Marketing Data Analysis 

Marketing Data Analysis 

INTRODUCTION

Globally, customer relationships play a vital role in the growth and development of companies, enterprises, and all business organizations regardless of their size- either small scale or large scale. It has been proven that the mode and how communication is being rendered out to the customers by these business enterprises determine the strength of these relationships. Significantly, business organizations are encouraged to create and maintain long-term relationships with their customers so as to have a brighter future for the organization (Chen & Popovich, 2003). To measure, develop, track, and maintain this relationship, customer base analysis is conducted to primarily focus and analyze the time interval and how different customers purchase and make transactions basing on a specific product. Consequently, Chen & Popovich (2003) agree that this has been promoted by the fast-developing technology which has developed the customer analytics software together with the customer relationship management software hence easing the process.

Moreover, customer analysis encourages increased productivity across the business organization since the software vividly shows the most probable product likely to be purchased by different customers (Storbacka et al, 1994). This, in turn, motivates the sales team to focus on the right and productive clients who value and are willing to purchase certain products and services. Majority of these software models rely on the previous transactions of the customers thus giving the most probable predictions. However, for the business organizations to exhibit continuous growth not only are the sales put in the first priority but also the quality of services being delivered, which is done by the quality assurance team, and customer satisfaction. Nowadays the customers’ satisfaction can be accessed by asking the customers perception of their services either through short text messages or email. Furthermore, suggestion boxes are availed so as to get the views and complaints from the customers.

Undoubtedly, Knox & Van (2014) affirms that these complaints create the platform of the turning point of any company or organizations’ relationship with the customers either towards the right or wrong direction. Managers of these organizations are advised to employ brilliant personnel to handle these complaints because they determine the subsequent customer behavior. For instance, extraordinary conclusions from the information conveyed help in turning customers into loyal ones leading to concurrent purchasing of products and services while ineffective handling of the information may lead to loss of some customers to other companies (Storbacka et al, 1994). Therefore, the aim of this project is to study how non-transactional elements such as customer recovery and complaints can be incorporated into customer base analysis and weigh their effect on the probability of a customer quitting the purchase of the company’s products.

Apparently, according to Knox & Van (2014), the non-transactional elements, customer complaints, and recovery have contributed to a great percentage in assisting majority businesses in making progress. In the light of the complaints made by the customers, we can see that it is easier to analyze them and come up with an effective remedy so as to avoid customer churn since they are much related. Besides, these complaints are boosted by previous customer experiences (Knox & Van, 2014). The customer analytic system is manipulated to allow room all these complaints and recovery options. Their effects can be felt far and wide to the extent of long-term impacts which consequently determine customer satisfaction. Clearly, failure to act upon customer complaints leads to consecutive drop out of the customers from the customer base and leaving those who can give the company another chance.

LITERATURE REVIEW

The building and rebuilding of trust and loyalty towards customers is accompanied by lots of challenges and hiccups throughout the journey towards success. Majority of the surveys which have been conducted in the recent past have justified that customer satisfaction is key in building strong long-lasting customer relations. Also, a recent study reveals that above 80% of satisfied customers have the great willingness of doing businesses with any company so long as they are provided with a positive experience (Gustafsson et al, 2005). Pursuing this further, customer complaints determine the satisfaction of the customers as well as how the process of customer recovery is going to be effective. Most of the complaints arise from poor services, poor customer service or failure to meet the customers’ expectations based on the actual services which are provided at different stages.

According to Bora et al (2015), both small and large-scale business enterprises globally have at one-point experienced customer churn due to the fact that retaining customers is one of the most critical challenges in any business. Furthermore, if the rate of customer churn is extremely high in a business it becomes cumbersome in coaxing consumers to purchase products and services from the enterprise. Also, the remaining customers lose trust and confidence in the business. For example, a customer churn of 9% means that a company is losing 9% of its customers annually. Consequently, Kisioglu & Topcu (2011) affirm that the reduction of profits and increased loses, the emergence of a bad reputation of the company, reduction in the company’s growth rate and the future validation of the business and also decreased office morale are some of the effects of customer churn.

