identified group of outcomes, for example, which does not identify a specific amount or value. CPIs producing the most value extraction from the commercial relationship between the company and the customers are identified as the strongest paths available to drive customer-centered growth in the marketplace. Through an integrated focus on both CPIs and Future Customer Value (FCV), organizational leaders are able to steer their business in ways that both operationalize and monetize these opportunities.

[00016] More specifically, the modeling of FCV as a function of custom selected, quantifiable CPIs helps to identify new opportunities for customer-centered company growth; score CPIs based on their value to customers to maximize value extraction by customers from the commercial relationship; valuate the identified CPI(s) to competitor CPI(s) or behaviors; calculate company growth in FCV based on adjusting one or more CPIs and evaluate those adjustments against competitor(s); make recommendations and predictions for company KPIs based on the value extraction, identified CPI(s), or FCV.

[00017] Further, the customer-centric CPIs identified and used to model FCV can be used to detect early warnings of competitive and market threats to company performance and identify which of those possible threats are most significant and can be most affected by adjusting one or more CPIs. It can also quantify the likely lost opportunity or loss prevention to the company that could be caused by those identified threats. Considering the same modeling from the opportunity side, the same customer-centric data is also used to identify opportunity for establishing desired market position, causing competitive or industry disruption, and driving quantifiable growth through increased customer acquisition, retention, or customer value or customer spend.

[00018] Even further, the customer-centric CPIs identified and used to model FCV help companies identify their most valuable customers and compare those customers to a more average or typical customer. The data on the difference between various customer groups helps focus selection of CPIs to adjust, company goals to set, and overall growth strategy.

[00019] The customer data used to generate and help identify the CPIs of interest is useful to track demographic, psychographic, and geographic information to generate unique customer profiles on individual customers, groups of customers (e.g., regional customers), or customer types (e.g. most valuable customers). This kind of profiling can help companies identify potential new customer acquisition opportunities, activate diagnostics that fuel the innovation pipeline, and identify priority areas for company investment. For example, each CPI that is selected as a driver of company growth - defined as increased future customer value in the model - can be used to determine company direction on overall performance (e.g. how a product is performing across the market) and brand-specific performance (e.g. how a product from a particular company performs against competitors offering the same or similar products in the market). These quantifiable CPIs identify areas for growth in the way in which companies provide meaningful value to customers, based on customer need and customer-centric parameters. The same data can also evaluate a company’s commercial relationship with its customers against other top industry performers and identify which of those competitors are providing different levels of value in key customer-centric, quantifiable ways.

[00020] FIG. 1 is a process flow diagram showing value extraction of a customer’s commercial experience with a product or brand based on Customer Performance Indicators (CPIs) 100. The CPIs are determined from ingested customer data relating to brand-specific performance.. This ingested customer data is received in any suitable manner from standardized customer responses to various survey data, analyzed customer behavior, free-form customer input, customer interviews, and the like. Some of the customer data is self-reported, which can be noisy. To correct or filter for this noise, the ingested data is compared to a standard to normalize it into distributions.

[00021] The system receives initial value data for multiple Customer Performance Indicators (CPIs) related to performance of the product or brand in the consumer market 102. Initial value data is raw data collected from any source relating to customer input around the product or brand, the market of the product or brand, the product or brand type or product category, experience with competitor product(s) or brand(s), and the like.

[00022] For example, the disclosed systems and method identified multiple CPIs that most represent human-focused customer outcomes that are associated with a product, a product type, a brand, or a market for the product. The human-focused customer experience is based on human need states that reflect the future of human consumerism of products. The human need states are categorized in this disclosure as functional, emotional, and social states for any consumer having a commercial relationship with a company. These human need states create the value for the consumer that is produced by the consumer using the product or services 104 - the “value creation” the consumer experiences by engaging with the company in the commercial relationship. The value creation is mapped to a set of desired commercial outcomes or goals for the company. The mapping correlates each human need state, a group of human need states, or a category or categories of human need states (functional, emotional, or social) with a commercial outcome 106. This is accomplished via a machine learning model which aims to learn the mapping for a brand. The model receives as input data measuring the CPIs, i.e. the value created by the brand for their consumers, and is trained to predict FCV, i.e. the value extracted by the brand from their customers. [00023] The CPIs each have a quantifiable impact on FCV, and, therefore generally drive commercial outcomes and goals because of the mapped data relationship. The mapped data relationship means that a positive adjustment of a CPI within the multiple CPIs produces a positive impact on FCV. That positive impact value could be a change that exceeds a threshold, is compared against empirical data, is evaluated against competitor data for the same CPI, or the like. The change can be defined in any suitable manner; however, it is often defined as an amount or value of change in CPI(s) that generates a statistically significant improvement in FCV. For example, performance improvement on a single CPI or group of CPIs produces a correlate improvement in FCV. That correlate improvement in FCV can be a ratio, exponential, direct value, or any other type of correlation of improvement between an improvement in the CPI(s) and the improvement in FCV.

