KR101205409B1 - Customer Equity Evaluation System and Method therefor - Google Patents

Abstract

PURPOSE: A customer asset measurement system and a method thereof are provided to consider relations between customers by measuring customer assets and a CLV(Customer Lifetime Value) through customer evaluation information for a customer asset driver. CONSTITUTION: A customer purchase information input unit(200) inputs customer purchase information. A purchase probability matrix element calculating unit(500) calculates a purchase probability matrix element for purchasing a brand by using a customer asset evaluation score. A CLV calculating unit(600) calculates a CLV by using the customer purchase information, a company sales information, and the purchase probability matrix element. [Reference numerals] (200) Customer purchase information input unit; (300) Customer asset evaluation score input unit; (400) Company selling information input unit; (500) Purchase probability matrix element calculating unit; (510) Effective value calculating unit; (520) Selection probability calculating unit; (530) Transition matrix calculating unit; (600) Customer lifetime value calculating unit; (700) Customer asset calculating unit

Description

Customer Equity Evaluation System and Method therefor}

The present invention relates to a customer asset measurement system and method, and more particularly, to a system and method for measuring customer assets using customer evaluation information on the customer asset driver including a social network driver.

In terms of PROFITABLE PRODUCT MANAGEMENT, sustainable assorted brand growth was recognized as an object to be eliminated. On the other hand, in terms of cus- tomer equity management, even if a small portion of the total sales is found, it is an object to be carefully managed if there are customers looking for products. Based on this paradigm, Rust et al. (2004) 's customer asset driver system estimates the probability that a particular customer chooses a purchasing brand using utility theory. Probability variables are used to estimate the future cash flows of a specific brand by converting the probability of purchasing the individual brand into a transition matrix.

In this theory, utility theory, Markov chain model, and net present value model are applied. Ultimately, the future cash flow value of individual brands is calculated, and the influence of drivers (value, brand, related assets) on the value is analyzed. This provides a comparison between competitive brands and administrative implications.

The most important part of Rust's model is customer asset drivers, which consist of value assets, brand assets, and relationship assets that customers provide. Value assets reflect consumers 'objective evaluation of the products or services provided by a company, brand assets reflect consumers' subjective views on the products or services provided by a company, and related assets arise between companies and consumers. It is the value of an asset that is assessed from the consumer's perspective. This model was first published in 2000 in the book Driving Customer Equity, and was published in 2004 in the Journal of Marketing under the title Return on marketing. By substituting these values into the utility theory, consumers evaluate the drivers for each of the major competitive brands, and they substitute the utility model to represent the utility value of each brand.

Utility theory is a theory widely used in all areas of economics and management. It uses the utility model to represent consumer's subjective views in stochastic numbers.

 Although the three drivers introduced in the previous studies, value, brand, and related assets, can be measured from the consumer's point of view, the intrinsic problem of drivers begins with companies. Each driver is scored by a customer evaluation, but the driver evaluation consists of only those parts that can be directly improved by the company's marketing activities. For example, an assessment of a company's value assets is a question about the quality of the product or service, and the score of this item can be a direct response, depending on the company's product or service quality efforts. Similarly, brand assets and related assets are made up of questions of this nature. In other words, Rust's model restricts the concept of customer assets to areas that can be managed from a brand perspective. The customer asset model can explain the measurable and controllable parts of a firm, but it is rather insufficient to explain the intrinsic factors that occur between consumers or between consumers and firms. For example, the impact of social interaction on consumer brand choice is beyond the scope of this model.

In addition, as communication methods available to consumers become more diverse, consumers have a problem in selecting a brand that is influenced by various factors in addition to the existing drivers.

The present invention is to solve the above-mentioned problems, an object of the present invention is to measure the customer lifetime value and customer assets by using the customer evaluation information on the customer asset driver including the social network driver provided from the customer, To provide a customer-oriented customer asset measurement system and method that consider not only the relationship between the customer and the customer but also the relationship between the customers.

In order to achieve the above object, a first aspect of the present invention provides a customer purchase information input unit for receiving customer purchase information including information related to a purchase brand, average purchase price, purchase frequency; A customer asset driver evaluation score input unit for receiving a customer evaluation score for the customer asset driver including a value asset driver, a brand asset driver, a relationship asset driver, and a social network driver; A corporate sales information input unit for receiving corporate sales information including information related to a discount rate and total number of customers; A purchase probability matrix element calculating unit for calculating a probability matrix element for purchasing a brand using the customer asset driver evaluation score received from the customer asset driver evaluation score input module; A customer lifetime value calculator for calculating a customer lifetime value using the customer purchase information, the company sales information, and the purchase probability matrix element; And a customer asset calculating unit configured to calculate a customer asset by multiplying the customer lifetime value by the total number of customers.

Advantageously, said social network driver comprises a word of mouth element and a brand-ego identity element.

Preferably, the word of mouth element includes word of mouth intention as an evaluation of after using the brand and a word of mouth effect item that is influenced by others in the purchase of the customer, and the brand-self-identity element includes cognitive consistency, emotional consistency And evaluative consistency.

