Embed Size (px)
Citation preview
Universidad Politécnica de Valencia
Facultad de Informática
Improving company effectiveness using e-CRM
Bachelor project
Author: Juozas Stundys
Executive: Dr. Ignacio Gil Pechuán
Valencia, 2008
2
Index
1. Introduction ............................................................................................................................... 3
2. Objectives................................................................................................................................... 5
3. Main structure ........................................................................................................................... 6
3.1. CRM description ...................................................................................................................... 6
3.1.1. What is CRM?....................................................................................................................... 6
3.1.2. Defining CRM ...................................................................................................................... 7
3.1.3. Defining the boundaries of CRM .......................................................................................... 8
3.1.4. Back office - front office ....................................................................................................... 9
3.1.5. Ten ways to initiate or improve your CRM ........................................................................ 12
3.1.6. Summary ............................................................................................................................. 14
3.2. Data mining in CRM .............................................................................................................. 14
3.2.1. What is data mining? .......................................................................................................... 14
3.2.2. Defining the goal................................................................................................................. 15
3.2.3. Applying Data Mining to CRM .......................................................................................... 16
3.2.4. Summary ............................................................................................................................. 19
3.3. Internet and CRM .................................................................................................................. 19
3.3.1. The Internet Enables CRM ................................................................................................. 19
3.3.2. Internet customer interaction .............................................................................................. 20
3.3.3. Metrics of CRM effectiveness ............................................................................................ 21
3.3.4. Web site as a measurement tool .......................................................................................... 24
3.3.5. Summary ............................................................................................................................. 26
3.4 e-CRM..................................................................................................................................... 27
3.4.1. The Emergence of e-CRM .................................................................................................. 27
3.4.2. Key Applications of e-CRM ............................................................................................... 28
3.4.3. Management Steps for e-CRM Integration ......................................................................... 30
3.4.4. Summary ............................................................................................................................. 32
3.5. Data mining and e-CRM ........................................................................................................ 32
3.5.1. Data mining in e-commerce ................................................................................................ 32
3.5.2. Integrated architecture ........................................................................................................ 32
3.5.3. Data collection .................................................................................................................... 35
3.5.3.1 Business event logging ..................................................................................................... 35
3.5.3.2. Measuring personalization success .................................................................................. 35
3.5.4. Analysis............................................................................................................................... 36
3.5.5. Summary ............................................................................................................................. 39
4. Conceptual decentralized supermarket model ..................................................................... 40
5. Conclusions .............................................................................................................................. 43
6. Bibliography ............................................................................................................................ 46
3
1. Introduction
The better a business can manage the relationships it has with its customers the more successful
it will become. Therefore IT systems that specifically address the problems of dealing with
customers on a day-to-day basis are growing in popularity.
Customer relationship management (CRM) is not just the application of technology, but is a
strategy to learn more about customers' needs and behaviours in order to develop stronger
relationships with them. As such it is more of a business philosophy than a technical solution to
assist in dealing with customers effectively and efficiently. Nevertheless, successful CRM relies
on the use of technology.
Any customer can be satisfied. The challenge is whether this can be done in a way that is
economically feasible for the enterprise. The challenge of balancing these (sometimes
conflicting) imperatives is the essence of customer relationship management.
Figure 1.1: Value proposition
Customer
profitability
Customer
satisfaction
The term customer loyalty is used to describe the behavior of repeat customers, as well as those
that offer good ratings, reviews, or testimonials. Some customers do a particular company a great
service by offering favorable word of mouth publicity regarding a product, telling friends and
family, thus adding them to the number of loyal customers. However, customer loyalty includes
much more. It is a process, a program, or a group of programs geared toward keeping a client
happy so he or she will provide more business.
Creating and nurturing loyal customers has always been a top priority for marketing teams within
companies large and small. But today's most innovative firms are looking for ways to go beyond
the frequent buyer program and transform the loyal customer into an essential extension of the
company's sales, marketing, and product development teams.
4
To realize the entire CRM concept needs a lot of secondary sources of theory. Therefore these
papers contain more information including data-mining, business strategy and so on.
The main structure of this document is organized in five parts. First two is preface to understand
objectives and structure of the paper. All primary information is placed in third part. Chapters
3.1-3.2 contain major theory about CRM, data-mining. Some general algorithms give a lot of
weight for further research. Chapters 3.3-3.5 discusses another part of CRM that is used for
electronic B2C (business to client) commerce (e-CRM). Moreover it is useful source of theory
for my conceptual decentralized supermarket model (Chapter 4). Each part ends with short
summary containing key facts of a chapter. Last part finishes paper listing primary conclusions
of the document.
5
2. Objectives
The main goal of this paper is to describe the importance of customer value for company
analyzing customer segmentation. Customer segmentation is a model dividing customers into
groups with similar characteristics. Implementation of this method requires both data-mining and
CRM knowledge. As an example I use simple customer profiling model (Figure 2.1) with a few
variables due to that data-mining contains enough complex mathematical and statistical
techniques.
The next step is to find some common theoretical suggestions for this type of model.
Figure 2.1
Rev
enue
Risk
Customers
segment 1
Customers
segment 2
Customers
segment 3
Customers
segment 4
The secondary objective is to show an example adjust new customer interaction methods in an
ordinary business market. The conceptual model visualizes mobile phone technologies
adaptation in supermarket sphere using e-CRM and data-mining fundamentals.
6
3. Main structure
3.1. CRM description
3.1.1. What is CRM?
CRM is short for Customer Relationship Management. Essentially1, CRM aims to put your
customers at the centre of the information flow of your company (Figure 3.1). The company or
organization is very rich in information about customers. It knows lots about them. But the
information is not shared. It‟s only available to specific job functions.
