Database Marketing and Its Measurement of Success - Dissertation Example

RESEARCH PROPOSAL: IN A NON-CONTRACTUAL SETTING, ARE LONG-LIFE S MORE PROFITABLE THAN NEW SHORT-LIFE S? Submitted by: College Name: Word Count: 3619 Table of Contents Table of Contents 1 Introduction 2 Research Questions and Objectives 2 Literature Review 3 Hypotheses 5 Methodology 6 Theoretical Model 6 Research Nature 8 Target Population 9 Sampling Technique and Size 9 Data Collection and Classification 9 Empirical Data Analysis 10 Semi-Structured Interviews 11 Validity, Reliability and Generalisability of the research 11 Reflections 12 Ethical Considerations 12 Other ISSUES 13 Timeline of the progression of dissertation 14 Conclusion 15 References 16 Introduction The basic principle of relationship marketing is the premise that firms should maintain long-term relationships with their customers. There has been enough conceptual evidence for this argument. Reichheld and Sasser (1990) state that, “Customer defections have a surprisingly powerful impact on the bottom line. As a customer’s relationship with the company lengthens, profits rise.” However, the idea has not been tested empirically in a non-contractual setting. The tenant that the loyal customers are always more profitable is a “gross simplification” (Dowling and Uncles, 1997). Many reasons have been put forward to support their position. Old customers always expect value-added relationships and bonuses for their continual purchases in order to buy more (Mohs, 1999). On the other hand, short-term and new customers do not have any reluctance on buying a product. This calls for a sound empirical research that can test the profitability of long-term customers as compared to short-term customers. Such a research will help organizations to understand the profitability of various categories of customers and help them focus on the correct ones. The researcher hopes of getting some useful insight into the basics of relationship marketing as a result of this research. Research Questions and Objectives As stated earlier, there is a lack of empirical research confirming the long-term profitability of customers in a non-contractual setting. This research aims to fill this gap by performing Lifetime analysis for a customer-vendor relationship in a non-contractual setting. Examples where this setting is prevalent are: department store purchases or mail-order purchases in the catalog and direct marketing industry. The objective of this research is to assess the profitability of long-term customers as compared to short-term customers in a non-contractual setting. Higher profits from a customer may arise because of the following reasons: lower costs or higher prices for long-term customers. The author wishes to find out the answers to the following questions: What is the nature of relationship between lifetime-duration and profitability? Do profits of a customer increase as time progresses? Are costs of serving long-term customers less than that of serving new customers? Whether long-term customers pay higher prices or not? Literature Review The concept of relationship marketing has attracted a lot of research in recent times. A stronger customer relationship management can be a distinct source of competitive advantage for companies (McKenna, 1993). The basis for this premise is that the long-term relationships are more profitable than short-term relationships. The higher profitability is credited to greater exchange efficiencies as a result of customer retention economics (Sheth and Parvatiyar, 1995; Sheth and Sisodia, 1995). Retained customers produce higher revenues and margin per customer as compared to lows or newer customers (Best, 2000). This implies an increasing trend of profits over time. Reichheld and Teal (1996) argue that the profits per time unit of a customer increase as time progresses. Another common claim is that serving a long-term customer involves less cost than serving short-term customers. Customers who are involved and have been retained for a long relationship are relatively low maintenance (Blattberg and Deighton, 1996). Long-term customers result in low marketing costs and therefore have higher contribution margins (Wang and Spiegel, 1994). The problem is that all these evidences have been proved in contractual setting. Dowling and Uncles (1997) claim, that these findings can not be generalized. They point out that keeping a customer for a longer period of time may result in costs such as reward programs, loyalty programs, and so on. These costs can be significant to the firm (Mohs, 1999). Hence, it can’t be concluded that the costs of serving long-life customers is less than that of short-life customers. Empirical research will give us a useful insight into this premise. There are various factors that play an important role in the retention of a customer (Blattberg and Deighton 1991). Some of these factors are the customer satisfaction, the competitive environment. It shall be noted that both the factors affect the relation of the other with customer retention. In specific case, different competitive environment affect the relation between