Churn analysis essentially deals with analysing the group of customers who have broken through from one company and joined the competitor or would be doing so in near future. It is imperative for companies to maintain the loyalty of their customer base which is also the basis of branding of products and services. But in some cases, customers are found not to be satisfied with the level of service that the company provides and are leaving or would be leaving it for joining the customer base of any other company’s customer base. (Borna, C., 2000)
Retention of customers is important to most of the organisations as the cost of acquiring new customers is significantly greater than the cost of keeping relationships with current customers. Understanding the customers is one of the most effective aspects of customer retention thus avoiding churning. It is important for organisations to understand the expectations, satisfaction levels, demographic trends, future trends in customer behaviour to be able to retain the existing flock of customers (Lu, J., 2002). The most effective customer churning avoidance would be to identify customers who are likely to be potential defections to rival organisations. Once this group of customers are being identified then the organisations can take adequate steps to prevent any customer attrition. Often organisations use profiling of the churning customers which can reveal the reasons of defection. The predictive models predict the likelihood of defection and perform segmentation based on defection scoring given to particular group of customers. Organisations like banking institutions take into account the life time value of the customers and decide on giving them perks which would avoid any churning when any such situation occurs. In addition, organisations also have to take into account the cost of acquisition of customers which they would have to invest in future if a customer is lost. Customer churning analysis is mostly relevant in industries where there is supposed to be a long term relationship with customers and the brand image plays an important role in buying behaviour of customers.
Propensity to Buy
Buying something is a very personal experience which involves a lot of thinking about the need of the item, the pricing range, the emotional attachment to the brand, etc. The propensity of buying a certain item is about finding the likelihood of the customer buying a particular product or service. It depends on various factors but sometimes it can be predicted and suggestions can be made to the customer which the company knows would be attractive to the customer (Keown, C.F., 1989). It is especially more feasible and likely in the e-commerce domain as companies have the previous track of customer purchases.
Based on the previous pattern of purchases, customers can be given specific information which would certainly be attractive to them and would eventually lead to an item being sold to the customer. Based on the previous experiences it is found that customers can be divided into certain groups which define their propensity to buy a particular product. The first group of customers know exactly what they want and would directly ask for specific information about the product which they want. These customers are the most likely buyers and companies should give maximum attention to these customers.
The second category of customers who sort of know what they want, but have not yet finalised that requirements, they are still in the process of narrowing down their choices and need further information about the products to make up their mind. They might turn out to be potential buyers given the company is able to provide them with accurate information and make them feel that this is the product that they are looking for.
The third set of customers are window shoppers who don’t enter the store or web portal with any specific intention to buying anything, but are looking for anything which they would find attractive and would buy it on their instincts. Most of the customers in super markets fall into this category of customers.
The fourth category of customers are the once who are just looking for products and have no intension of buying in the current date. They just want to check out the new products and would certainly not buy it today but would probably buy in sometime in future.
All these category of customers have different propensity to buy certain products, hence the companies need to concentrate on the once which are the most likely buyers from the crowd of customers.
Clustering in Business to Customer Websites
The marketing is about the art of spreading the message across the board, as well as sending the desired message to the customers which would eventually drag the customer towards the products. The sale of goods in an e-commerce setup essentially involves interaction between the customers and the web portal. Each customer’s interaction with the site can be represented as a session between the customer and the web portal and it can be stored for future profiling of the customers preferences. Hence a customer group in e-commerce can be considered as a session group.
The data that the web portal can gather through the various browsing session can be stored and data mined for analysing the customer behaviour to identify in which group a particular customer can be put into. Using this data, the web portal can merge web pages which are represent a common characteristic which appeals to particular group of customers, for example, the web pages which have a common set of products like DVD players and portable music players (iPod’s) can be liked together to give an integrated look to the customers who would be interested in browsing through those products as well.