[00024] Further, adjusting more than one CPI can have a compounding effect on the correlate improvement of FCV. In some examples, the disclosed systems and methods output recommendations to adjust more than one CPI to realize an optimal improvement in FCV. Some CPIs have different levels of importance to customers with a commercial relationship with the target company compared to the commercial relationship the same customer has with a competitor company because the relationship the customer has with each company is dependent upon the functional, emotional, and social human needs discussed above. In this example, the target company has a different set of value created by its products and value extracted by the customer from use of the products as they relate to the customers functional, emotional, or social human needs than the same customer’s commercial relationship with a competitor. More specifically, a customer extracts value from its commercial relationship with the target company in functional and emotional human needs while it extracts value with a competitor company in a social need. The impact to FCV in improving the functional and emotional human needs the customer extracts - the CPIs - with the target company is greater than the impact to FCV in improving the social human need the customer extracts from its relationship with the competitor. [00025] Still further, the target company can evaluate its commercial relationship with its most valuable customers against the commercial relationship that a competitor or group of competitors has with its own most valuable customers. The identified CPI(s) that drive the target company’s most valuable customers to maintain the commercial relationship with the target company may differ from the CPI(s) that drive the competitor’s most valuable customers to maintain their own commercial relationship with the competitor. In this example, the identified CPI(s) that drive most change in FCV for the target company and the competitor often impact FCV in a substantial way, especially as those CPIs are compared to the change in FCV that adjusting other CPIs might produce.

[00026] The identified CPIs that drive the most change to FCV for a company can be included in a “brand fingerprint” for the company that indicates the unique foundation and drivers for the commercial relationship between the company and the customers. Logically, companies that understand their brand fingerprint on customers correlate the CPIs producing the brand fingerprint to determine KPIs, mitigate risk, avoid industry threats and disruption, improve market position, and generate targeted customer acquisition strategies.

[00027] The multiple CPIs are optionally validated by overlaying data from companies that provide quantitative input on the multiple CPIs that are mapped to commercial outcomes or goals 108. While this ingestion of company-provided data is optional, it helps to validate the multiple CPIs evaluated to develop the target CPIs, the FCV, and ultimately drive growth and goals in a customer-centric approach. The validation process seeks to confirm the relationship between the CPIs evaluated and the company-provided customer success metrics, e.g. customer experience score, or an internal calculation of customer lifetime value. This helps better contextualize and understand customer goals in the way they relate to legacy KPIs the business may be more interested in measuring in lieu of FCV. Examples of this type of study could emerge in two ways. The first being a confirmatory analysis that the FCV scores produced by the Vex model correlate with one or more legacy KPIs. This would give the brand confidence that the demonstrated impact of CPIs on FCV would indeed translate to similar impacts on these legacy KPIs. The other type of study would involve removing FCV from the model and replacing it with one or more of the brand’s legacy KPIs. This allows us to operationalize an analytical model of value exchange with respect to the legacy KPIs and CPIs. However, the brand’s legacy KPIs in each scenario will potentially be unavailable for competitor brands, which prevents any comparative analysis to these competing brands.

[00028] Before deciding which CPIs are most important to generate the target lift in particular commercial outcomes or goals, an initial or baseline FCV can be calculated based on CPI data of aggregate of total CPIs 110. This baseline value is used to compare additionally generated FCV values against to ensure that the suggested CPI adjustment generates a quantifiable target lift or progress towards a commercial outcome or goal. In some examples, the initial FCV is calculated based on one or more categories of quantifiable data whether the customer is likely to do business with the brand in the future, the current share of the customer’s wallet held by the brand compared to direct competitors, or frequency of spend by the customer on the brand or product. Other example systems and methods can exclude calculating an initial FCV even though it is a validating measure to help normalize future calculated FC Vs. Instead of evaluating absolute values of FCVs to each other in the future, each of the FCVs could be evaluated against the initial FCV if the initial FCV is calculated. However, one of skill in the art can also understand that using absolute values of future FCV (those calculated based on one or more adjusted CPIs) without comparing each against an initial FCV is also included in the scope of this disclosure. [00029] Referring again to FIG. 1, the disclosed systems and methods determine a level of overall importance and a level of brand-specific importance for each of the multiple CPIs based on the respective initial value data of past customer value for each of the multiple CPIs 112. The initial value data of past customer value for each of the multiple CPIs relates to the ingested data around the value extracted by the customer in past transactions, experiences, or engagements with the company in general or specifically with the product or with another product the company offers.