customer satisfaction and customer retention (Jones and Sasser 1995). For example, in an environment where there is little competition, the customer has not many choices and will not defect even at levels of dissatisfaction. On the other hand, high satisfaction can’t guarantee retention of the customer in environment of high competition. Various methods have been proposed to test the profitability of long-term customers. The most basic of them is a simple correlation between the lifetime and profitability of various customers. Although useful, this measurement gives only a static representation. Longitudinal approach can be used to make correct inferences about customer behaviour with the development of relationship (Fournier, Dobscha, and Mick, 1998). Useful findings call for the development of an accurate model for measuring customer life-time, profitability, and defection. The measurement of customer life-time is more difficult in case of a non-contractual setting as the purchases are made at the discretion of the customer. Based on a customer’s activity in the past, a firm may know of the active and inactive customers. This model suggested by Schemitllein, Morrison and Colombo (1987) and then later modified by Schemitllein and Peterson (1994) has found sufficient support in the marketing literature. Their model which is based on the Negative Binomial Distribution (Pareto Model) enables firms to answer this key question: Which individual customers are most likely to be active customers in the future? With the knowledge of a customer’s probability of being ‘alive’, and a cut-off probability level, we can determine a reasonable measurement of the customer’s lifetime. Blattberg and Deigbton (1991) propose a graphical model that suggests that the customer bases for a firm shall be divided into various categories depending upon their behaviour and attitude. This model is as shown in Figure 1. The two major dimensions-lifetime and revenues are considered appropriate by many managers in the field of relationship marketing. Most of the research done above caters to the case of contractual settings. There has been no substantial evidence for non-contractual settings. This research attempts to cover this gap. The way firms treat their short-term and long-term customers have a lot of impact on their profitability. Both long-term and short-term customers shall be treated differently (Garbarino and Johnson, 1999) Hypotheses On the basis of the brief review of the literature and the objectives of the study, the author proposes to test the following hypotheses for the test: H1: The relation between lifetime and profitability is positive H2: The profits of a customer increase with the progress of time H3: The costs of serving long-term customers are less than that of small-term customers. H4: Long-term customers pay higher prices as compared to small-term customers. Methodology The methodology chosen for the research shall be according to the research objectives and the research questions that need to be answered. The first step in the methodology will be a literature review. The literature review will be aimed at gaining thorough background knowledge about the area of study. The author will try to study past models and research that will act as a basis for the present study. Based on some preliminary literature review, the author has studied the theoretical model that will be used in the study. Theoretical Model Measurement of Customer life time for non-contractual setting The measurement of the customer life-time for non-contractual setting is based on the NBD/Pareto Model. This model is based on the premise that purchase is a random activity and defection is not observable directly. The model returns the probability of a customer being alive. This probability depends upon the customer’s purchase history through the number of purchases and the time t at which the most recent purchase was made. The equation for determining the probability of a customer being alive is (Schemitllein and Peterson (1994). Where, ; ; F (a1. b1; c1; z1) is the Gauss hyper-geometric function; r, α, β and s are model parameters x is the number of purchases t is the time of the most recent purchase since the trial T is the time since trial The continuous probability as derived from the Schmittlein and Peterson’s model can be extended to a dichotomous alive/dead measure. On the basis of a person’s time of birth (the time of first purchase) and a cut-off threshold level, we can get to know the time when a customer has left the relationship. Thus life-time of the customer will be then the difference between the cut-off value and the time of birth of the customer. This lifetime can then be used to measure the profitability of the customer. Analysis of the method: This model is based on the premise that the time of birth of a customer is known. This can be assumed because of the popularity of customer databases in organization (Petrison, Blattberg, and Wang 1997). The parameter estimation is the essential part of the life-time calculation. The parameters can be estimated using the bootstrap method of moment estimation