Preferably, the customer lifetime value calculation unit calculates a Customer Lifetime Value (CLV _ij ) for individual customers (i) and brands (j) by the following equation.

<Expression>

Where T _ij is the number of purchases made by the customer i during a specific period, d _j is the discount rate of the brand j, f _i is the average number of purchases of the customer i over a period of time (for example, in years), and V _ijt is the time at which the customer i Π _ijt is the expected profit per j brand purchase unit at time t of customer i, B _ijt is the brand j purchase probability matrix element of customer i at time t, and t / fi is discounted according to the time of purchase. Means that.

Preferably, the customer asset calculation unit calculates the customer asset using the average value of the customer lifetime value for all customers of the brand.

Advantageously, the purchase probability matrix element calculating unit calculates the utility value for each brand of the customer according to the driver by applying the customer asset driver evaluation score received from the customer asset driver evaluation score input module to the utility model. ; A selection probability calculation unit for calculating a selection probability for each brand by comparing the utility value of each brand of the customer calculated by the utility value calculation unit with the utility value of all brands; And a transition matrix calculation unit configured to apply a Markov chain model to a selection probability for each brand calculated by the selection probability calculation unit to calculate a transition matrix for each type of brand movement of the customer.

Preferably, the purchase probability matrix element calculation unit further comprises a principal component analysis unit for calculating a principal component coefficient for determining the relative influence of the value asset driver, brand asset driver, relationship asset driver, social network driver.

According to a second aspect of the present invention, there is provided a customer asset measurement method using a customer asset measurement system, comprising: customer purchase information input step of receiving customer purchase information including information related to a purchase brand, an average purchase amount, and a purchase frequency; A customer asset valuation score input step of receiving a customer valuation score for the customer asset unit including a value asset unit, a brand asset unit, a relationship asset unit, and a social network unit; A corporate sales information input step of receiving corporate sales information including information related to the discount rate and the total number of customers; A purchase probability matrix element calculating step of calculating a probability matrix element for purchasing a brand using the received customer asset valuation score from the step of receiving the customer asset valuation score; Calculating a customer lifetime value using the customer purchase information, the company sales information, and the purchase probability; And a customer asset calculation step of calculating a customer asset by multiplying the customer lifetime value by the total number of customers.

Preferably, the social network portion includes a word of mouth element and a brand-ego identity element.

Preferably, the word of mouth element includes word of mouth intention as an evaluation of after using the brand and a word of mouth effect item that is influenced by others in the purchase of the customer, and the brand-self-identity element includes cognitive consistency, emotional consistency And evaluative consistency.

Preferably, the step of calculating the customer lifetime value calculates a Customer Lifetime Value (CLV _ij ) for each customer (i) and the brand (j) by the following equation.

<Expression>

Where T _ij is the number of purchases made by the customer i during a specific period, d _j is the discount rate of the brand j, f _i is the average number of purchases of the customer i over a period of time (for example, in years), and V _ijt is the time at which the customer i Π _ijt is the expected profit per j brand purchase unit at time t of customer i, B _ijt is the brand j purchase probability matrix element of customer i at time t, and t / fi is discounted according to the time of purchase. Means that.

Preferably, in the customer asset calculation step, the customer asset is calculated using an average value of the customer lifetime values for all the customers of the brand.

Preferably, the purchase probability matrix element calculating step calculates the utility value for calculating the utility value for each brand of the customer according to the driver by applying the customer asset valuation score received in the step of receiving the customer asset valuation score to the utility model step; A selection probability calculation step of calculating a selection probability for each brand by comparing the utility value of each brand of the customer calculated in the utility value calculation step with the utility value of all brands; And a transition matrix calculating step of applying a Markov chain model to a selection probability for each brand calculated by the selection probability calculator to calculate a transition matrix for each type of brand movement of the customer.

Preferably, the purchasing probability matrix element calculating step further includes a principal component analysis step of calculating a principal component coefficient for determining a relative influence of the value asset, brand asset, relationship asset, social network portion.

A third aspect of the present invention is to provide a computer readable recording medium having recorded thereon a program capable of executing the above-described method of measuring customer assets by a computer.

With the above-described configuration, the customer asset measuring system and the customer asset measuring method according to the present invention complement the enterprise-oriented brand-consumer relationship structure, which is a limitation of the conventional customer asset driver model, and apply customer-oriented factors rather than companies or brands. By doing so, the multilateral customer-brand relationship can be measured by a realistic and realistic customer asset model. In other words, it is possible to increase the explanatory power of the variable market structure by adding customer-oriented factors that the asset value of the brand-oriented view cannot explain.

In addition, by applying the customer asset measurement system and the customer asset measured by the method to the marketing system according to the present invention, it is possible to create a new customer to the company away from the conventional management-oriented marketing system such as enterprise-wide resource management or customer relationship management And the strategic tools needed to find new growth engines.