If a sales person wants to know about what issues are outstanding with customer service for a
particular customer, then they have to make contact with the holders of that information and wait
for a response. If the salesperson is chasing the information in response to a question from the
customer, then the customer also has to wait.
So, although many companies are information rich, the information is compartmentalized. It is
not corporate knowledge and the ability to access information and to deliver it rapidly to
customers is low - high quality customer service is compromised.
In a customer focused company, the information flow and the ability to access information is
very different.
1 CRM as a theory has a lot of characterizations, but in this paper I have used a description from
http://www.SalesAgility.com paper “Introduction to CRM”
7
Figure 3.1
Source: http://www.salesagility.com “Introduction to CRM”
CRM is an application that enables companies to make the move towards being a customer
centered organization by putting the customer at the centre of all the information that relates to
them and allowing authorized people within the organization to access the information.
In a customer centered organization, salespeople would have access to all the information that
affects their relationship with their customer. The conversations, the emails, the complaints, the
complaint resolutions, all the information that had been sent to the customer, who else in the
company the customer had spoken to … everything that affects their ability to service the
customer and sell more product or services to them.
Customers of a customer centered organization feel more valued. Their requests are dealt with
more rapidly and accurately because all the information required to service the request is in one
place. Customer centered organizations may have a higher customer retention rates than
competitors organized along traditional lines because of this.
3.1.2. Defining CRM
Customer relationship management in its broadest sense simply means managing all customer
interactions. In practice, this requires using information about your customers and prospects to
more effectively interact with your customers in all stages of your relationship with them. We
refer to these stages as the customer life cycle.
8
The customer life cycle has three stages:2
Acquiring customers
Increasing the value of the customer
Retaining good customers
Sometimes a distinction is made between analytical versus operative CRM (Figure 3.2).
Analytical CRM has to do with modeling, campaign management, and long-term decisions on
customer development strategies. Operative CRM deals with execution of customer contact
strategy.
Figure 3.2
Source: web page of Future of identity in the information society - http://www.fidis.net
It might be interesting to predict customer behavior, but only when one manages to act upon
these insights does this knowledge become commercially relevant. Prediction and analysis is
important, but only deployment will make this useful. Exactly this is where the most difficult
challenge seems to lie at the moment. When deployment results are consistently fed back, the
organization will learn from its past actions and truly adapt to customer needs as displayed by
response. We make the assumption here that one may infer relevance of the offer to the customer
from response.
In large organizations, specialties of analysis versus execution are likely to become dispersed -
even more so as the level of sophistication in modeling advances. The interface between
operative and analytical CRM relies heavily on good metadata for using and reusing data mining
model score code. Also, the ability to evaluate and monitor models in a time efficient and error-
free manner will lead to improved targeting. As an invaluable side effect of this, more
knowledge on customer behavior will be gained. This customer knowledge is considered by
many to be the most important asset in competitive markets these days.
3.1.3. Defining the boundaries of CRM
2 Herb Edelstein “Building profitable customer relationships with data mining”, Two Crows Corporation
9
CRM suffers from a lack of clear definition. There is no consensus about what is meant by the
term CRM. Different constituencies having an interest in CRM have different emphases.
Technology firms want to sell software solutions into client organizations; consultants want to
profit from helping clients generate strong business outcomes from their CRM investments;
clients typically want CRM to improve both cost and revenue sides of the profit equation.
In the context of these different perspectives, we have found it useful to conceptualize CRM at
three levels of abstraction: strategic, operational and analytical.
At a strategic level, CRM is seen as a core business strategy. As such it competes against other
possible core business strategies. CRM is consistent with becoming more customer - or market -
centric3. Other strategies might be preferred in different conditions of economic and market
development. For example, as emerging national economies have sought to generate wealth, they
have typically adopted a production - oriented approach to doing business where the goal is to
compete on price supported by low unit manufacturing costs. For example, this is true of many
businesses on mainland China today. Similarly, in developed economies, as new markets
emerge, companies typically adopt a sales oriented.
At an operative level, CRM is concerned with automating chunks of the enterprise. CRM
vendors have developed products that enable automation of selling, marketing, and service
functions. A major driver of CRM implementations has been channel integration. Whereas it was
common for companies to have single or few routes to market, now it is commonplace to have
many. In business-to-business markets, channels have multiplied: distributors, catalogues, on-
line, electronic exchanges/auctions, direct selling. Under such circumstances, the creation of a
single-view of the customer using data captured across all channels and exploitation of that data
has been a huge challenge. It is a challenge that is difficult to meet without IT. Most CRM
projects also involve a number of smaller, but also very challenging projects, such as: systems
integration, data quality improvement, process reengineering, data analytics, and market
segmentation. CRM implementations such as these require strong project management and
change management skills.
At an analytical level, CRM is focused upon exploitation of customer data to drive more highly
focused sales and marketing campaigns. Analytical tools such as decision trees, neural networks
and clustering can be used to improve the effectiveness and efficiency of customer acquisition,
customer development and customer retentions strategies. We conclude this section by reporting
our preferred definition which reflects all 3 perspectives: CRM is the core business strategy that
integrates internal processes and functions and external business networks to create and deliver
value to targeted customers at a profit. It is grounded on high quality customer data and enabled
by information technology.
3.1.4. Back office - front office
3 Kotler, P. (2000). “Marketing Management: Analysis, Planning and Control”
10
To deal with the challenges of customer relationships in the fast-evolving Internet world, even
the most customer-focused companies have to understand the three essential insights to getting
customer relationships right4:
1. That building CRM in the front office is just the start, and that it must involve the back
office functions like manufacturing, fulfillment, and billing as well as the analytical
functions like data warehousing and pushing customer insights back up to the front
office.
2. That conducting relationships across multiple media requires the correct technical
infrastructure, allowing companies to deal with their customers in a consistent way across
multiple media, and even add new media as required without the need to develop every
interface separately and from scratch.