Basically, the objective of this report is to discuss in depth the non-transactional elements in the business world which are inclusive of customer complaints and recovery in TESCO PLC. A data collection from TESCO PLC, a multinational supermarket which is situated in the United Kingdom. This chain store has emerged the largest UK’s supermarket chain due to the introduction of the customer loyalty scheme in the mid 90’s which assists in the customer base analysis. The initiative is accompanied by loyalty cards which aid in the assessment of customers’ purchase behavior through the data which is collected every time the customer shops at the supermarket. In addition, the company offers different categories of products such as liquor, meat, general merchandise, apparel, deli, dairy, fresh produce, and bakery products. A random sample of about thirty thousand customers (20,000 training set and 10,000 test set) from the year of 2015 with different types of variables was obtained. The given data will help in the analysis of the main objective of this report.

The report integrates analytical tasks whereby models are constructed to predict the customer churn using binary classification trees such as C&R tree uses metrics related to the confusion matrix to evaluate the supermarket’s performance. Also, the performance of the constructed model is evaluated using the Recency, Frequency, and Monetary (RFM) method. Moreover, discussion of the methodology used and the results obtained are well elaborated in the report as well as some conclusions and recommendations.

METHODOLOGY

In this situation of customer churn prediction, one of the methodologies used is the Classification and Regression (C&R) tree node which is an example of the binary classification trees. This type of methodology incorporates the use of recursive partitioning to divide the training records (20,000 training set) into smaller segments with the same field values. It majorly deals with categorical predictor variable (classification of customers) and the continuous dependent variable (probability of customer churning- regression) (Breiman, 2017). The C & R tree initiates by scrutinizing carefully the inputs in order to find the fittest split according to the reduction of the impurity index from the split. Pursuing this further, the split splits into two more splits (binary splits) and so on until one of the halting criteria has been triggered.

Generally, According to Breiman (2017), there are different algorithms for predicting the rate of customer churn using the continuous variables or categorical variables. For instance, the analysis incorporated the General Linear model and the General Regression model. This led to the combination of both the categorical variable and the continuous variable effects to predict the continuous dependent variable which in our case is the customer churning. The C & R tree was recommended because it determines a set of if-then logical conditions hence generating the splits that permit accurate predictions of the customer churn.

According to the surveys which have been conducted in the logistic department of prediction analysis, this methodology has been rated as one of the best since the decision tree implicitly sorts out the variables via screening or feature selection (Ekşi, 2011). This is one of the few techniques whereby a training data set such as the Tesco training data set is fit in the decision tree and few nodes are split from the tree with the dataset and feature selection done automatically. Also, they are time efficient because they require little efforts from the analytics for data preparations. In addition, the tree performance is not affected by the nonlinear relationships between different parameters since the linearity of the data is assumed. Besides, Ekşi (2011) notes that the results from the analysis are easy to interpret and explain due to the fact that they are clear to understand why classifications and predictions are done in a certain manner. However, not only does the C & R tree method incorporate some strengths but also some weaknesses too such as pronate to errors in classification of many classes and the smaller number of training examples and the computational training is a bit expensive.

Consequently, another model type used in the prediction of customer churn is the Recency, Frequency and Monetary (RFM) method which is one of the most recommended technique of marketing by business advisers because it determines how recently a particular customer purchases product, how often and the amount of money the customer spends when they start shopping (Girolami et al, 2006). With the help of customer relationship management, the RFM method is very vital because it helps in figuring out the target mailing to customers to ensure customer satisfaction and recovery thus aiding in retaining its customers. The RFM method was used so as to evaluate the performance of the constructed model through the lift chart so as to determine which customers were the topmost, average, and bottommost shoppers. Girolami et al (2006) adds that with the integration of the RFM method analysis, it has been proven that only 20% of the customers contribute to 80% of the growth of a business since these customers tend to shop more often purchasing goods and services worth a lot of money in the business enterprises. In summary, the recency of the customers exhibits their level of engagement, the frequency can predict the time the customer is expected to purchase goods, and the monetary value predicts the value of the business.

Besides that, in order to perform the RFM analysis, the customers are divided into four groups according to their recency, frequency, and monetary expenditure. Thereafter, the four groups are further divided into three different customer segments. The customers are then awarded some model evaluation metrics such as 1 for the presence of any of the three elements and an x for lack of any of the elements (Khajvand & Tarokh, 2011). For example, a customer who has purchased most recently, most quantity and spent the most will belong to the (1-1-1) RFM segment while that who have not purchased for some time, but purchased more frequently and spend the most will belong to the (3-1-1) RFM segment.