1 is a block diagram of a customer asset measurement system according to an embodiment of the present invention.
2 is a flowchart of a method of measuring customer assets according to an embodiment of the present invention.
3 is a detailed flowchart of a purchase probability matrix element calculating step of FIG.

The following detailed description of specific embodiments provides several descriptions of specific embodiments of the present invention. However, embodiments of the present invention illustrated below may be modified in many different forms, and the scope of the present invention is not limited to the embodiments described below.

1 is a block diagram illustrating an overall customer asset measurement system according to an embodiment of the present invention.

Referring to FIG. 1, the customer asset measurement system 100 according to an embodiment of the present invention includes a customer purchase information input unit 200, a customer asset evaluation score input unit 300, a company sales information input unit 400, and a purchase probability matrix. An element calculator 500, a customer lifetime value calculator 600, and a customer asset calculator 700 are included.

Here, the customer purchase information input unit 200 performs a function of receiving customer purchase information including a recent purchase brand, a purchase probability of a competitive brand, an average purchase amount, a purchase frequency, and the like.

In this case, the recently purchased brand may select a recently purchased brand within a brand category. Preferably it is possible to select from the designated competition brands. In addition, the purchase probability of the competitive brand may indicate that the purchase probability for each brand is 100% if there is a certain number of competitive brands in the corresponding industry. In addition, the average purchase amount may be defined as the average amount of expenditure when purchasing a product in a category once. In addition, the purchase frequency may be defined as a purchase frequency of products in an annual category.

Here, the customer asset valuation score input unit 300 performs a function of receiving a customer asset valuation score for the customer asset unit composed of a value asset unit, a brand asset unit, a related asset unit, and a social network unit. In this case, the brand asset part includes quality and price as measurement elements, the brand asset part includes recognition, image, etc. of the brand as a measurement element, and the related asset part includes customer treatment and special programs as measurement elements. In addition, the social network unit may include word-of-mouth effect, word-of-mouth intention, and brand-ego correspondence (cognitive, emotional, and evaluative correspondence) as measurement elements.

Hereinafter, the value assets, brand assets, relationship assets, and social network assets constituting the consumer-oriented customer asset unit applied in one embodiment of the present invention will be described in detail.

1) Value Asset

A company's value assets, perceived by consumers, measure how reliable they are, how high the quality is, and what the goods are worth. A value asset reflects a relatively objective consumer's perceived opinion of a product or service and is a unilateral asset recognized by the company's product or service capabilities.

2) brand assets

It means the value added to a product or service by evaluating how the image is perceived by consumers and whether the image of the brand suits my taste. Brand assets are valued as monetary asset values as intangible values in the modern market economy, including corporate value assets, consumer awareness, and brand expansion.

3) relationship assets

It is a measure of how consumers perceive the value of the relationships that occur at customer contacts as a company provides products or services to consumers. Mainly, the quality of the face-to-face relationship is assessed, and consumers are assessing whether the company gives me a relationship benefit or special treatment.

The above client asset divisions form complementary structures according to their characteristics and influence the formation of customer assets. However, many researches on the value of these assets have been conducted before the customer asset research, and many have been criticized for artificially distinguishing the characteristics of brand assets. Aaker divides the brand equity component into four levels: brand awareness, brand loyalty, perceived quality, and brand association. Here brand royalties are related assets in customer assets and perceived quality is value assets. Despite the confusion of similarly similar factors, the reason for distinguishing these relationships in customer assets is that brand assets are valued only by brand factors, while in the customer asset model each device is used to explain more dynamic relationships. It is a product to distinguish the relationship between them and to evaluate the value of assets through the relationship between consumers and brands.

However, despite the efforts to express the customer-brand relationship of customer assets in terms of asset value, the customer assets have little explanation for factors that occur among consumers. This can be explained in two ways, first, because it was developed as a conceptual framework as an initial model. The initial customer asset model was developed to form an integrated management system for the long-term brand management aspects of the company and the brand-consumer relationship. Therefore, it can be said that the role of the system as a strategic management system due to the dynamic market environment change is excluded. Second, it could not reflect the drastic change in the online environment. This is a natural result. Since the mid-late 2000s, there has been a drastic change in the online environment, especially the influence of social media. This is an environmental change unimaginable at the time of model development, and it is difficult to predict how and when such rapid market changes will be predicted in the future.

Marketing theory and its strategic methods do not remain static, but integrate theoretical and point-of-view views of what is happening in the market and what can happen in the future to achieve a reciprocal effect between theory and practice. I need to. Despite the introduction of the systematic customer asset model of the previous study, the new customer-oriented global customer asset model proposed in this study is to provide a more realistic theoretical basis by presenting a dynamic marketing model.

4) social network assets

In terms of consumer-oriented customer assets, the concept of social is recognized as the most important concept in the fields of economy, society, psychology and management, despite considerable theoretical disputes. Social, that is, social aspects in my field, is hard to define, but confusion arises in the quantitative approach.