[00030] In some examples, the company selects a business outcome improve. The consumer market goals can be any suitable goal, including but not limited to KPIs. They are goals or outcomes that are quantifiably measured and quantifiably impacted by adjusting one or more CPIs. The consumer market goals have at least one and in some examples multiple parameters or characteristics, such as subjective and objective metrics related to the commercial relationship between the company and the customer. For example, when predicting FCV, a machine learning model helps us understand the ways CPIs influence that outcome. If we were to model a different outcome such as revenue growth, the model would then help us understand how CPIs affect revenue growth. In scenarios where a KPI other than FCV is being modeled, that data may only be available for the brand being studied (e.g. Grocery Chain A won’t have data on the revenue growth of Grocery Chain B).

[00031] The systems and methods then evaluate a level of overall importance and a level of brand-specific performance for each of the multiple CPIs 114, 116 in the manner discussed above. This evaluation of overall and brand-specific goals can be done based on a parameter or characteristic of the consumer market goal. One or more CPIs within the multiple CPIs is identified as a target group based on the evaluation of the level of overall importance and the evaluation of the level of brand-specific importance for each of the CPIs 118. The target group is the subset of one or more CPIs that most impacts FCV when adjusted. While most or potentially even all of the CPIs have a positive impact on the FCV if adjusted upward, the target group of CPIs are directly correlated with the consumer market goal. This direct correlation produces a target group of CPIs that, when adjusted and improved, result in optimal impact to lift in FCV and resulting improvement to the consumer market goal. Changing the consumer market goal or combination of consumer markets goals changes the CPIs that need adjusting. Additionally, when one or a group of CPIs is adjusted, it can affect another CPI or group of CPIs that were not adjusted. When this happens, the effect to the CPI(s) that were not adjusted could have positive or negative effects. This kind of secondary or “halo effect” evaluation of CPIs that were not adjusted can also be included to identify the CPI(s) to adjust that results in the optimal FCV target lift. The secondary evaluation of CPIs that were not adjusted are also considered from quantitative data evaluation in a similar way that the adjusted CPIs are evaluated. These can be intended or unintended secondary consequences and may be positive, negative, or a combination across multiple CPIs that were not adjusted in some examples.

[00032] For example, if a company identifies two business outcomes - revenue growth and market share - to improve, the target group of CPIs differs for the combination than it would for either of the goals if individually identified or if combined with a different goal. Specifically, the CPI or CPIs that most impact(s) improvement in revenue growth may very well be different than the CPI or CPIs that most impact(s) market share. Furthermore, the CPI or CPIs that most impact(s) the combined goals of revenue growth and market share may yet again be different. The target CPIs with the largest effect on revenue growth correlate to increasing satisfaction among existing customers and finding ways for these customers to explore a wider variety of the brand’s offerings than these customers currently explore. On the other hand, the CPIs having the largest impact on market share would relate more to acquiring brand new customers.

[00033] The disclosed systems and methods then determine value creation data related to the performance of the product or service in a consumer market based on the evaluation of level of overall importance and evaluation of level of brand-specific importance for target group of CPIs 120. As discussed above, the value creation data is the value provided by the product or brand through the commercial relationship between the customer and the company. The value creation is linked to the customer-centric or quantifiable metrics or measures of the commercial relationship between the customer and company that relate to the functional, emotional, or social human needs of the customer - the target CPIs that most affect the value of the product or brand. That value creation is then translated into value extraction data. To determine the value extraction data, the value creation data is matched, aligned, correlated, or otherwise associated with a characteristic or value of the initial FCV 122. In some examples in which the initial FCV is not determined, the value extraction data can be matched, aligned, correlated, or otherwise associated with a characteristic of a gold standard or sample high-performing competitor FCV based on the same or similar target CPIs. In yet another example, the value extraction data is matched, aligned, correlated, or otherwise associated with a characteristic or value of the initial FCV and a gold standard FCV or other empirical data from high-performing competitors. A machine learning model can track this data and match the value extraction data analysis with the best fit gold standard and/or empirical data based on one or more of the target CPIs. In some examples, the value creation data is matched, aligned, correlated, or otherwise associated with multiple FCVs. In this example, the multiple FCVs can include the initial FCV and one or more of the additionally generated FCVs based on adjusting different CPIs or groups of CPIs or on an iterative process. The multiple FCVs can also include the initial FCV and a gold standard or high-performing competitor FCV or any combination of multiple FCV types and options.

[00034] The disclosed systems and methods then correlate the value extraction data, and in some examples also the value creation data, to the KPI related to performance of the product or brand in the consumer market 124. As discussed above, the company outcomes or goals, such as a KPI, are based on the value extracted from the commercial relationship between the customer and the company. The value extraction data is directly linked to the customer human needs based on quantifiable values relating to the functional, emotional, or social needs of the customer.