on the sample. The bootstrap method will give us an increased understanding of parameter sampling properties. Another important factor in the life-time calculation is the calculation of threshold-value. An intuitive value of the cut-off probability is 0.5 (Sharma, 1996). The value has been used by Helsen and Schmittlein (1993) in the prediction of purchase events. To confirm if this value is appropriate, the research proposes to perform a sensitivity analysis. Calculation of Profit Profit can be calculated as the net-present value of the profit of an individual customer basis for the period of 36 months using the equation (Berger and Nasr, 1998) Where, LTπi = individual net-present lifetime profit of the customer i for the period of 36 months. GCti = Gross contribution in month t by customer i. Cti = the mailing cost for the customer i in month t 0.125 is the discount rate on the basis of assumed rate of 15%. The GCti can be calculated on the basis of the revenue of each household for every month for the observation period. The gross contribution is assumed to be 30% of the revenue. Cti which indicates the mailing cost for the customer i in time t include the cost of catalogue production, letter shop, and mailing costs. Research Nature The main objective of this research is to find if long-term customers are more profitable than short-life customers or not. These characteristics make it a descriptive research. The research starts with a review of the literature, and ends with the comparison of the empirical findings to the past research. Since the objective is not to formulate any theories, deductive approach will be used to complete the research. The study is longitudinal in nature as it covers a period of time. Target Population The research is meant to be examining the profitability in a non-contractual setting. Examples of non-contractual settings include department store purchases or mail-order purchases in the catalog and direct marketing industry. Of these, catalogue marketing is a very common form of marketing. It involves selling through catalogues that are mailed to the list of customers who have purchased something at the store previously. The target population is the set of people who have purchased something in the recent past from catalogue retailer. Sampling Technique and Size The author will collect the data from a catalogue retailer for the period of three years. The sampling will first be of all those customers who have made their first purchase in the three year window. Of these, simple random sampling will be used so that every element in the population will have an equal chance of being selected. From the retail store’s past sales data, the average number of customers added each year is found to be 2800. Therefore over a period of three years, the store can be expected to add 8400 customers. The sample size therefore can be estimated to be around 8500. Data Collection and Classification The researcher will be using the sales data from a catalog retailer for a period of three years. The entire purchase history of a household in the time window will be referred to as an observation. The sample elements are such that the first purchase of the element in within the 3 year window. The data will be collected from the database available with the retailer. The retailer has a database containing the information of each customer that is added to its database in the last 3 years. The entire sample will be divided into two groups. Cohort 1 will be the sample of households that have started shopping right from the beginning of the time window. Cohort 2 will be group of households that has started shopping one after the beginning of the time window. Cohort 2 will serve as validation sample for the results observed with Cohort 1. Empirical Data Analysis The researcher intends to use a segmentation scheme based on behaviourally different subgroups as suggested by Blattberg and Deighton (1991). We consider the profit as the dependent variable and lifetime duration as the independent variable. We use median lifetime duration as it is a better representation of the lifetime as compared to mean lifetime (Collet, 1994). Thus the entire dataset is categorized into a shorter and a longer ‘lifetime’ and smaller and larger ‘revenue’. Our special interest would be on customers who fall in Segment 2 (customer with long lifetime and small revenues) and Segment 3 (customers with small lifetime and large revenues). To test if the profits from a customer increase over time, we analyze the sign of the slope coefficient between profits and time. A positive slope indicates that profit do increase over time. A linear regression equation that is applicable in the setting is: Where, as is the constant term (the interception point of the line) s represents the segment t is the month bs is the regression coefficient. To test the hypothesis that long-term customers have lower costs than that of short-term customers, we compute the ratio of promotional costs to revenues for each household for each month. The promotional costs include the cost of printing and mailing the promotion catalogs and other