A society in a dictionary is a group of two or more individuals and associations based on common interests, beliefs, and understandings that are independent of the political and cultural institutions. It refers to small families, workplaces and schools, the whole society, and part of it, political, economic, civil, labor, and cultural societies.

Interest in the concept of social networks began to grow as social media such as myspace and facebook, and Internet companies such as yahoo and google actively set up social networks. Social networks are nothing new. Concepts such as word of mouth intention and consumer self-congruity have been used to explain social phenomena since the pre-com dot-com period. However, as the communication paths that companies use to spread their products or product information have diversified, the concept of social networks has begun to emerge. The business-to-consumer path by traditional mass communication has been greatly reduced by oral communication, the activity of exchanging consumption information between consumers. In addition, as the exchange of information between consumers through the use of social communication channels, the area of social groups in which consumers operate is expanding. For example, in the past, local communities were often limited to belonging to social groups by offline activities, but the development of information and communication not only spreads the range of social groups exponentially online but also belongs to the groups. The types and numbers of are also becoming diverse. In other words, consumers are reconstructing social units online and expanding them into areas of consumption. Accordingly, a lot of fragmentary studies on brand-ego coordination of psychological and semiotic aspects have been conducted. Recently, the importance of brand-self coordination of social networks has been increasing.

Therefore, in the present invention, the scope of social is defined as an aggregate of individuals, which is an entire social aspect, and the intangible value generated by the relationship between individuals and individuals or groups is defined as social network assets.

A network is a collection of objects and can be called a node that connects them. Newman (2003) categorized networks into social, informational, technical and biological aspects. He defined social networks as people or groups with similar types of contacts or interactions. People form social networks and create links that exchange discourses with each other, which in turn creates an information network (Dwyer 2007). In other words, social networks do not simply exist, but actual drivers such as information are created among people. A good example is word-of-mouth and brand-ego identity.

The word of mouth is a value that has a great influence that is created between the individuals who make up the social network, and the influence of online has widened the scope of the network and its influence. In the sense of the network representing the link, the value can be grasped by observing the movement of messages between individuals along the path. In other words, the value that occurs between the individuals who make up social networks is being expressed by the word of mouth.

The linkage of the network of brand-ego identity is distinct from word of mouth. According to the theory of social identity, consumers create brand-ego coherence by identifying the image between them and the brand. In other words, by purchasing a brand, the identity between the brand and the individual is transmitted. In addition, consumers tend to connect with their images and social groups. In other words, consumers identify brand and self as a means of connecting with social groups, and vice versa.

(1) Oral Elements

The word of mouth is one of the important sources of information in customer decision making because it can affect people's beliefs and expectations of a particular organization. In general, word of mouth is defined as "informal communication between customers about the characteristics of an organization or product." To present concretely the typical characteristics of the oral text, the oral text is verbal communication as the term itself implies. It is in this sense that word-of-mouth differs from non-verbal communication such as written correspondence, memos, and gestures. Second, word of mouth is mostly informal. Word-of-mouth is sometimes official, but it is mostly informal among intimate people, such as friends and relatives. Third, the oral subject (sender) is either a past or present customer who has direct or indirect experience with a particular object, and the object (receiver) is most likely a future customer with availability. Finally, the messages contained in the oral text mainly consist of the evaluation of the dissatisfaction of the specific organization's characteristics, services and products. On the other hand, the intention to act means a clear possibility to lead to the act. Therefore, oral intention can be defined as the possibility of informally conveying an evaluative intention (mainly related to dissatisfaction) about a specific organization's characteristics, services and products, orally to others.

Moreover, online word of mouth is an informal and interactive communication act or process that allows the free and indirect experience or information related to products and services to be communicated between consumers via e-mail or text on the Internet. Recently, testimonial information, in which consumers freely share their experiences and opinions, has become a representative type of online word of mouth.

Testimonial information is classified into positive information and negative information according to the direction. Recently, research has been conducted on the effects of positive Chinese New Year and negative Chinese New Year information on consumers. Oral directionality studies do not provide consistent results, but research is conducted in two directions, negative and negative, affecting consumers' decision-making rather than positive ones. To understand these conflicting results, it is necessary to consider other variables that affect online word of mouth. In other words, the persuasive effect of the word of mouth may vary depending on the characteristics of the consumer. Therefore, it is necessary to consider the moderating effect by consumer personal characteristics.

The word-of-mouth effect on the consumer can be measured by dividing the receiving state and the sending state according to the direction of information. According to the general purchasing decision making process, the information search process is performed before the actual purchase behavior of the consumer occurs. In this case, the consumer is affected by the reference group in addition to the objective product or service information. In other words, word-of-mouth effect occurs as an effect on consumer's decision making. After the actual buying behavior, the consumer evaluates the product or service. At this time, the evaluation has the most influence on the purchase of the next product or service, but it is an important variable that changes the attitude of consumers by affecting the information search process in the purchasing behavior of others. As a result, consumers are influenced by word of mouth and at the same time have a two-way effect on others.