The RFM system helps in identifying and analyzing the best customers in the business. Significantly, strategies such as rewarding the best customers through loyalty programs are essential regardless of the business model (Khajvand & Tarokh, 2011). This is done in order to retain and motivate the customers to increase their loyalty towards the business enterprise. For instance, the TESCO company ranges the customer loyalty across two loyalty stages such as silver and gold thus assisting their analysis panel in the determination of the best customers. Other strengths of this methodology include increased customer retention, response rate, conversion rate, and increased revenue collection.

However, the RFM model comes with limitations too such as lack of an intelligent link between frequency, recency, and monetary value when it comes to customers having the same recency, frequency, and spending the same amount of money but in different transactions. Also, the segmentation applied to the system is inferior (McCarty & Hastak, 2007). In conclusion, the RFM model is simple, easy to understand, and allows higher flexibility in predicting different phenomena in any business enterprise.

Moreover, the analysis integrated the use of the IBM SPSS modeler and Microsoft Excel.

EMPIRICAL STUDY

The empirical study is a methodological quantitative and qualitative approach used to collect and analyze data based on observations and experiences. For instance, the case in this report will focus on the RFM model which will highlight the three aspects of the model such as the frequency, recency and how customers spend on the supermarket’s nine products. From the earlier discussion of evaluation metrics, the following table gives a detailed information on all the metrics.

Table 1:  Evaluation metrics

Table 2: Key RFM segments

Table 3:  A sample of the working data

The Model Building Process

Step 1: The database (TESCO.training) file is exported from CloudDeakin

Step 2: A .CSV file is prepared with all the orders

Step 3: The RFM-analysis.py script in the directory is executed

RESULTS AND DISCUSSION

RFM Model

From the above working data, the results of the RFM model were as shown below

Table 4: A sample of the results from the RFM model

The above data implies that;

3. Customer 159712 purchased more recently thus R=1, bought few times therefore F=4, spent somewhat little money hence M=3.
4. Customer 173678 purchased more recently thus R=1, bought few times hence F=4, spent very little money thus M=4.
5. Customer 187164 purchased more recently hence R=1, bought few times thus F=4, spent somewhat little money thus M=3.
6. Customer 145986 purchased more recently hence R=1, bought more frequently hence F=1, spent more money thus M=1.

C & R Tree

From the above binary C & R Tree, it is prudent to say that the TESCO Company is doing great in the chain store business. This is due to the fact that the supermarket has a lower customer church of around 4% thus a higher customer retention rate. Besides that, the social economic influence in the industry is very high (71,6%). This concludes that it offers relatively cheaper and efficient products, therefore, the trust and loyalty of its customers is gained day by day. With an average frequency of 7% per customer, it is clear that the probability of new and the existing customers shopping in the chain store is very high.

Screenshot 1: product contribution towards the total monetary value

From the above screenshot, sales from the Apparel category performed the worst while that of the grocery category performed the best being followed closely by those in the Dairy category.

In the previous business world, random guessing was the order of the day due to the lack of the current technology which is being incorporated in the business industry. Random guessing gives an exaggerated and unrealistic data which thereafter brings a lot of inconveniences and loss to the business (McCarty & Hastak, 2007). With the aid of C & R tree and the FRM method of prediction, the predicted analysis of business data is more accurate and extensive than the normal guessing. This is due to the fact that the results of the analysis are complex but easier to understand because they are given in percentages, bar and pie charts. In addition, other offer different recommendations in terms of comments according to the progress of the business. Since the results of the C & R tree method are more detailed than those from the RFM method, C & R method is the best because it gives its results in a binary tree which are easily understood. Generally, the new technology around the globe has brought a lot of pros than cons in the business sector (Ekşi, 2011). Undoubtedly, from the above statistics and analysis, it can be concluded that the TESCO company is doing much more in making their customers feel comfortable and also invested a lot in retaining its customers.

The following shows the results from the IBM SPSS modeler

Screenshot 1: The comment section

screenshot 2: The analysis outline showing customer churn, text analysis, and current customers analysis

Screenshot 3: Customer chart analysis

screenshot4:Customer Segmentation analysis

Screenshot 5: Customer cluster analysis

Screenshot 6: An excel document

RECOMMENDATIONS AND CONCLUSSION

Unquestionably, TESCO company is doing great in the chain store business in the United Kingdom due to its higher percentage of customer retention and keeping a very low customer churn percentage. As discussed earlier in the introduction of this report, customer satisfaction is very vital to the growth of a business. Pursuing this further, customers are satisfied when the management looks into details of the non-transactional elements in the business such as customer complaints and recovery. Discussed below are some recommendation to the Tesco company.