promotional costs such as advertisements in the newspapers, billboards etc will be divided by use of prorated measures for each customer. The mean promotional cost for each household will be calculated for each segment and then the costs would be compared across segments to determine if the costs of long-term customers are less than short-term customers. Similarly, we will be calculating the average prices paid by each customer for each segment and then compare these averages across segments to determine if long-term customers pay higher prices or not. The tools that will be used for data analysis and formatting will be MS Excel and SPSS 16.0. Semi-Structured Interviews After the completion of the literature review and the quantitative analysis, the author proposes to conduct semi-structured interviews. The author expects to interview certain managers in the direct marketing industry as well as some general customers. The semi-structure interviews will be in the form of an open questionnaire that the author intends to conduct on his own. The reason for conducting these semi-structured interviews is that the author expects to find certain qualitative intelligent inputs that can’t be derived from the quantitative analysis alone. The quantitative analysis will then be used to explain the possible reasons behind the findings. Validity, Reliability and Generalisability of the research The author has tried to validate the results. The classification in two cohorts was done for the same reason. The results from cohort 1 were validated using the data in cohort 2. The data is collected from the past purchase data of a genuine retail store. The data gathered will be reliable. However, the author will try to verify each data set and will remove data that appears false. In order to find out if the findings of the research can be generalized or not, the author will try to find out the possible explanations of the findings. Based on the explanations, the author expects to find the settings where the findings can be generalized. The author also expects to find suitable extensions of the study on the basis of these explanations. Reflections The prime objective of this research is to test the basic principle of relationship marketing that long-term customers are more profitable than short-life customers. The researcher has tried to take extreme caution regarding various aspects of the research. Ethical Considerations The data collected for the study consists of historical details about each customer. The details collected will be the purchase history, items bought, the monetary value of items bought and the time. All this is very sensitive information and can be used by spam marketers. The author will ensure that the information is not leaked. It will be ensured that the data collected is used for the purpose of this research only. Another ethical issue that arises as a result of the research design is the security of the data. The customer data collected from the retailer about the customer is sensitive and can be used for some unethical purposes if it falls into the hands of the competitors. So, the researcher will protect the data from such loss. Same is the case with the data about the costs of the promotion and catalogues. The author will make sure that none of the data is used for purposes that are unethical in any nature. Another issue that arises because of the research design is the violation of copyright rules by authors who have done prior research in the field. The methodology of the research consists of the review of literature of previous studies in the same field. The author respects the contribution made by previous scholars and will make sure that every author whose work is used is referenced properly using appropriate referencing style. Semi-structured interviews by the author also raise ethical issues. The interviewees will be the managers of direct marketing industry and therefore there are chances that they might divulge details that are trade secret and can be used in unethical manner by competitors. To remove this issue, the author intends to conduct the interview in private and will ensure that the responses are not accessible to others and used by others for unethical purposes. Other ISSUES Other issues that may arise in the research are that of data collection. The data that needs to be collected is confidential in nature and it will be hard to get the data. However, the researcher expects to have collaboration with some retail store. The store will be told about the objectives of the study and the advantages of the study that can be used by the store. The author will try to convince the store managers that the findings of the research can be used by the store to increase their profitability. Another issue is that of conducting the semi-structured interviews of store managers. Store managers might be busy and may not be able to devote much time towards the interview. The author expects to remove this issue by explaining the importance of the study and will inform them the value of their contribution. There are many limitations to the design of the study chosen by the