Looking at the flow of previous studies related to word of mouth effect, it can be seen that word of mouth effect is bidirectional. After using the brand, word of mouth as an evaluation of the word is referred to as word-of-mouth intention or word-of-mouth behavior and is a dependent variable according to the preceding factors in the consumer behavior model. On the other hand, the word of mouth influenced by others when a consumer purchases a product or service may be called an effect of word-of-mouth.

Such word-of-mouth that has concomitant characteristics in consumer behavior occurs simultaneously in word-of-mouth effect and word-of-mouth intention in the study of consumer behavior models. The customer asset is the value of the future value of the consumer and is a model for forecasting not only the current purchase behavior but also the future purchase behavior. Therefore, it is desirable to measure the word-of-mouth effect that affects the current buying behavior of consumers and the word-of-mouth intention that affects each individual or other consumers.

(2) brand-ego identity factor

Self-image is the sum of thoughts and feelings about one's own self that the individual considers himself to be an object or an object. In other words, the self image actually builds an integrated image of several egos. For example, a female consumer who has both a home and a job shows not only a self-image as a mother to a child but also a wife, a housewife, a boss and a co-worker, a daughter-in-law, a member of a community such as a friend and a neighbor. It can have different self images. Each self image is not separate but builds an integrated self that forms the overall image of the individual. For these properties, self-image can be implied in one unified sense, in which all properties are organically related. Therefore, self-image can be said to have two-sided, both interactive and interactive. In addition, various self-images influence the selection of products and brands by preferring products or brands that reflect their self-image as purchasing decision-making variables in the consumer purchasing decision-making process. It also affects your self-image. Because people often look at the products or brands that others consume or use, they tend to think that someone is a person. This extension of self-image theory is also called symbolic reciprocity.

When a product and a brand are selected based on symbolic reciprocity, the product and brand are purchased not as an objective entity but as a subjective symbol. And consumers are more concerned about what they mean than what practical functions they have. Therefore, the concept of self can be understood from a multidimensional point of view, and the self can be said to have its own internal judgment and social influence at the same time or separately. In addition, it can be seen that the entity may change dynamically depending on the social consumption situation. The self image is divided into actual individual self image, ideal individual self image, actual social self image, and ideal social self image from a multidimensional perspective that can explain the connection with the social environment.

The actual individual self image is the image that the individual has about himself, and the ideal individual self image is the image he wants to have. The actual social self-image means the belief in how an individual looks by others, and the ideal social self-image is the image that others want to have about themselves. Self-congruity is explained by the process of harmonizing the image between the consumer, the self and the brand. The greater the degree of harmony between the brand and the self image, the more likely the consumer is to maintain a relationship with a specific brand.

Real self-image and real social self-image are the perceptions of how the consumer himself or others see themselves. The consumer tries to maintain self-image by acting to prove self through feedback process. Therefore, if the actual individual self image and the actual social self image match with the brand image, the use of the brand contributes to self-validation.

Social self-theory divides identity into cognitive, evaluative and emotional factors. In other words, from a social point of view, brand-ego coherence refers to a psychological state in which consumers perceive, feel, and add value to the brand. This psychological state that consumers perceive is caused by the existence of the brand, but homogeneity with the brand is a means to express oneself and one who wishes to be projected in society. Therefore, when purchasing a product or service, consumers buy in consideration of how much the product reflects me in addition to the tangible and intangible values that the brand provides. In other words, it is a psychological relationship between consumers and brands, which is different from brand-oriented assets. So brand-ego coherence matches cognitive (how brand and individual identities match), emotional (brand is personally or socially valuable to me), and evaluative (benefit from connecting with a brand) Can be measured by sex.

Here, the corporate sales information input unit 400 performs a function of receiving corporate sales information including information related to a discount rate for each brand of a company and the total number of customers in a brand category.

The customer purchase information input unit 200, customer asset valuation score input unit 300, corporate sales information input unit 400 is connected to the individual customer terminal and the enterprise terminal via the Internet, the homepage, bulletin board specifically provided for measuring customer assets Information from customers and companies, or by forming a community network through blogs, mini homepages, online cafes, or social networks with unspecified people online through social media such as myspace, facebook, twitter, etc. Can be input. Table 1 summarizes the information collected through the customer purchase information input unit 200, customer asset valuation score input unit 300, corporate sales information input unit 400.

division Item Measuring tool Collection Customer Purchase Information Recently purchased brand Recently purchased brands in a brand category (choose from the specified competition) customer Probability of Competitive Brand If there are five competing brands in the industry, indicate the probability of purchasing 100% for each brand Average purchase amount Average amount spent on a single purchase of a product in a category Purchase frequency Frequency of purchases of products in categories per year Customer Asset
Score Value Asset Division Quality, price Brand asset department Awareness, image Relationship asset department Customer treatment, special program Social Network Department Oral Effect, Oral Intention, Brand-Self Congruence (Cognitive, Emotional, Evaluative Congruence) Company Sales Information discount rate Discount rate by brand Enterprise Total number of customers Total number of customers in brand category

Here, the purchase probability matrix element calculating unit 500 performs a function of calculating a purchase probability matrix element for purchasing a brand using the customer asset valuation score received from the customer asset valuation score input unit 300. Meanwhile, the purchase probability matrix calculating unit 500 applies a customer asset valuation score received from the customer asset valuation score input unit 300 to a utility model to calculate a utility value for each brand of the customer. A selection probability calculation unit 520 for calculating a selection probability for each brand by comparing the utility value of each brand of the customer calculated by the utility value calculation unit with the utility value of all brands; And a transition matrix calculator 530 for applying a Markov chain model to the selection probability for each brand calculated by the selection probability calculator to calculate a transition matrix for each brand movement type of the customer.