The success of the incorporation of diversification in the enterprise is felt when all nine categories of products can strive through during bad economy (Spiess et al, 2014). Significantly, it will be of great will to the company if the apparel, general merchandise, and the liquor category increase their sales either through promotions so as to meet the standards of the competitive market. Additionally, the company ought to make a wise decision in exploring the market of at least a majority of the continents so as to make larger profits and justify its position on the global face.

Moreover, in order to retain its customer-base, the company should take the advantage of bringing the marketing resource closer to their customers by clearly defining its vision, mission, values, and the roles of each employee towards retaining the customer base (Spiess et al, 2014). In the case of customer churn, the company’s committee should work on decreasing the percentage to almost null (Bora et al, 2015). This will be accomplished by understanding and measuring why the customers are leaving in the first place. After understanding the causes of the churn, the committee should come up with solutions such as escalating the treatment of the customers by relating to them at a personal level, appreciating the customers with thank you gifts, thank you notes so as to motivate them to visit the store again and again (Gustafsson et al, 2005). Besides that, the employees can welcome the customers at the entry and on their way out ask them for any constructive feedback to look for any complaints. Creation of customer participation forums such as the board of customer can be of great help (Hill & Alexander, 2017).

In conclusion, the TESCO company is just an example of all the chain store worldwide which is doing a tremendous job in increasing its profits through the incorporation of ideas that benefit customers. The discussion above has clearly shown the effectiveness of handling the non-transactional elements in any business which was the core aim of the project.Companies should emulate the work being done by the management team of the TESCO company in customer churn reduction, improving customer satisfaction, taking customer complaints seriously, and increasing their rate of customer retention. I would recommend that the TESCO company ought to appreciate its management team for the good work they have exhibited.

References

Bora, D., Pathak, B., & Bezbaruah, S. (2015). Customer churn analysis. Tezpur University.

Breiman, L. (2017). Classification and regression trees. Routledge.

Chen, I. J., & Popovich, K. (2003). Understanding customer relationship management (CRM) People, process and technology. Business process management journal, 9(5), 672-688.

Ekşi, İ. H. (2011). Classification of firm failure with classification and regression trees. International Research Journal of Finance and Economics, 76, 113-120.

Girilami, M., Mischak, H., & Krebs, R. (2006). Analysis of complex, multidimensional datasets. Drug Discovery Today: Technologies, 3(1), 13-19.

Gustafsson, A., Johnson, M. D., & Roos, I. (2005). The effects of customer satisfaction, relationship commitment dimensions, and triggers on customer retention. Journal of marketing, 69(4), 210-218.

Hill, N., & Alexander, J (2017). The handbook of customer satisfaction and loyalty measurement.  Routledge, Canada

Knox, G., & van Oest, R. (2014). Customer Complaints and Recovery Effectiveness: A Customer Base Approach. Journal Of Marketing, 78(5), 42-57. doi: 10.1509/jm.12.0317

Kisioglu, P., & Topcu, Y. I. (2011). Applying Bayesian Belief Network approach to customer churn analysis: A case study on the telecom industry of Turkey. Expert Systems with Applications, 38(6), 7151-7157..

Kumar, V., & Reinartz, W. (2016). Creating Enduring Customer Value. Journal Of Marketing, 80(6), 36-68. doi: 10.1509/jm.15.0414

Khajvand, M., & Tarokh, M. J. (2011). Estimating customer future value of different customer segments based on adapted RFM model in retail banking context. Procedia Computer Science, 3, 1327-1332.

McCarty, J. A., & Hastak, M. (2007). Segmentation approaches in data-mining: A comparison of RFM, CHAID, and logistic regression. Journal of business research, 60(6), 656-662.

Rust, R. T., & Zahorik, A. J. (1993). Customer satisfaction, customer retention, and market share. Journal of retailing, 69(2), 193-215.

Spiess, J., T’Joens, Y., Dragnea, R., Spencer, P., & Philippart, L. (2014). Using big data to improve customer experience and business performance. Bell labs technical journal, 18(4), 3-17.

Storbacka, K., Strandvik, T., & Grönroos, C. (1994). Managing customer relationships for profit: the dynamics of relationship quality. International journal of service industry management, 5(5), 21-38.