author such as a time window of just 3 years, analysis of data from just one retail store, inability to integrate customer’s behaviour, satisfaction levels and attitudes in the framework and so on. The author will try to remove these limitations in the future research. These limitations and the possible area of improvement will be mentioned in the final dissertation proposal. Timeline of the progression of dissertation The author expects to start working on the dissertation from 15th October 2009. The dissertation will take about 233 days. Buffer Period of 14 days is kept to cover any kind of unforeseen circumstances. MS Project will be used to keep track of the activities, their start time, and delay, if any in the activities. The timeline for the complete process of preparing this dissertation has been shown below using a Gantt chart. Most of the resources will be required for the data collection and analysis part. For collection of the customer data from the retail store, the author proposes to formulate a data collection plan and format. The author will also prepare the brief outline of the semi-structured interviews. Other resources that will be required include access to scholarly journals that will be useful during the literature review part. Some of the tools that will be required for the dissertation are MS Word (for preparing the dissertation), MS Project (for maintaining the project schedule), and SPSS & MS Excel (for Data collection, modification and Analysis). Conclusion The research proposal is a brief outline that will act as the basis for the actual dissertation research. It contains the details of various steps, the issues that may arise in conducting the research and the possible ways to do the research. However, it shall be noted that the research design and methodology formulated in the research proposal is not the end point. The research design and methodology may also be changed if the author finds certain problems in the Literature review step of the research. Based on the research proposal, the author expects to complete the proposal within the given time frame and achieve the objectives of the research. The author expects that the findings from the research enable him to reach to a conclusion that is beneficial to the direct marketing industry, the retail stores and the area of relationship marketing. References Best, Roger J. (2000). Market Based Management. Upper Saddle River. NJ: Prentice Hall. Blattberg, Robert C. and John Deighton (1991). "Interactive Marketing: Exploiting the Age of Addressability." Sloan Management Review, 33 (I), 5-14. Collett, D. (1994). Modeling Survival Data in Medical Research. London: Chapman and Hall. Dowling, Grahame R. and Mark Uncles (1997). "Do Customer Loyalty Programs Really Work?" Sloan Management Review, 38(summer), 71-82. Fournier, Susan, Susan Dobscha, and David G. Mick (1998). "Preventing the Premature Death of Relationship Marketing," Harvard Business Review, 76 (January/February), 42-51. Garbarino, Ellen and Mark S. Johnson (1999). The Different Roles of Satisfaction, Trust, and Commitment in Customer Relationships," Journat of Marketing. 63 (April). 70-87. Helsen, Kristiaan and David C. Schmittlein (1993). "Analyzing Duration Times in Marketing: Evidence for the Effectiveness of Hazard Rate Models." Marketing Science, 11 (Fall). 395-414. Jones, Thomas and W. Earl Sasser Jr. (l995)."Why Satisfied Customers Defect," Harvard Business Review, 73 (November/ December). Pp. 88-89. McKenna, Regis (1993). Relationship Marketing: Successful Strategies for the Age of the Consumer. Boston: Harvard Business School Press. Mohs, Julia (1999). "Frequency Marketing." Retail Report, 12 (4), 3. Petrison. Lisa. Robert C. Blattberg. and Paul Wang (1997). "Database Marketing: Past. Present and Future," Journal of Direct Marketing. 11(4), 109-23. Reichheld. Frederick, and Earl W. Sasser (1990). "Zero Detections: Quality Comes to Services," Harvard Business Review, 68 (September/ October). Pp 105-11. Reichheld. Frederick and Thomas Teal (1996). The Loyalty Effect. Boston: Harvard Business School Press. Schmitilein. David, Donald G. Morrison, and Richard Colombo (1987). "Counting Your Customers; Who Are They and What Will They Do Next?" Management Science, 33 (January). 1-24. Schmitilein. David, and Robert A. Peterson (1994). "Customer Base Analysis: An Industrial Purchase Process Application." Marketing Science, 13 (1). 41-67. Sharma, Subhash (1996), Applied Multivariate Technique.’ New York: John Wiley & Sons. Sheth, Jagdish N. and AtuI Parvatiyar (1995). "Relationship in Consumer Markets: Antecedents and Consequences," Journal of the Academy of Marketing Science, 23 (4), 255-71. Sheth, Jagdish N and Rajendra Sisodia (1995). "Improving the Marketing Productivity." in Marketing Encyclopedia: Issues and Trends Shaping the Future. Jeffrey Heilbrunn, ed. Chicago; American Marketing Association/NTC Publishing, 217-37. Wang, Paul and Ted Spiegel (1994). "Database Marketing and Its Measurement of Success; Designing a Managerial Instrument to Calculate the Value of a Repeat Customer Base." Journal of Direct Marketing, 8 (2). 73-81. Read More