[00035] The disclosed systems and methods then outputs the target group of CPIs, the value creation data, the value extraction data, the FCV, or the KPI 126. The output can also be used to develop predictions, forecasts, and diagnostics of the product, the company, and the commercial relationship between the customer and the company at the brand level and the product type level, for example. [00036] The developed predictions, for example, can be generated by adjusting a single CPI, in isolation, or adjusting multiple CPIs to generate respective FCVs. The predictions can also adjust differing combinations of CPIs as well. The selected target CPIs and the adjustments or predictions made to result in the optimal FCV can also be determined on subjective input from the company or customer. To determine the selected target CPIs, a FCV from a first CPI adjustment can be evaluated against a FCV from a second CPI adjustment. The respective CPI adjustments can include a single CPI or an adjustment of multiple CPIs. The FCVs can be compared to each other or against a threshold or empirical data, such as a gold standard or competitor values, for example. The target lift value produced by the first CPI adjustment can be compared to a target lift value produced by the second CPI adjustment. The comparison can be used to determine whether the first CPI adjustment, the second CPI adjustment, or both adjustments in combination (series or parallel) produces the optimal target lift value for the company.

[00037] The model could produce a set of CPIs or an optimal FCV that recommends adjusting the target group of CPIs in a particular way that does not align with the company’s resources, for example. In this case, the company would alter the target CPIs to be supported by existing or attainable resources available to execute the suggested adjustments.

[00038] For example, a target CPI or CPIs may be selected in order to drive improvement in a specific business KPI such as FCV, revenue growth, market share or others. Information from company can inform which CPIs are most likely to be impacted, the cost of driving improvement, and other business matters. The disclosed systems and methods would allow for a prediction of the improvement on business outcomes from improving CPI(s) and improvements to different set(s) of CPI(s) can be modeled to predict which CPI or CPIs would result in improved net business outcomes.

[00039] FIG. 2 shows an example model showing a functional relationship between Customer Performance Indicators (CPIs), Future Customer Value (FCV), and key performance indicators (KPIs) 200. In this example, the human needs 200 are categorized into functional 204, emotional 206, and social need 208 states, as discussed above. Functional human need states include: easy access to the information the consumer needs or wants; an ability to make the consumer’s life easier in any defined parameter; an ability to save the consumer time compared to non-use of the product; improvement of the consumer’s physical wellbeing; an ability to save the consumer money; and an ability to give the consumer options. Emotional human need states include improvement to the customer’s mental well-being; an ability to make the customer feel good; an ability to reduce the customer’s anxiety or risk; an ability to motivate the customer; and an ability to provide the customer with a sense of accomplishment. Social human need states include providing a sense of belonging to the customer; allowing the customer to connect with others; allowing the customer to help others; and elevating the customer’s perceived status.

[00040] The human needs are quantifiable data generated from customer input, such as surveys, behavior tracking, free form or standardized feedback, and the like. As discussed above, the data is quantifiable into an absolute or relative value and can be normalized to a standard, in some examples. These CPIs are used to determine a measured value creation, which is value creation data that relates to the value of the product or brand to the customer based on the human needs or CPIs.

[00041] The value creation data is categorized into three types - importance, performance, and penetration 210. For example, CPI performance and penetration may be measured via customer input in the forms of data listed above. CPI importance with respect to FCV or another KPI reflecting the value received from customers is estimated via a machine learning model.

[00042] As discussed above, value extraction data is generated as a function of the value creation data 212. Alternatively, the CPIs can directly correlate with the value extraction data without first generating the value creation data. The value creation data, however, provides another example of the data-driven, quantitative link between value provided to the customer and value extracted by the brand from the commercial relationship - the value created by the product or brand is a function over time of the value extracted by the customer by use of the product or relationship with the brand. As discussed above, the value extraction data is correlated, aligned, matched, or otherwise associated with a company outcome or goal, such as the KPIs 214 shown in FIG. 2. The KPIs 214 shown in FIG. 2 include outcomes and goals related to company revenue, profitability, and asset efficiency 216. The company revenue, profitability, and asset efficiency 216 KPIs have parameters or characteristics that include: recurring revenue rate, average revenue per customer, revenue growth, gross profitability, cost of goods and services sold, net profit, customer conversion rate, customer lifetime value, employee retention, and employee engagement level. FCV can be a proxy for revenue, was assigned to be the de facto KPI in that it is composed of the drivers of revenue: share of wallet; frequency of purchase; intent to purchase into the future. [00043] FIG. 3 is an example process flow diagram showing how CPIs and FCV are used to provide predictive, prescriptive, and diagnostic data 300. The example shown in FIG. 3 is a value exchange model that includes three integrated algorithms: (1) an FCV machine learning (ML) algorithm; (2) a simulation and forecasting algorithm; and (3) a natural language processing (NLP) algorithm powered by artificial intelligence (Al).