In order to derive the purchasing probability matrix element, the customer's asset valuation score obtained from the consumer survey is first applied to the utility model to find the utility value for each individual brand. The individual selection probability can be obtained by comparing the utility value of each customer's brand with the utility value of all brands.

This selection probability can be stochastically calculated for each brand movement model of the customer through the Markov chain model. The Markov chain model uses the transition probability to statistically verify the change in probability value from the time t-1 to the time t and the time t + 1. In other words, it is possible to explain the transition form of the past purchase probability of the individual customer and the current purchase probability. Finally, the probabilities of purchasing probabilities for each brand of each customer, calculated through the Markov chain model, are applied to the customer lifetime value model.

Hereinafter, a methodology for calculating a purchase probability matrix element by using a customer asset valuation score for a consumer-oriented customer asset driver applied in an embodiment of the present invention will be described in detail.

5) Utility Model

Utility theory was introduced to score individual customer drivers of customer assets. The utility model included brand inertia as well as individual driver scores.

_{Incorporating the} above model, U _ijk represents the utility of brand k of customer i who recently purchased brand j. Here, the dummy variable LAST _ijk is 1 when j = k, and O otherwise. X is the row vector of the driver. The utility model is shown in the following equation.

Here, U _ijk = utility of k brand for i-customer who recently purchased j brand, β _0k = logit regression coefficient for inertia, β _ik = row vector of logit regression coefficient for driver, ε _ik = residual term. .

The utility of individual levels leads to transition matrices of individual levels. Each row of the transition matrix is assumed by a recently purchased brand, so each row represents a different utility. That is, the first column assumes that the first brand was most recently purchased, and the second column assumes that the second brand was most recently purchased. Since the effects of brand inertia appear only when they are repurchased, the utility of each row gives different results.

At the same time as the polynomial logit model, the choice probability model for customer i is

P _ijk = Pr [individual i chooses brand k ^* , given that brand j was most

     recently chosen]

The results of the utility model at the individual level can lead to the lifetime value of the customer at the individual level because it creates an individual-level brand transition matrix.

6) Transition Matrix Model

The most fundamental reason for using transition matrices in this model is to maintain the customer mobility model. The first consideration in calculating lifetime customer value is to define the relationship with the customer. In the preceding study, the relationship between brand and customer is regarded as a contract situation, and classified into contractual or non-contractual situation. The contractual situation applies the probability of customer retention to the model from the time an individual customer begins to maintain a relationship with a particular brand and continues from the end. In this case, it is assumed that the lost customer does not come back because the relationship with the customer is not once renewed.

Non-contract situations, on the other hand, apply the survival probability of the customer to the model. Instead of ending the relationship with the brand, the customer remains alive as a probability of purchase. The Markov Chain Model through Transition Probability is a model for customer retention, churn, and reacquisition, which is a non-contractual situation because it includes not only the probability of purchase for a particular brand, but also the factors that move between competing brands. To implement the consumer-brand relationship.

B _it is the 1ⅹJ row vector, and B _ijt is the probabilities matrix for purchasing the brand j when the customer i makes a purchase at time t.

B _it is the 1ⅹJ row vector, and B _ijt is the probability matrix element where the customer i _buys the brand j at the time t.

7) Principal Component Analysis Model

Principal component analysis is needed to reduce the variables obtained through surveys from consumers to the dimension of customer asset drivers. The independent variables involved in principal component analysis are the customer asset driver and the LAST variable. In addition, the principal component analysis is carried out firstly to block the problem of multicollinearity between variables. Second, coefficients calculated through principal component analysis can be measured as coefficients of independent variables.

This model measures competitive factors between brands in the market when measuring customer lifetime value. This competitive factor consists of the relative influence of the drivers perceived by consumers, namely, value assets, brand assets, related assets and social network assets, and serves as a factor in determining the selection probability of a competitive brand.

Independent variables for principal component analysis are driver variables and LAST variables. The vector X _ijk is represented by the combination of the previous j-brand purchase and the next k-brand purchase of the individual customer i. Rotate the varimax to extract the principal component of X _ijk , which can be used as a tool for dividing dimensions and making decisions. F _ijk represents the rotated factor matrix. In other words, F _ijk is a combination of F _{value assets} , F _{brand assets} , F _{value assets} , and F _{social network assets} .