The clustering of customers can also be done through their buying patterns. Is the customer is a regular buyer, or an occasional buyer? Is the customer a frequent browser, but not a frequent buyer? Is the customer just a browser who does not buy products but is used to checking the product prices and their features. All these buying patterns divulge lot of information about the customers psyche.
In addition, customers can be classified according to their pricing range (low, medium or high end), frequency of buying, number of products purchased in one transaction, etc. All this categories can be used to classify the customers and help companies to be able to manage a targeted approach towards the potential customers.
In today’s competitive market, the struggle for customers retaining and increasing the customer count has become a prime motive of organisations which is central to the achievement of its goals. The fierce competition between rival organisations has given birth to heighten customer expectation. In addition to it, the delusion of borders for business has added to the pressure on organisations to come up with innovative techniques and schemes to attract and maintain the current customer pool. All the above factors have forced companies to adapt a customer centric approach to the working of the organisation. (Adebanjo, D., 2003; Chen, Q. & Chen, H.M., 2004)
The e-CRM helps in identifying the most valuable customers for an organisation which can be best suited for the lifetime value that the organisation can provide to its customers. It allows companies to do thorough data mining and finding the patterns of customer behaviour. The results from the data mining exercise could give crucial information for the management to take informed decision in real time. The real time decision process could reduce the response time for the customer’s queries or complaints thus helping overall customer satisfaction which leads to customer loyalty towards a particular brand or organisation. The e-CRM could also enable markers, executives or sales people to gain an insight into the business operations. e-CRM could provide some effective tools for organisations to maintain and enhance the customer base. Author would like to propose the e-business CRM strategy using the following steps:
Customer transactions generate a huge database during their interaction with digital offerings. Various data mining techniques could be employed for segmenting, profiling of the customers. These profiles could be used for customised marketing according to the customer profile. This could enhance the customer retention as well as could attract new customers.
The data provided to the customers could be personalised according to his/her choice. The data could be passive, active, inferred or a real-time customisation based of customer preference. It is a kind of content management where the customer gets the data according to its previous transactions or set preferences. In addition, merchandising and cross selling could also be considered as part of the personalisation.
One thing that has occurred is that the term personalisation has been applied to techniques that, although beneficial to marketers and their audience, may not truly be personalisation. For example, the technique of allowing a user to customize such things as the background colour and or the position of objects on the page is called personalization by many sites; however, these features are more akin to customization (Pan, S.L. & Lee, J.N., 2003). Personalisation essentially provides control over things which the customer wants the information about. Such as stock price which should be displayed, where to place news head lines, and other formatting features.
e-CRM concept helps organisations to understand its customers or in general all the customers within that particular market domain. It helps the organisations which are implementing e-CRM based on the web sites on the Internet to address the key audiences and providing them with relevant information which is a key aspect of the working of organisations. The electronic CRM helps in identifying the most valuable customers for an organisation which can be best suited for the lifetime value that the organisation can provide to its customers. It also allows companies to do thorough data mining and finding the patterns of customer behaviour. The results from the data mining exercise could give crucial information for the management to take informed decision in real time. The real time decision process could reduce the response time for the customer’s queries or complaints thus helping overall customer satisfaction which leads to customer loyalty towards a particular brand or organisation. The e-CRM could also enable markers, executives or sales people to gain a insight into the business operations.
Customer Contact Management
Various channels of communication could be set such as e-mail for in-bound and out-bound communication. e-CRM could help in handling interaction with customer and build online agents and help in live online support, also web chat. It could be helpful in effectively maintaining communication between customer and organisation. When 24/7 help is available through e-CRM, it helps in building a faith based relationship where customer always thinks that there is someone to help him out at every time throughout (Jutla, D., Craig, J. & Bodorik, P., 2001). Companies could take advantage of the automated e-mail system to make the customers aware of the recent deals available which could be of interest to them based on their previous purchases. It could also help organisation to concentrate its human resources in other productive work. The e-CRM implementation also helps in segmenting the customers based on their particular interests. Thus organisations could channel the right information towards the relevant customers.