3. Building the correct strategy for directing customers to different media. For a few
organizations the strategy “we will deal with customers on whatever medium they prefer”
is right; but for the vast majority of organizations it is a recipe for disaster.
Getting it right in CRM across multiple channels means that you can deal with customers in and
across multiple media and still have a unified up-to-date view of the customer, with no gaps.
Ideals such as one-to-one marketing and the market one have been widely written about but
rarely realized, except in the occasional corner florist‟s. Getting CRM right is the closest
approach to achieving these ideals that a large organization can make. Doing this across multiple
media is a major achievement that will make the organization ready to face the future.
In CRM, there is a “virtuous triangle (Figure 3.3). The purpose of this is to ensure that
organizations know their customer fully, and then act according to their needs and the
organization‟s interest. Important information is generated and used in other areas. Any company
that is doing CRM properly must integrate the front office, the back office, and analytical
systems.
The back office executes the customer requirements. Generally the only customer contact
functions in the back office are billing and logistics, and in even these functions, the
customer contact is moving into the front office environment.
Analytical software allows the organization to look for patterns in the customer data
which they have collected. The outputs from this are strategic and tactical information.
The strategic information can be used to determine future strategy, while the tactical
information will help to modify existing practice. Increasingly the tactical information is
generated and used on the fly in customer interactions.
4 Bradshaw, D. & Brash, C. (2001), Managing customer relationships in the e-business world:
how to personalize computer relationships for increased profitability. International Journal of Retail & Distribution Management, Vol. 29, No 12, pp. 520-530
11
Figure 3.3: The “virtuous triangle” of CRM
Source: Bradshaw & Brash (2001), pp. 525
The current focus on CRM tends to be almost entirely on the front office. This is not harmful -
almost all organizations could improve their performance in this domain - but it is not optimal in
the long run. Extending CRM into multiple media means integrating the front office and aspects
of the back office with different communication channels (Figure 3.4). This has to be done in a
methodical way and organizations that get this part right will have thought carefully about the
technical infrastructure they need. Standards are rapidly developing in this area and many
vendors are building media portals that allow the organizations using them to deal consistently
with customers across multiple media.
Just because organizations can deal with customers across multiple media does not mean that
they should offer the same facilities on each medium, or the same level of service. Doing so
would be a disaster. The media are different and demand being handled differently. For example,
people are generally willing to wait on hold for a reasonable time to speak to a call center agent.
However, they are certainly not prepared to wait to do the same transaction with an interactive
voice response (IVR) machine.
12
Figure 3.4: CRM and multiple media
Source: Bradshaw & Brash (2001), pp. 526
Some interactions are far more costly than others, and there is an especially steep differential
between interactions involving humans and those that are automated. Business models built on
automated transactions cannot sustain large volumes of transactions switching to human - based
interactions. Organizations must therefore decide two main issues:
1. For which customers and on what occasions they want to use specific media
2. How they are going to direct customers to the companies chosen medium.
The latter point tends to be the most problematic - directing customers to the chosen medium.
The most problematic part is directing customers away from the live agents to the automated
media. A way of doing this is price - live interactions cost more or discounts are available only
via automated media. But this is not the only way. As already pointed out, one way to do this is
with level of service - customers have to wait to speak to an agent, but can connect immediately
with the IVR (interactive voice response). Other choices are the levels of facilities, offered over
the different media. For example, the Web is very good for conveying large amounts of data and
graphics. If a company offers its customers all the data or graphical information they require
from the Web, customers who can will use the Web as their chosen medium.
3.1.5. Ten ways to initiate or improve your CRM
CRM is a business strategy5, which requires planning, commitment and change, and any
employee that has any point of contact with a customer at any time should be considered a
“CRM user”. Excellence in CRM is not achieved with a software product or a marketing
campaign. However, technology such as telephone systems and Web sites can be used wherever
it aids a CRM strategy. Below, ten ways to improve or initiate a CRM program is stated:
5 Vinas, T. (2003), Industry Week/IW, May 2003, Vol 252 Issue 5
13
1. Defining CRM
The heart of CRM is about knowing your customers and the way they want to interact. CRM is
about marketing and customer knowledge, not about great software. CRM can have a wide
scope, but it can also be as simple as managing your activities and keeping your promises.
2. Top management commitment
CRM defines what is happening in an organization so CEOs (chief executive officer) can move
in several areas with more business intelligence. CEO and board commitment is also a critical
factor that influences the impact of CRM initiatives and diminishes resistant to change. To state
it simply, if a CRM strategy is simple and sanctioned by the CEO, it will probably work.
3. Staff involvement
The biggest problem with CRM is that it requires staff to capture more data to do more things.
Some employees dislike the way that it slows them down and if they have not been shown the
value, they reject it. Organizations must get CRM users excited and start with a couple of key
supporters. The employees need to be assured that if they put information into CRM systems
they will get value out of it.
4. Integration of CRM systems
Different people in an organization have different views of the same customer. The marketer
might think, “He keeps buying, we need to keep selling to him” while the financial manager
thinks, “he is not paying his bills, we will have to stop selling to him”. This is why you need to
identify customers and why CRM systems need to be integrated. Furthermore, if customer data
does not easily reveal which customers bring in value, it can be too late to adapt marketing
practices once the information is realized. Successful CRM is about having real time access to
the right information.
5. Researching CRM tools and technologies
It is important to spend time researching the IT market to find the tools and technologies
appropriate for the organization. Customer relationship tools in the mid-90s did not include SMS
messaging, the Internet, or digital phones. There was only mail and fax. However, now we have
got interactive TV, interactive Web, digital telephony systems, and e-mail. It is a learning curve.