[00044] Each algorithm takes a different combination of inputs in the form of the ingested data related to CPIs and outputs from the other algorithms in the system. As discussed above, the ingested data can be in the form of various customer survey data, analyzed customer behavior, free-form customer input, customer interviews, and the like. The outputs from the other algorithms help refine and improve the generated data. For example, FCV helps diagnose company health and use predicted or simulated adjustments to CPIs 315 as an ingested input to generate new scores based on the predictions and simulations. The CPI improvement simulation or prediction algorithm 308 predicts FCV using simulated scores for one or more CPIs.

[00045] In the example shown in FIG. 3, the FCV ML algorithm 302 receives survey response data measuring both initial FCV 304 and survey results from fifteen CPIs. The system can additionally ingest other indicator variables for various brands. The FCV ML algorithm 302 is trained using cross-validation on the prepared survey response data to predict FCV. The degree of influence each predictor variable has on FCV is obtained by analyzing given values of the CPIs and their average marginal effect on FCV. This analysis is applied at the overall market level as well as the individual brand level to assess CPI importance at these different granularities. Additional ingested data for the FCV ML algorithm include: predictor variables in the model (i.e. independent variables); top-box recodes (i.e. conversion of discrete ordinal measurement of survey data to binary variables) of the 15 CPIs; brand-level indicator variables to distinguish brands from each other; sector-level indicator variables to distinguish between groups of similar brands; and variables the model is trained to predict (i.e. the dependent variable that change in response to another changed variable).

[00046] As discussed above, FCV is created as a composite score from three separate dimensions captured in the survey: purchase frequency (e.g. a quantitative measure of how frequently a customer purchases from or spends money with the company); likelihood of future purchase (i.e. retention probability); and share of wallet (i.e. the share of spend the company receives, relative to the other competitors the customer engages). [00047] Purchase frequency can be quantified by asking customers to provide numerical responses to a survey inquiry such as “Thinking about the future, how often are you to continue to purchase products, services, accounts/subscriptions from the company over the next 12 months?” The customer can be offered responses to the survey on an available eight-point scale, such as:

1 Daily = 5

2 Few times per week = 5

3 Weekly = 4

4 Few times per month = 4

5 Monthly = 3

6 Few times per year = 3

7 Yearly = 2

8 Less than yearly = 1

[00048] Further consolidation is applied to normalize the resulting purchase frequencies to make these values more comparable across industries with varying purchase patterns - e.g. customer do not purchase insurance daily no matter their affinity for the company. To do this, for each company, the purchase frequencies with incidence below 10% are consolidated into an adjacent level, collapsing toward the middle of the scale. If values of 1 (less than yearly) occur in less than 10% of a brand’s responses, these values are recoded to a value of 2 (yearly). Looking at the frequency for values of 2 (including any previously recoded values of 1), if these occur in less than 10% of a brand’ s responses, these values are recoded to a value of 3 (monthly/few times per year). These steps are then repeated for both values of 5 (daily/few times per week) and values of 4 (weekly/few times per month). The remaining values are consolidated into three distinct levels: the max value remaining receives a value of 3 indicating the highest purchase frequency; the lowest value remaining receives a value of 1; all other values in-between receive a value of 2.

[00049] The likelihood of future purchase (i.e. retention probability) can be quantified by asking customers to provide numerical responses to a survey inquiry such as “Thinking about the future, how likely are you to continue to purchase products, services, accounts/subscriptions from the company over the next 12 months?” The customer can be offered responses to the survey on an available five-point scale, such as: 1 Extremely unlikely = 0.1 probability of retention

2 Unlikely = 0.3

3 Unsure = 0.5

4 Likely = 0.7

5 Extremely likely = 0.9

[00050] Share of wallet can be quantified by evaluating the share of spend the brand receives, relative to the other competitors the respondent shops with:

(l-Rank/(# Brands+l))*(2/(# Brands)) where Rank is how the brand being evaluated ranks relative to the competitors the respondent also uses (1 = best experience, 2 = second best, etc.) and # Brands is the total number of brands the respondent uses (the current brand being evaluated plus the number of competing brands they use). [00051] The possible values as a result of this calculation range from 0.005 (if out of the maximum allowed 20 competitor brands, for example, the respondent ranks this brand #20 out of 20) up to 1.0 (the respondent uses zero competitors).

[00052] FCV is computed from these three variables as:

(Purchase frequency * Likelihood of future purchase) + Share of wallet

[00053] To help remove noise around FCV and improve model fit, a transformation is employed to micro-aggregate the raw data. This is the variable the predictive model 302 is trained to predict. For each company, deciles are computed from FCV. The deciles are calculated by partitioning the underlying distribution of values into ten numeric ranges, each containing a similar proportion of the data. The average FCV within each company’s deciles are then used to replace the raw values of FCV. Outputs from the model includes each company’s predicted FCV score; the relationship between FCV and each CPI at the brand-level (partial dependence plots; the FCV by CPI curves); the size of the impact each CPI has on FCV, on average; and the cumulative CPI score.