The e-CRM implementation reduces the cost of communication between organisation and customers. As the customer’s information is readily available, any enquiry from the customer could be responded within few seconds. This helps organisations in free themselves for other productive work, thus reducing administrative and operational costs. As Internet based CRM’s result in service oriented transactions, customers are empowered to configure, track and change their orders themselves (Liver, B. & Dermler, G., 2000). The e-CRM implementation integrates the workflow between the e-CRM application and the back office systems such as production, finance, and supply chains; hence it results in delivering cost savings. It improves the overall customer interaction which leads to providing better service, and improve customer satisfaction and customer loyalty to gain a lifetime customer for the organisation
Customer Community Building
e-CRM could help in creating forums where customers can share there views and experiences. Such open forums could help in building faith in the brand and organisation. As the forums are open for all the customers, it could help in build transparency within the system. This service helps in building inter-user interaction where the communication is based on discussion in forums, user-generated content, messaging, ratings, auctions, etc. In addition to equipping the organisations with crucial customer behaviour insight, the e-CRM could also provide the organisations the capability to give a personalised and more customised service to its customers (Scullin, S. et al., 2002). The personalised approach towards customers leads to better customer service and greater customer loyalty. The organisations could implement intelligent agents through their e-CRM implementation to give the personalised touch to the service and gain customer insight. The facility of using agents is limited to e-CRM as compared to general CRM. The intelligent service agents can search for the best deals for customers by searching through the database. An example of this could be the insurance deals provided by web based search engine providers which take the customer input and search through the various online insurance deals and find the best deal for the customer based on his/her choices. This concept saves a lot of time and effort of the customer and increases the customer satisfaction. It also increases the cross-selling opportunities for organisations which could help the customers finding the related products which could be purchased with the current purchase. It is beneficial for both customers as well as organisations.
Supply Chain Management
Now the focus should be shifted towards the supply chain management. The most efficient supply chain in the world would mean little if the customer cannot track the order he/she placed with the web portal. The supply chain management technologies are decreasing production costs is practically a given in today’s business. Hence the author would recommend that the e-business strategy should place a significant attention to maintaining an efficient supply chain management system. The resulting improvements in service at the sales, marketing, and contact centre levels are to be monitored continuously (Lee-Kelley, L., Gilbert, D. & Mannicom, R., 2003). The company’s partners should be able to continuously monitor the level of sales and should be able to translate the resulting information into better customer service (Xue, M. & Harker, P.T., 2002). Companies, particularly those selling commodity products, are increasingly doing away with their approved supplier lists and posting proposal requests directly on the Web, inviting all potential suppliers to bid and thereby increasing competition and lowering prices
Integration of Channels
The author would also like to mention the importance of maintaining as many sales points as possible. The organisation should not limit its sales points to its own particular web domain; it can attract customers through various web domains and even go for the retail infrastructure which would complement the web portals. For example, according to the local preference of having a feel of the products in Japan, Dell Inc. has opened up retail outlets in exhibitions which would provide the facility for customers to have a feel of the computers before they buy them (Luck, D. & Lancaster, G., 2003). The e-CRM implementation is marred by three perspectives, viz, complexity, time frame, and configurability. The complexity of the implementation cannot be under estimated as it could lead to chaos within the organisation. The cost of purchase of e-CRM applications is also an aspect of the complexities involved in the implementation of e-CRM’s. As the business processed involved in organisations vary, each implementation has to be customised according to that particular organisations needs. For this purpose organisations often take the help of external consultants which have professional expertise in customising e-CRM’s which could potentially increase the cost involved in the implementation of e-CRM, in addition to the actual cost of the CRM itself.
Profiling of Partners
The author would like to suggest that the company should profile its partners as well, in addition to its customers. The analysis would reveal the sales pattern through a particular channel. With the electronic business set up it becomes more feasible to be able to analyse the data in real time and take steps accordingly.