6. Long-term view
Some organizations benefit from expensive and complex high-end CRM systems. Those systems
can take a long time to roll out and require consultant input. In this case, management must be
patient and look at the big picture in the long-term instead of just focus on short-term costs. It is
all about what the organization needs, if you do not have a vision behind the sticker price, you
are in trouble.
7. Managing consultant and vendor relationships
Although consultant and vendor relationships can be fractious and expensive, these partners are
often the only one that can see the “bigger picture”. It is therefore important for organizations to
manage the relationships and develop mutually reasonable expectations.
8. Measuring the success of your CRM strategy
14
It may take time for a CRM strategy to show return. Often organizations will spend a lot of
money in CRM, but not allocate a budget to change culture or establish training. So they regroup
and sometimes there is a lot of trial and error. CRM is not a five-minute wonder. Sometimes it
takes decades for return on investment to come in.
9. Keeping it simple
While IT people need to help with technical CRM decisions, each investment should have a
business requirement. Remove what the organization do not need and implement CRM systems
where they add value for users before trying to add value to the organization. A big system might
take two years to roll out and then suddenly the organization‟s requirements and directions have
changed. Also remember that CRM needs can differ between departments.
10. Outsourcing
If the organization does not feel up to the CRM learning curve, consider outsourcing.
Outsourcers will often recruit, train, and manage contact centers on behalf of clients. The
customer does not care about where the person they are talking to is sitting as long as their needs
are met. Today, outsourcers that get customer information which they sell back to the
organization or uses to manage CRM for the organization are becoming more and more
common.
3.1.6. Summary
We have discussed about CRM fundamental steps and principles but it should be noted that for
our further research the most important part is front-office systems supporting company-
customer relationship.
Despite CRM involves wide range of techniques analytical part of CRM is more important for
our research. Analytical CRM is typically more difficult to implement, makes accurately
predicting ROI more difficult and may require a longer period of time to realize returns.
However, unlike operational CRM, the potential ROI of analytical CRM continues to grow over
time. The analytical process delivers the ability to continually improve customer knowledge and
relationships, which leads to continuing growth and profitability.
3.2. Data mining in CRM
3.2.1. What is data mining?
Data mining is a term that covers a broad range of techniques being used in a variety of
industries. Due to increased competition for profits and market share in the marketing arena, data
mining has become an essential practice for maintaining a competitive edge in every phase of the
customer lifecycle.
Historically, one form of data mining was also known as “data dredging”. This was considered
beneath the standards of a good researcher. It implied that a researcher might actually search
through data without any specific predetermined hypothesis. Recently, however, this practice has
15
become much more acceptable, mainly because this form of data mining has led to the discovery
of valuable nuggets of information.
Another form of data mining began gaining popularity in the marketing arena in the late 1980s
and early 1990s.6 A few cutting edge credit card banks saw a form of data mining, known as data
modeling, as a way to enhance acquisition efforts and improve risk management. The high
volume of activity and unprecedented growth provided a fertile ground for data modeling to
flourish. The successful and profitable use of data modeling paved the way for other types of
industries to embrace and leverage these techniques. Today, industries using data modeling
techniques for marketing include insurance, retail and investment banking, utilities,
telecommunications, catalog, energy, retail, resort, gaming, pharmaceuticals, and the list goes on
and on.
3.2.2. Defining the goal
The use of targeting models has become very common in the marketing industry (in some cases,
managers know they should be using them but aren‟t quite sure how). Many applications like
those for response or approval are quite straightforward. But as companies attempt to model
more complex issues, such as attrition and lifetime value, clearly and specifically defining the
goal is of critical importance. Failure to correctly define the goal can result in wasted money and
lost opportunity.
Each CRM application will have one or more business objectives for which you will need to
build the appropriate model. Depending on your specific goal, such as “increasing the response
rate” or “increasing the value of a response”, you will build a very different model. An effective
statement of the problem will include a way of measuring the results of your CRM project.
The first and most important step in any targeting-model project is to establish a clear goal and
develop a process to achieve that goal. Figure 3.5 displays the steps and their companion
chapters.
In defining the goal, you must first decide what you are trying to measure or predict. Targeting
models generally fall into two categories, predictive and descriptive. Predictive models calculate
some value that represents future activity. It can be a continuous value, like a purchase amount or
balance, or a probability of likelihood for an action, such as response to an offer or default on a
loan. A descriptive model is just as it sounds: it creates rules that are used to group subjects into
descriptive categories.
6 http://en.wikipedia.org
16
Figure 3.5: Steps for successful target modeling
Define Goal
Select Data
Prepare Data
Select and Transform Variables
Process Model
Validate Model
Implement Model
Source: Olivia Parr Rud “Data Mining Cookbook” John Wiley & Sons, Inc. (2001), pp. 5
Companies that engage in database marketing have multiple opportunities to embrace the use of
predictive and descriptive models. In general, their goal is to attract and retain profitable
customers. They use a variety of channels to promote their products or services, such as direct
mail, telemarketing, direct sales, broadcasting, magazine and newspaper inserts, and the Internet.
Each marketing effort has many components. Some are generic to all industries; others are
unique to certain industries.
3.2.3. Applying Data Mining to CRM
In order to build good models for your CRM system, there are a number of steps you must
follow. The Two Crows data mining process model described below is similar to other process
models, differing mostly in the emphasis it places on the different steps.
While the steps appear in a list, the data mining process is not linear – you will inevitably need to
loop back to previous steps. For example, what you learn in the “explore data” step may require
you to add new data to the data mining database. The initial models you build may provide
insights that lead you to create new variables.
The basic steps of data mining for effective CRM are7:
1. Define business problem
2. Build marketing database
3. Explore data
4. Prepare data for modeling
5. Build model
6. Evaluate model
7 Herb Edelstein “Building profitable customer relationships with data mining”, Two Crows Corporation
17
7. Deploy model and results
1. Define the business problem
Each CRM application will have one or more business objectives for which you will need to
build the appropriate model. Depending on your specific goal, such as “increasing the response
rate” or “increasing the value of a response,” you will build a very different model. An effective
statement of the problem will include a way of measuring the results of your CRM project.