[00054] For the company’s predicted FCV score, a validation model can be used to predict the FCV score. For example, a support vector regression is trained to predict FCV using 10-fold cross- validation using support vector regression (SVR). SVR is like another possible option to predict the FCV score, ordinary least squares (OLS) regression, in that it is defined to minimize the prediction error but also the size of the model’s parameters. SVR also does not aim to find the smallest average prediction error possible like OLS, but instead aims to reduce the average prediction error below some threshold. The SVR is trained not to find the best model, but instead to find one that is good enough (i.e. a model with error below a toleration amount according to a threshold). Using the SVR helps mitigate the confounding effects of multicollinearity (i.e. correlation among the independent variables) and allows for excellent generalization capability while maintaining high accuracy. This technique also works well in high dimensions (i.e. when the number of inputs to the model are sufficiently large). Any suitable validation model can be used or no validation model can be used in alternative examples.

[00055] Ten-fold cross-validation used in predicting FCV is a strategy for training a generalizable machine learning model. One drawback of machine learning models is that they can sometimes “overfif ’ the data. This means the model is very accurate when predicting the same data on which it was trained but fails to be as accurate when predicting out-of-sample observations that were not available during training. This usually means the model cannot generalize outside of the sample used for training, which means the model has not “learned” the relationships in the data, but has more or less “memorized” the training data and cannot determine what to do with a new data point. [00056] The ultimate FCV score is the average predicted FCV scaled as a percentage of what is possible based on the FCV ML algorithm. The FCV ML algorithm produces predicted FCV values for every customer, which are then scaled based on the lowest predicted FCV for the company and the highest FCV:

(Predicted FCV-Min)/(Max-Min)

[00057] This equation expresses the model’s predictions as a percentage of what is possible for the company. Taking the average of these scaled values across respondents gives us the company’s ultimate FCV score.

[00058] As shown in FIG. 4, a value exchange curve can be plotted, which is a representation of the relationship between CPIs and FCV for each company 400. The horizontal axis represents the CPI score for all CPIs. The noted percentage lift in FCV represents the predicted increase that can be realized when the company focuses on improving its top three CPIs. [00059] The disclosed systems and methods adjust or suggest various changes in the predictor variables (ex: CPIs), in isolation, to see how those changes impact predicted FCV. To do this, the disclosed systems and methods first focus on a single predictor variable (e.g. one of the CPIs) in this example. For each unique value of that CPI, a copy of the data used to train the model and then replace all original values of the predictor CPI with each unique value (i.e. the data set is copied once for each unique value of the predictor; each copy has a constant value for the predictor of interest). Next, the modified data set copies to the predictive model to obtain predicted values of FCV. The disclosed systems and methods then average across the predicted FCV values from each copied data set, resulting in the average prediction for each unique value of the predictor CPI being interrogated.

[00060] These relationships are charted to visualize or “predict” the average marginal effect each CPI has on FCV 310. To align these predicted values more closely with the observed scores in the data, the predicted changes in FCV can be re-calibrated in some examples so that the same baseline FCV is predicted at each individual CPFs aggregate top-box score, as shown in FIG. 3. For example, if a brand’ s aggregate top-box score for the CPI ‘ Saves me time’ is 28, and the system predicts a 5% lift from baseline FCV at this value of the CPI, the system then subtracts 5% from the entirety of predicted lifts for this CPI in order to show how predicted FCV changes from the brand’s actual CPI score. In alternative examples, no re-calibration is needed or desired.

[00061] The size of the impact each CPI has on FCV, on average is also calculated using the prescriptive simulation model that helps forecast and make recommendations based on the ingested and analyzed data. We determine the impact of each CPI for a given brand by analyzing the difference in predicted FCV when the CPIs are at their minimum possible value and their maximum possible value. After altering a given CPI from its baseline down to the minimum or maximum possible value, we send these updated ratings through the simulation model to account for the “halo effect” of the movement from baseline. The resulting predicted FCV after accounting for the full constellation of changes in the CPIs is used as that CPI’s impact on FCV for the given brand. The difference between the predicted lifts in FCV obtained from the simulation model when the CPI is maximized and minimized is assigned as that CPI’s importance score. Any CPI with a difference smaller than 1% lift on FCV is considered to have a negligible impact and is assigned an importance score of zero. Once the importance scores for all 15 CPIs are determined, these scores are normalized to sum to 1. [00062] A cumulative CPI score is calculated based on the size of the effect each CPI has on FCV and whether the CPI is non-negligible. Negligible CPIs are given a weight of zero while the others are weighted proportionally based on their impact on FCV. These weighting values are applied at the respondent-level - the individual responses received in a customer survey - and the resulting linear combination of CPI top-box scores and weights (i.e. multiply the CPI effect sizes by their corresponding top-box scores and take the sum of these products) are averaged across customer input, which results in the company’s ultimate cumulative CPI score.