When the utility model is calculated in the rotated component dimension, it can be expressed as follows.

Is the coefficient vector.

Here, the customer lifetime value calculator 600 calculates a customer lifetime value using the customer purchase information, the company sales information, and the purchase probability matrix element.

The customer asset model translates the future cash flows that a company's current or potential consumers bring to the company at its present value. Firstly, customer asset value can be obtained by multiplying customer lifetime value by the total number of customers at the individual customer level. At this time, the Customer Lifetime Value (CLV _ij ) for each customer (i) and brand (j) is calculated by the following equation.

Where T _ij is the number of purchases made by the customer i during a specific period, d _j is the discount rate of the brand j, f _i is the average number of purchases of the customer i over a period of time (for example, in years), and V _ijt is the time at which the customer i Π _ijt is the expected profit per j-brand purchasing unit at time t of customer i, B _ijt is the brand j purchase probability matrix element of customer I at time t, and t / fi is discounted according to the time of purchase. Means that.

Here, the customer asset calculator 700 functions the customer asset using the customer lifetime value. The customer lifetime value at the individual customer level is multiplied by the total number of customers, yielding Customer Equity (CE).

In order to calculate customer assets as shown in the above formula, customer lifetime value at the individual customer level should be calculated first. Here, customer lifetime value represents the average value of individual levels for all customers of a brand. In general, the customer life value calculated by using the customer data is calculated as the data for each customer by individually calculating the value of each customer, but in this study, the final purpose is to know the customer's customer assets. The average value of customer lifetime value is sufficient.

2 is a flowchart illustrating a method for measuring customer assets according to an embodiment of the present invention.

In addition, the customer asset valuation score input unit 300 receives a customer asset valuation score for the customer asset unit consisting of a value asset unit, brand asset unit, relationship asset unit, social network unit (S200).

In addition, the customer asset driver evaluation score input unit 300 receives a customer evaluation score for the customer asset driver composed of a value asset driver, a brand asset driver, a relationship asset driver, and a social network driver (S200).

Next, the purchase probability matrix element for purchasing the brand is calculated using the customer asset driver evaluation score received from the customer asset evaluation score input unit 300 (S400).

Next, a probability matrix element for purchasing a brand is calculated using the customer asset driver evaluation score received from the customer asset driver evaluation score input unit 300 (S400).

Subsequently, the customer purchase information input to the customer purchase information input unit 200, the company sales information input to the company sales information input unit 400, and the purchase probability matrix calculated by the purchase probability matrix element calculation unit 500. Using the elements to calculate the customer lifetime value (S500).

Finally, the customer lifetime value of the individual customer calculated by the customer lifetime value calculation unit 500 is multiplied by the total number of customers to calculate a customer asset (S600).

Referring to FIG. 3, first, a utility value for calculating a utility value for each brand of a customer is calculated by applying a customer asset valuation score received from the customer asset valuation score input unit 300 to a utility model (S410).

Referring to FIG. 3, first, a utility value for calculating a utility value for each brand of a customer according to a driver is calculated by applying the customer asset driver evaluation score received from the customer asset driver evaluation score input unit 300 to a utility model. S410).

Thereafter, the utility value for each brand of the customer calculated by the utility value calculator 510 is compared with the utility value of all the brands to calculate a selection probability for each brand (S420).

Next, by applying the Markov chain model to the selection probability for each brand calculated by the selection probability calculator 520, a transition matrix for each type of brand movement of the customer is calculated (S430).

On the other hand, the customer asset measuring method according to the present invention can also be implemented by a computer through a program recorded on a computer-readable recording medium. Computer-readable recording media include all kinds of recording devices that store data that can be read by a computer system. The computer readable recording medium can also be stored and executed as distributed readable code distributed over a computer system connected via a computer communication network.

100: Customer Asset Measurement System
200: customer purchase information input unit
300: customer valuation score input unit
400: corporate sales information input unit
500: purchase probability matrix element calculation unit
510: utility value calculation unit
520: selection probability calculator
530: transition matrix calculation unit
600: Customer Lifetime Value Calculator
700: Customer Asset Calculator

100: Customer Asset Measurement System
200: customer purchase information input unit
300: customer asset driver evaluation score input unit
400: corporate sales information input unit
500: purchase probability matrix element calculation unit
510: utility value calculation unit
520: selection probability calculator
530: transition matrix calculation unit
600: Customer Lifetime Value Calculator
700: Customer Asset Calculator

Claims (15)