2. Build a marketing database.
Steps two through four constitute the core of the data preparation. Together, they take more time
and effort than all the other steps combined. There may be repeated iterations of the data
preparation and model building steps as you learn something from the model that suggests you
modify the data. These data preparation steps may take anywhere from 50% to 90% of the time
and effort of the entire data mining process!
You will need to build a marketing database because your operational databases and corporate
data warehouse will often not contain the data you need in the form you need it. Furthermore,
your CRM applications may interfere with the speedy and effective execution of these systems.
When you build your marketing database you will need to clean it up – if you want good models
you need to have clean data. The data you need may reside in multiple databases such as the
customer database, product database, and transaction databases. This means you will need to
integrate and consolidate the data into a single marketing database and reconcile differences in
data values from the various sources. Improperly reconciled data is a major source of quality
problems. There are often large differences in the way data is defined and used in different
databases. Some inconsistencies may be easy to uncover, such as different addresses for the same
customer. Making it more difficult to resolve these problems is that they are often subtle. For
example, the same customer may have different names or – worse – multiple customer
identification numbers.
3. Explore the data.
Before you can build good predictive models, you must understand your data. Start by gathering
a variety of numerical summaries (including descriptive statistics such as averages, standard
deviations and so forth) and looking at the distribution of the data. You may want to produce
cross tabulations (pivot tables) for multi-dimensional data.
Graphing and visualization tools are a vital aid in data preparation, and their importance to
effective data analysis cannot be overemphasized. Data visualization most often provides the
Aha! Leading to new insights and success. Some of the common and very useful graphical
displays of data are histograms or box plots that display distributions of values. You may also
want to look at scatter plots in two or three dimensions of different pairs of variables. The ability
to add a third, overlay variable greatly increases the usefulness of some types of graphs.
4. Prepare data for modeling.
This is the final data preparation step before building models and the step where the most “art”
comes in. There are four main parts to this step:
18
First you want to select the variables on which to build the model. Ideally, you would take all the
variables you have, feed them to the data mining tool and let it find those which are the best
predictors. In practice, this doesn‟t work very well. One reason is that the time it takes to build a
model increases with the number of variables. Another reason is that blindly including
extraneous columns can lead to models with less rather than more predictive power.
The next step is to construct new predictors derived from the raw data. For example, forecasting
credit risk using a debt-to-income ratio rather than just debt and income as predictor variables
may yield more accurate results that are also easier to understand.
Next you may decide to select a subset or sample of your data on which to build models. If you
have a lot of data, however, using all your data may take too long or require buying a bigger
computer than you would like. Working with a properly selected random sample usually results
in no loss of information for most CRM problems. Given a choice of either investigating a few
models built on all the data or investigating more models built on a sample, the latter approach
will usually help you develop a more accurate and robust model of the problem.
Last, you will need to transform variables in accordance with the requirements of the algorithm
you choose for building your model.
5. Data mining model building.
The most important thing to remember about model building is that it is an iterative process. You
will need to explore alternative models to find the one that is most useful in solving your
business problem. What you learn in searching for a good model may lead you to go back and
make some changes to the data you are using or even modify your problem statement.
Most CRM applications are based on a protocol called supervised learning. You start with
customer information for which the desired outcome is already known. For example, you may
have historical data because you previously mailed to a list very similar to the one you are using.
Or you may have to conduct a test mailing to determine how people will respond to an offer.
You then split this data into two groups. On the first group you train or estimate your model. You
then test it on the remainder of the data. A model is built when the cycle of training and testing is
completed.
6. Evaluate your results.
Perhaps the most overrated metric for evaluating your results is accuracy. Suppose you have an
offer to which only 1% of the people will respond. A model that predicts “nobody will respond”
is 99% accurate and 100% useless. Another measure that is frequently used is lift. Lift measures
the improvement achieved by a predictive model. However, lift does not take into account cost
and revenue, so it is often preferable to look at profit or ROI. Depending on whether you choose
to maximize lift, profit, or ROI (return on investment), you will choose a different percentage of
your mailing list to which you will send solicitations.
7. Incorporating data mining in your CRM solution.
In building a CRM application, data mining is often only a small, albeit critical, part of the final
product. For example, predictive patterns through data mining may be combined with the
knowledge of domain experts and incorporated in a large application used by many different
kinds of people.
19
The way data mining is actually built into the application is determined by the nature of the
customer interaction. There are two main ways you interact with your customers: they contact
you (inbound) or you contact them (outbound). The deployment requirements are quite different.
Outbound interactions are characterized by your company originating the contact such as in a
direct mail campaign. Thus you will be selecting the people to whom you mail by applying the
model to your customer database. Another type of outbound campaign is an advertising
campaign. In this case you would match the profiles of good prospects shown by your model to
the profile of the people your advertisement would reach.
For inbound transactions, such as a telephone order, an Internet order, or a customer service call,
the application must respond in real time. Therefore the data mining model is embedded in the
application and actively recommends an action.
In either case, one of the key issues you must deal with in applying a model to new data is the
transformations you used in building the model. Thus if the input data (whether from a
transaction or a database) contains age, income, and gender fields, but the model requires the
age-to-income ratio and gender has been changed into two binary variables, you must transform
your input data accordingly. The ease with which you can embed these transformations becomes
one of the most important productivity factors when you want to rapidly deploy many models.
3.2.4. Summary
Effective data mining is a delicate blend of science. Every year, the number of tools available for
data mining increases. Researchers develop new methods, software manufacturers automate
existing methods, and talented analysts continue to push the envelope with standard techniques.