All the above discussion helps in maintaining better customer service, effective customer lifecycle management, getting customer feedback, customer loyalty. e-CRM could be helpful in all the above categories, but it’s the ultimate product which is going to have the final verdict by the customer.
Difference between Online Analytical Processing, Data-mining, and Query reporting
Essentially, all the above mentioned techniques are data analysis techniques but they differ in the aspects of how they are applied. For example, data mining is driven by data where the queries are being applied on the data directly to get the hidden patterns in the data which are not obvious on the overlook. On the other hand, online analytical processing is driven by the user or the user’s intention to verify his or her queries. The major distinction between online analytical processing and data mining is based on how they operate on the data. The OLAP technologies focus on the reporting of the facts to the user, while on the other hand data mining is all about finding hidden patterns within the database or data warehouse. Another distinction can be made on the basis of the approach to data analysis. The OLAP technologies stress on the top-down approach to data analysis since it basically comes down to the operation of data via addition. In other words, OLAP applications typically operate on summary data that have been aggregated in complex tables for fast and easy analysis. With an OLAP application, a marketing manager might slice through a segment of recent high-value customers to report on what they are buying, how much they are spending and how much time they typically spend at the company site. (Goil, S. & Choudhary, A., 1997; Hand, D.J., 1998; Han, J. & Kamber, M., 2006)
Data mining incorporates analytical procedures grounded in traditional statistics, mathematics and business and economic theory. Much of mining methodology takes its roots from what is referred to as econometrics. Econometrics is an analytical methodology that involves the application of mathematics and quantitative methods to historical data in conjunction with traditional statistics with the focus of testing an established economic or business theory. (Mannila, H., 1996)
Data mining algorithm runs on sales and visit data to discover patterns of behaviour in the high-value customers. Data mining can find what factors influence certain types of sales most. Therefore, it can be said that data mining operates on division of data rather than summation. While end users follow their intuition with an OLAP application to investigate data, IT or marketing departments perform data mining to analyse the data, then report back to business managers about what they found.
Data mining is generally done on the offline data by creating data warehouses, while OLAP is mostly done online by the user. Another difference lies in the amount of data being used. In data mining applications, the amount of data that is generally being analysed is in level from few giga-bytes to hundreds of giga-bytes, which is not the case in terms of OLAP applications where the database is quite smaller. The complexity of data is also an issue of difference between the two methodologies. While data mining is able to handle very complex data bases, OLAP is generally limited to lower-to-medium complexity levels.
The data mining technologies give a complete picture of the customer behaviour while the OLAP can only provide a slight insight into the customer behaviour, hence data mining techniques are far more effective in analysing the customer patterns than OLAP technologies. Basically they are designed for different purposes hence they do not match up in scale or complexity that can be handled.
In some cases, the data mining techniques may involve heuristics when the amount of data to be analysed is huge typically in the case of scientific computing where the data is in the range of tera-bytes. It is not the case in terms of OLAP or query reporting which gives an actual account of the data which is analysed. The higher end methodologies are used to develop fine tuned business strategies through the use of sensitivity analysis and predictive modelling which differs from those techniques which merely attempt to uncover patterns and relationships in more generic and often voluminous data. One could consider directed data mining methodologies as part of the information mining spectrum.
A major difference between query reporting and data mining is that it is interactive querying of the database which delivers powerful query, analysis and reporting capability of the database. It can be used to analyse database through an intuitive and highly interactive interface that also lets users design their own dashboards which is not the case in the data mining approach. The data mining is a complete approach in the sense that it definitely allows user inputs but only after the whole analysis is being performed. Another aspect of difference is in the level of mining of the data. Generally, data mining is performed on the whole of the database to find the hidden patterns which are not obvious from the simple querying of the database while query reporting is generally performed on certain section of the data base looking