A customer purchase information input unit for receiving customer purchase information including information related to a purchase brand, an average purchase amount, and a purchase frequency;
A customer asset valuation score input unit for receiving a customer asset valuation score for the customer asset unit including a value asset unit, a brand asset unit, a relationship asset unit, and a social network unit;
A corporate sales information input unit for receiving corporate sales information including information related to a discount rate and total number of customers;
A purchase probability matrix element calculating unit for calculating a purchase probability matrix element for purchasing a brand using the customer asset evaluation score received from the customer asset evaluation score input unit;
A customer lifetime value calculator for calculating a customer lifetime value using the customer purchase information, company sales information, and purchase probability matrix elements; And
And a customer asset calculating unit configured to calculate a customer asset by multiplying the customer lifetime value by the total number of customers.
The method of claim 1,
The social network unit customer property measurement system, characterized in that it comprises a word-of-mouth element and brand-ego identity element.
The method of claim 2,
The word of mouth element includes a word of intention item as an evaluation of the after use of the brand and a word of mouth effect item that is influenced by others when the customer purchases the word of mouth,
Wherein said brand-ego identity component comprises cognitive consistency, emotional consistency, and valuation consistency.
The method of claim 1,
The customer lifetime value calculator calculates the Customer Lifetime Value (CLV _ij ) for individual customers (i) and brands (j) by the following equation.
<Expression>

here,
T _ij is the number of purchases made by customer i over a specified period of time,
d _j is the discount rate of brand j,
f _i is the average number of purchases (i.e. yearly) of customer i over a period of time,
V _ijt is the quantity to purchase the j brand while the customer i is at time t,
π _ijt is the expected profit per unit of brand j purchased at time t of customer i,
B _ijt is the brand j purchase probability matrix element of customer I at time t,
t / fi means discounted at the time of purchase.
The method of claim 4, wherein
The customer asset calculating unit calculates the customer asset using the customer lifetime value which is an average value of individual levels for all the customers of the brand.
The method of claim 1,
The purchase probability matrix element calculating unit
A utility value calculator for calculating a utility value for each brand of a customer by applying the customer asset valuation score to a utility model;
A selection probability calculation unit for calculating a selection probability for each brand by comparing the utility value of each brand of the customer with the utility value of all brands; And
And a transition matrix calculator configured to apply a Markov chain model to the selection probability for each brand calculated by the selection probability calculator to calculate a transition matrix for each type of brand movement of the customer. .
The method according to claim 6,
The purchase probability matrix element calculating unit
And a principal component analyzer configured to calculate principal component coefficients for determining relative influences of the value asset, the brand asset, the relationship asset, and the social network.
In the customer asset measurement method using the customer asset measurement system,
Receiving, by the customer purchase information input unit, customer purchase information including information related to a purchase brand, an average purchase amount, and a purchase frequency;
Receiving, by the customer asset valuation score input unit, a customer asset valuation score for the customer asset unit consisting of a value asset unit, a brand asset unit, a relationship asset unit, and a social network unit;
Receiving, by the corporate sales information input unit, corporate sales information including information related to a discount rate and the total number of customers;
Calculating, by a purchase probability matrix element calculating unit, a purchase probability matrix element for purchasing a brand using the customer asset valuation score;
Calculating a customer lifetime value using the customer lifetime value calculation unit using the customer purchase information, corporate sales information, and purchase probability matrix elements; And
And a customer asset calculator calculates a customer asset by multiplying the customer lifetime value by the total number of customers.
The method of claim 8,
And wherein the social network driver comprises a word of mouth element and a brand-ego identity element.
The method of claim 9,
The word of mouth element includes a word of intention item as an evaluation of the after use of the brand and a word of mouth effect item that is influenced by others when the customer purchases the word of mouth,
Wherein the brand-ego consistency factor comprises cognitive consistency, emotional consistency, and valuation consistency.
The method of claim 8,
The customer lifetime value calculator calculates a customer lifetime value (CLV _ij ) for each customer (i) and brand (j) according to the following equation.
<Expression>

here,
T _ij is the number of purchases made by customer i over a specified period of time,
d _j is the discount rate of brand j,
f _i is the average number of purchases (i.e. yearly) of customer i over a period of time,
V _ijt is the quantity to purchase the j brand while the customer i is at time t,
π _ijt is the expected profit per unit of brand j purchased at time t of customer i,
B _ijt is the brand j purchase probability matrix element of customer I at time t,
t / fi means discounted at the time of purchase.
12. The method of claim 11,
Wherein the customer asset calculator calculates the customer asset using the customer lifetime value, which is an average value of individual levels of all the customers of the brand.
The method of claim 8,
Mechanism cost in the purchase probability matrix element calculation unit,
Calculating, by a utility value calculation unit, the utility value for each brand of the customer by applying the customer asset valuation score to the utility model;
Calculating, by a selection probability calculating unit, a selection probability for each brand by comparing the utility value of each brand of the customer with the utility value of all brands; And
And a transition matrix calculation unit calculating a transition matrix for each type of brand movement of the customer by applying a Markov chain model to the selection probability of each brand.
The method of claim 13,
The apparatus cost in the purchase probability matrix element calculation unit,
And a principal component analyzing unit calculating a principal component coefficient for determining a relative influence of the value asset unit, the brand asset unit, the relationship asset unit, and the social network unit.
A non-transitory computer-readable recording medium having recorded thereon a computer program for executing the method of measuring a customer asset according to any one of claims 8 to 14.

Images