Data mining and, more specifically, data modeling, is becoming a strategic necessity for
companies to maintain profitability.
The recent chapter defines the term “data-mining” and main steps for implementing customer
model which is very close to analytical level of CRM. It is useful for making segmentation
models and grouping customers. Applying new data to effective models enables to use certified
business information and experience. Besides data-mining steps is fundamental modeling
customer segments.
3.3. Internet and CRM
3.3.1. The Internet Enables CRM
The theory of one-to-one marketing, or CRM, has been discussed for many years, however, the
Internet provides a uniquely cost - effective mechanism for delivering it. CRM requires a vendor
to identify its customers individually, not just as a demographic group; it must then interact with
them to understand their individual preferences and needs; and it must then customize its
products and services to meet those individual needs.
If it does this successfully, the vendor thereby raises the cost of switching. Firstly, the customer
is satisfied and is therefore less inclined to switch; but, more importantly, even if a motivation is
provided - such as a special offer from a competitor - the customer is faced with the effort of
having to teach their preferences to the new supplier.
20
As the vendor has now established a close relationship with the customer, it is in a privileged
position to identify and satisfy other needs. CRM is therefore a mechanism both for improving
customer retention and for increasing revenues per customer. It also allows the vendor to
segment its customers into the more profitable and the less profitable, and to concentrate its
efforts on the former.
In the real world, these processes of identifying and communicating with customers can be
prohibitively expensive. They have tended to be confined to vendors with a small number of
clients. Loyalty cards are an attempt to move CRM into the mass market, but the level of
personalization is typically fairly limited: your supermarket sends you appropriate coupons, and
your airline knows if you are vegetarian. Those are hardly impenetrable barriers to switching.
In contrast, the Internet reduces the interaction and personalization costs dramatically. If
customers can be persuaded to serve themselves from a Web site, then most of the work can be
done by information systems. Software can actively solicit or passively monitor each customer‟s
preferences; and the Web site itself can change dynamically to offer a different service to each
customer. The effect on customer loyalty can be huge: for example, 73 percent of Amazon‟s
book sales in 1999 were to repeat customers.8
3.3.2. Internet customer interaction
While the first wave of organizational Internet sites were little more than online brochures, it is
now crucial that Web sites give customers options for interacting with the organization. Internet
access gives customers three new ways to get in contact with organizations: Web chat, Web
callback, and e-mail.
Web chat
Web chat allows a Web site visitor and organizational representative to have a text-based
“conversation” in near real-time, by alternately typing sentences in the window provided by a
chat program. This allows organizations to offer customers one-to-one contact with a
representative without them having to disconnect from the Web, which is important for
households that use the same telephone line for Web access and voice calls. As representatives
can often conduct more than one chat session simultaneously, providing chat can also save
businesses money in comparison with staffing a conventional call centre.
Web callback
A Web callback facility allows customers to enter their telephone numbers and be called by a
representative from the organization. Furthermore, organizations can use a callback form to
establish the customer‟s interests, and ensure that a representative with relevant product
knowledge telephones the customer. This contrasts with customers being repeatedly transferred
after contacting a conventional call centre until someone who can answer the inquiry is reached.
Web users who access the Internet using their one and only telephone line cannot accept the call
until they have disconnected their Web connection. They should therefore be able to specify
when they would like to be called. However, it should also be possible for those who have
separate connections to be called immediately. 8 Gartner Group “Market Analysis. B2C Internet Business Models” (2000), pp. 5
21
Sending e-mail from the Web site
The case for allowing e-mail to be sent by customers is elementary to make: customers like it
because they do not have to wait for an available representative, as is often the case with a call
centre and organizations like it because agents can typically turn around more e-mails per hour
than they can handle telephone interactions. There are two approaches Web sites can take to
inviting written correspondence electronically: a feedback form or a hyperlinked e-mail address
which launches the customer‟s own e-mail client.
While feedback forms offer organizations a tempting opportunity to gather personal data about
their customers, making them too long or intrusive will deter some customers, and therefore be
counter-productive. Moreover, the approach of launching the user‟s own e-mail client allows
customers to be as brief or detailed as they like, but does not let the organization collect specific
data, which could be used either to route the current enquiry or for marketing purposes in the
future.
Although they are convenient in many circumstances, callback, e-mail, and chat are not
appropriate for every interaction – a customer in a hurry will often prefer direct telephoning and
to be connected with an agent immediately. Many organizations look to the Web as a means to
reduce the number of live agents they need to employ. While this should be achieved by
providing a high-quality Web experience which means few customers need to call for more
information, some organizations adopt the tactic of making it virtually impossible for customers
to telephone them – clearly unwise, for example when a customer is just about to order several
hundred dollars worth of goods and just has a small query on the delivery arrangements.
3.3.3. Metrics of CRM effectiveness
There are some reasons why performance measurement is so powerful in enhancing business.
First, measurement removes the ambiguity and disagreement that surround high-level strategic
concepts. Second, measurement provides the precise language for clearly communicating at all
levels what the organization wants to accomplish and how it intends to accomplish it. Third,
measurement allows the continual evaluation of organizational alignment on strategic objectives.
Last, measurement not only improves the probability but also speeds the pace at which change
occurs. The four perspectives are customer knowledge, customer interaction, customer value,
and customer satisfaction.
Customer knowledge
In order to adopt the current customer-centric business environment, organizations use data
mining and data warehousing technology. A major problem is filtering, sorting, manipulating,
analyzing, and managing this data in order to extract information relevant to CRM activities.
Data mining tasks are used to extract patterns from large data sets.
Technology learning is also important towards understanding customers. It is required, therefore,
to assess employee skills to use customer information effectively. Security is another basic and
critical prerequisite when dealing with customer information. Security, in particular, has been a
serious issue concerning online purchases and an impediment to the acceptance of the e-channel.