[00063] Referring again to FIG. 3, a simulation and forecasting algorithm recommends the next best step a company should take to improve FCV 310 based on changes in the predictor CPI variables and the FCV modeling. The simulation and forecasting algorithm generates the recommendations in two ways: how a change in any single CPI impacts the others, and how the full constellation of changes in the CPIs ultimately impacts FCV based on those quantified values discussed above. A probabilistic model is used to determine - given changes in a single CPI - how the distributions of the related CPIs should also be impacted. Once the full effect of the single-CPI shift has been accounted for in all other CPIs, these new simulated scores are fed to the FCV ML algorithm to generate a predicted FCV 310, which shows how the single-CPI shift ultimately impacts FCV.

[00064] Next, the system determines the CPI or CPIs whose improvements have the greatest impact on FCV after accounting for changes in the dependent CPIs, the degree of impact on FCV, and the successive next-best steps after the initial improvements are made. As discussed above, the CPI values and micro-aggregated FCVs (both discussed above) are used to develop companylevel Bayesian networks (BNs) capturing the dependencies among the CPIs. A Bayesian network is a type of model that contains a qualitative and a quantitative component. The qualitative component is the set of relationships among the variables in the model (e.g. A influences B). The quantitative component is the amount of influence present in each relationship.

[00065] The system uses the BNs to estimate CPI dependencies at the company level using the generated BNs. For each company, nonparametric bootstrap estimation is used to repeatedly sample the data and generate a network structure from each sample. From this data, the system then computes the strength and direction of every possible CPI dependency. For example, is the CPI ‘Improves my mental wellbeing’ dependent on the CPI ‘Makes me feel good’ or is the dependency reversed. [00066] The system also measures the frequency of the connection during the bootstrap estimation to determine whether the connection is valid. Dependencies that are estimated above a certain probability threshold are used to build the final network of dependencies. The threshold can be predefined or user controllable, as needed. The final network of dependencies is then used to fit a model quantifying the amount of impact present in each relationship.

[00067] The BNs and given changes in CPIs to determine the “next best action” are then used to improve FCV. First, a selected level of impact on a given CPI (e.g. +5 percentage points) is used to alter the observed CPI scores. This change is then pushed through the BN to update all dependent CPIs. Then, using the full suite of updated CPIs, the predicted FCV is obtained using the predictive model. The impact of the improvements in the given CPI are quantified by computing the percent change in FCV from the company’s current baseline FCV (described above in Section 1):

% Change in FCV= (Updated FCV-Baseline FCV)/(Baseline FCV)

[00068] This is repeated for all CPIs to determine which allows for the greatest improvement in FCV. The CPI whose assumed improvements have the largest resulting % change in FCV is suggested as the CPI to improve in Step 1 of the company’s efforts. The following steps in the simulation use the updated CPI scores 315 and updated FCV 310 as the starting point. This allows us to see after the suggested improvement has been realized in Step 1, where the company should focus their improvement efforts next.

[00069] A natural language processing (NLP) algorithm 316 determines, based on this quantified data, how a company can act on high-impact CPIs in relevant ways to improve FCV. Following the conclusion of the algorithm that identifies the most important variables that impact the other predictors and FCV 310, 312, 314, the NLP -based algorithm determines how the simulated changes in individual CPIs could be realized based on open-ended responses 318 collected with the survey. Using term-based sentiment analysis, the NLP algorithm 316 detects different terms mentioned in the response and scores the sentiment polarity based on the context of that term. Terms with sentiment distributed more positively suggest attributes of the customer experience the company is delivering well, indicating, or validating reasons for the high CPI ratings. Terms with sentiment distributed more negatively suggest attributes of the customer experience where the company can improve, potentially driving the improvements simulated in the CPIs, and thus increasing FCV. [00070] Though certain elements, aspects, components or the like are described in relation to one embodiment or example, such as an example diagnostic system or method, those elements, aspects, components or the like can be including with any other system or method, such as when it desirous or advantageous to do so.

[00071] The foregoing description, for purposes of explanation, used specific nomenclature to provide a thorough understanding of the disclosure. However, it will be apparent to one skilled in the art that the specific details are not required in order to practice the systems and methods described herein. The foregoing descriptions of specific embodiments are presented by way of examples for purposes of illustration and description. They are not intended to be exhaustive of or to limit this disclosure to the precise forms described. Many modifications and variations are possible in view of the above teachings. The embodiments are shown and described in order to best explain the principles of this disclosure and practical applications, to thereby enable others skilled in the art to best utilize this disclosure and various embodiments with various modifications as are suited to the particular use contemplated. It is intended that the scope of this disclosure be defined by the following claims and their equivalents.