Many customers are concerned about the amount of personal information that is contained in
22
databases and how it is being used. Customers perceive safety of transactions and seller empathy
as important. Table 3.6 shows the metrics of customer knowledge.
Table 3.6: Measures for customer knowledge
Objectives Measures
Collecting appropriate customer information
Analyzing customer data
Acquiring new customers
Understanding customers needs
Improving skills of employee
Improving CRM techniques
Customer acquisitions (No.)
Number of customers (No.)
Web marketing
Page views per day
Net sales/employee (%)
Technological capacity (No.)
Frequency of hardware upgrade (No.)
R&D investment (€)
Support R&D
Data warehouse, Data mart, Data mining
Multi-dimensional analytical
Service R&D
Customer segment personalization
Recommendation
Web service
Customer profile research (€)
Security level (%)
Source: Kim, Suh & Hwang (2003), pp.12
Customer interaction
Many communication channels are developed to interact with the customer effectively. To
manage various communication channels effectively, managers make an effort to monitor the
business processes. The processes can be divided into internal and external processes. The
internal processes refer to the handling of the processes in the organization internally, whereas
the external processes describe the interactions between suppliers and customers. Internal
processes determine operational excellence and external processes determine channel
management effectiveness. The customer relationship can be reinforced by effective customer
interaction. Customer interaction has the following components9.
Contacts with organizational staff-front line and other
Outbound contact management-mail, telephone, sales visits, and deliveries
Physical service environment
Transaction - price, value, and terms
To analyze customer interaction, some important measures need to be considered, such as the
number of marketing campaigns, total cost for promotion, frequency of contents updates,
payment, response channels, and so on. Communication channels not only include classic
communication channels such as letters, fax, and telephone but also emerging new channels such
as call centers, service centers, Web sites, and virtual Internet communities. It is vital to manage
various channels efficiently and immediately. Internal processes need to connect and integrate
diverse channels effectively.
9 Jan Johansson, Jörgen Sparredal “CRM in e-Business”, Master‟s thesis, Luleå University of Technology (2004),
pp. 22
23
Furthermore, organizations need to analyze the business process to evaluate measures such as
payment methods, delivery channels, and product diversity. Customer satisfaction can be
increased by improving channel management and maximizing operational excellence. Therefore,
it is necessary to analyze such information as delivery time, response time, and product diversity.
Table 3.7 shows the metrics of customer interaction.
Table 3.7: Measures for customer interaction
Objectives Measures
Appropriate response to customer request
Integration of business processes
Improving channels management
Maximizing the effectiveness and efficiency of
organizational operations
Customizing products and services
Marketing campaign (No.)
Total costs for promotion (€)
Frequency of contents update (No.)
Number of payment methods (No.)
Number of response channel to customer inquiry (No.)
Total cost for managing channel (€)
Avg. delivery time after order fulfillment (No.)
Response time to customer inquiry (No.)
Transaction conducted by members (%)
Product diversity
Detailed product information
Timeless sales in popular product
Source: Kim, Suh & Hwang (2003), pp.12
Customer value
Customer value describes tangible and intangible benefits gained from CRM activities, which
help to arrange the relationship with the customer successfully. Customer value can be achieved
through, for example, value added by relevant information in virtual communities, a loyalty
program, and an attractive bundling of different products.
In order to determine the customer value, organizations need to analyze such information as
marketing campaigns, number of retention customers, and net sales. CRM initiatives should
provide mutually beneficial value to the customer and the organization. Current customer
profitability should be calculated, establishing a baseline and comparing new calculations to that
baseline periodically. Calculating customer value potential and using it as a guideline will be
profitable in the future. Table 3.8 shows the metrics of customer value.
Table 3.8: Measures for customer value
Objectives Measures
Improving customer retention
Profit increase
Improving customer service and support
Building an attractive virtual community
Number of retained customers (No.)
Net sales (€)
Ordinary sales (€)
Asset/employee (€)
Profit/employee (€)
Channel interface
Usability
Attractiveness
Navigation efficiency
Contents search
Consistency of site structure
Source: Kim, Suh & Hwang (2003), pp.13
24
Customer satisfaction
`Customer satisfaction is difficult to measure because it is hard to quantify the satisfaction level.
It represents a modern approach for quality in organizations, and serves the development of a
truly customer-focused management and culture. Measuring customer satisfaction offers an
immediate, meaningful, and objective feedback about customer preferences and expectations.
Among the four perspectives, the customer satisfaction perspective is the most important because
customer satisfaction is directly linked to an organization‟s profits. Service delivery via various
channels of IT applications has emerged as an important attribute in satisfying customers. Proper
CRM practices can potentially impact customer satisfaction ratings and can potentially lead to
increased customer retention. Table 3.9 shows the metrics of customer satisfaction.
Table 3.9: Measures for customer satisfaction
Objectives Measures
Improving service quality
Establishing relationship with customers
Brand image (%)
Service level (%)
Number of daily inquires (No.)
Customer satisfaction (%)
Assurance
Reliability
Empathy
Responsiveness
Tangibles
Source: Kim, Suh & Hwang (2003), pp.15
3.3.4. Web site as a measurement tool
In 2001 was stated that the advent of the Internet has added various new dimensions to marketing
communities. One dimension is that the Internet can be used to move customer through the
phases of the buying process. Making contact with potential interested customers, converting
some those into customers, and then supporting the purchase and post-purchase phase of the
supplier-customer relationship follows this.