Claims

CLAIMS What is claimed is:

1. A method of quantifying extracted value of a product or brand to a customer, comprising: receiving initial value data for multiple Customer Performance Indicators (CPIs) related to an ability of a brand to deliver on CPIs to its customers; determining a level of overall importance and a level of brand-specific importance for each of the multiple CPIs based on the respective initial value data of past customer value for each of the multiple CPIs; evaluating the level of overall importance for each of the multiple CPIs based on a parameter or characteristic of a desired business outcome evaluating the level of brand-specific importance for each of the multiple CPIs based on the parameter or characteristic of the desired business outcome identifying a target group of CPIs within the multiple CPIs based on the evaluation of the level of overall importance and the evaluation of the level of brand-specific importance for each of the CPIs; determining value creation data related to performance of the product or brand in the consumer market based on the evaluation of the level of overall importance and the evaluation of the level of brand-specific importance for each of the target group of CPIs; determining value extraction data related to the performance of the product or brand in the consumer market based on the value creation data and a characteristic or value of a Future Customer Value (FCV) for each of the target group of CPIs; and correlating the value creation data and the value extraction data to a key performance indicator (KPI) related to performance of the product or brand in the consumer market; and outputting one or more of the target group of CPIs, the value creation data, the value extraction data, the FCV, or the KPI.

2. The method of claim 1, wherein the FCV for each of the target group of CPIs includes a quantifiable value or a qualitative value compared against a standard.

22

3. The method of claim 1, further comprising determining value extraction data related to the performance of the product or brand in the consumer market based on the value creation data and a characteristic or value of multiple FCVs for each of the target group of CPIs.

4. The method of claim 1, wherein the FCV includes one or more of a customer share of wallet, frequency of customer purchase, and a customer likelihood to continue to purchase into the future.

5. The method of claim 1, further comprising adjusting a first CPI of the target group of CPIs based on one or more of the value creation data, the value extraction data, and the characteristic or value of the FCV to correlate to a target lift value in the KPI.

6. The method of claim 5, further comprising determining a target lift value in the first CPI based on the target lift value in the KPI.

7. The method of claim 6, further comprising outputting one or both of the target lift value of the KPI and the target lift value of the first CPI.

8. The method of claim 5, further comprising adjusting a second CPI of the target group of CPIs based on one or more of the value creation data, the value extraction data, and the characteristic or value of the FCV to correlate to a target lift value in the KPI.

9. The method of claim 8, wherein the second CPI is adjusted sequentially after adjustment of the first CPI.

10. The method of claim 8, further comprising comparing an adjusted value of the first CPI with an adjusted value of the second CPI, and selecting either the first CPI or the second CPI based on a magnitude of the resulting change in FCV given the CPI adjustments.

11. The method of claim 1, further comprising: adjusting a first CPI of the target group of CPIs based on one or more of the value creation data and the value extraction data, and predicting a first target lift value for the FCV based on the adjusted first CPI.

12. The method of claim 11, further comprising: adjusting a second CPI of the target group of CPIs based on one or more of the value creation data and the value extraction data; predicting a second target lift for the FCV based on the adjusted second CPI; and outputting one or both of the first target lift value and the second target lift value.

13. The method of claim 12, further comprising: comparing the adjusted first CPI to the adjusted second CPI, the adjusted first CPI and the adjusted second CPI to a threshold, or the adjusted first CPI and the adjusted second CPI to empirical data; and predicting the first target lift value for the FCV and the second target lift value for the FCV based on the comparison of the adjusted first CPI to the adjusted second CPI, the adjusted first CPI and the adjusted second CPI to a threshold, or the adjusted first CPI and the adjusted second CPI to empirical data.

14. The method of claim 12, further comprising selecting the first CPI, the second CPI or both for adjustment based on a comparison of the first target lift value to the second target lift value.

15. The method of claim 1, wherein the multiple CPIs relate to quantifiable data or values associated with a customer functional, emotional, or social experience with the product or brand in the consumer market.

16. The method of claim 1, wherein the product or brand is a company product or brand, and further comprising: comparing the value creation data for the company product or brand with value creation data for a competitor product or brand; and comparing the value extraction data for the company product or brand with value extraction data for a competitor’s product or brand.

17. The method of claim 16, further comprising: generating a comparison between the company product or brand and the competitor’s product or brand based on the comparison of the value creation data for the company product or brand with the value creation data for the competitor’s product or brand and the comparison of the value extraction data for the company product or brand with the value extraction data for the competitor’ s product or brand; and outputting the comparison.

18. The method of claim 17, further comprising determining a recommendation to employ a quantifiable enhancement to one or more of the target group of CPIs based on the comparison.

19. The method of claim 17, further comprising determining a recommendation to employ a quantifiable enhancement to the KPI based on: the comparison of the value creation data for the company product or brand with the value creation data for the competitor product or brand, and the comparison of the value extraction data for the company product or brand with the value extraction data for the competitor product or brand.

20. The method of claim 1, further comprising correlating the value creation data and the value extraction data to the KPI based on CPI values relating to revenue, profitability, or asset efficiency of the KPI.

25