25
Figure 3.10: Internet process model and assessment tools
Internet users attracted to
Web
Subsequent visitor repeat
purchase
Visitors placing and order
Visitors entering onto
dialogue
Active visitors seeking
information
Process model Effectiveness measurement
Web site awareness
Purchase effectiveness
Promotional effectiveness
Web site attractiveness
Loyalty effectiveness
=
=
=
=
=
No. of Web site visitors
No. repeat customers
No. placing order
No. initiating dialogue
No. seeking information
Total internet users
No. of Web site visitors
No. seeking information
Total customers ordering
No. initiating dialogue
Source: Chaston (2001), e-Marketing Strategy, pp.174
As illustrated, as potential customers progress through each stage phase of the buying process it
is theoretically possible to assess the effectiveness of the Web site. Application of the
measurement tools posited in Figure 3.10 assumes that a Web site is capable of recording all hits
and that data can be acquired about the nature of these hits. For example, to determine loyalty
effectiveness, organizations can do that by calculate the number of customers that does a repeat
purchase and divide it with the total customers ordering, in order to measure loyalty
effectiveness.
Web site measures10
Visitor count
How many people have visited a Web site.
Unique visitor count
How many unique people have visited a Web site. This measure does not double-count users
who visit a site multiple times in a period. Web sites can have difficult in accurately determining
unique visit counts, especially for those visitors who have chosen not to identify themselves by
10
Kotzab, H. & Madlberger, M. (2001), European reatailing in e-transition?: An empirical evaluation of Web-based retailing – indications from Austria. International Juornal of Physical Distribution & Logistic Management, Vol. 31, No. 6
26
not registering with a site, visitors who use multiple machines to visit a Web site, and visitors
who disable cookies in their browser preventing the system from anonymously identifying them.
Page hits
How many pages have been downloaded from Web site, or how many times a single page has
been visited in the site.
Duration
Total time a visitor spent on a page or a Web site.
Click-through-rate
What percentage of visitors clicked on a banners ad or other form of Internet marketing to visit
the advertised Web site.
Impressions
How many visitors viewed a Web page that contained an advertisement of some kind.
Registered users
How many visitors registered with the Web site.
Breakage
What percentage of visitors started interacting with a Web site (for example, by starting a survey
or purchasing a product), but chose not to complete the interaction.
Click stream
Not a measurement per se, but a source of many measurements. The click stream is the
sequential history of all interactions with a visitor on a Web site usually stored within log files in
the Web server. This behavioral data is used for example, to derive page hits, visitor counts of
images and advertisements viewed
Most of the measures within a Web site are designed to review the health of the Web site.
However, with the wealth of customer information embedded within the click stream data, many
CRM software products include the ability to tie these measures to other off-line customer
measures, for example, survey responses.
3.3.5. Summary
Internet mainly expands CRM boundaries to new customer interaction channels using site
interaction analysis systems. It is one of the fundamental method reducing company‟s expenses.
When customers transfer all or most of their buying behavior to a Web site, the Web site will
become a primary source of customer information. Customer-profiling services will then become
an important revenue source for the Internet business, to the detriment of some traditional
consumer research companies.
However, the value of those services will depend on the amount and longevity of the data
collected, and its value to vendors. For example, demographic information - such as income,
location and ages of children - is relatively enduring and will have some marketing value to a
27
wide variety of companies. In contrast, knowing who is planning to get married in the next three
months is of great short-term sales value to a relatively narrow group of vendors.
This Market Analysis examines the types of business that offer products and services to
consumers via Web sites. There are many other companies involved in business-to-consumer
(B2C) e-commerce. For example, there are companies offering hardware, software, networking
and marketing services to the B2C Web site operators. However, only companies offering
services directly to consumers are addressed here.
3.4 e-CRM
3.4.1. The Emergence of e-CRM
In order to fully understand how a unified view of the customer can be achieved through the
strategic use of e-CRM, we must make a distinction between the terms CRM and e-CRM. We
consider CRM an approach or business strategy providing seamless integration of every area of
business that touches the customer - namely marketing, sales, customer service and field support-
through integration of people, process, and technology. On the other hand, taking advantage of
the revolutionary impact of the Internet, e-CRM expands the traditional CRM techniques by
integrating technologies of new electronic channels, such as Web, wireless, and voice
technologies, and combines them with e-business applications into the overall enterprise CRM
strategy. In other words, what the traditional CRM delivers can be considered only a fraction of
an e-CRM solution (as shown in Figure 3.11).
Figure 3.11: The differences between CRM and e-CRM.
Data warehouse
- Customer information
- Transaction history
- Products information
Transaction analysis
- Customer profile
- Past transaction
- History
Target marketing
- Static service
- One-way service
- Time and space limits
CRM
WebHouse
- Customer information
- Transaction history
- Products information
- Click stream
- Contents information
Transaction analysis
- Customer information
- Past transaction history
Activity analysis
-Exploratory activities
(navigation, shopping
cart, shopping pattern
etc.)
I-to-I marketing
- Real time service
- Two-way service
- At any time
- From anywhere e-CRM
Customer dataAnalysis of customer
characteristicsCustomer service
Source: Shan L. Pan, Jae-Nam Lee “Using e-CRM for a unified view of the customer”, Communications of the
ACM (April 2003), pp. 96
The traditional CRM has limitations in supporting outside multichannel customer interactions
that combine telephone, the Internet, email, fax, chat, and so on. Unlike the traditional CRM, the
current e-CRM solution (front - office suites) supports marketing, sales and service. Integration
28
between CRM systems and Enterprise Resource Planning (ERP) systems is becoming more
common. The integration of all channels across all areas in the company is critical.
With the advancement of Web-based technology, market dynamics are driving companies to
adopt e-CRM. A fundamental motivator is the speed and unparalleled cost-effectiveness of the
Internet are making the implementation of e-business possible and relatively cost-effective.
Customer retention has replaced cost-effectiveness and cost-competitiveness as the greatest
concern of business executives today - it costs approximately five to ten times more to acquire
new customers than to retain established customers. It‟s going to take more than Web
interactions to keep the customer brand-loyal. Furthermore, over the last few years, the speed of
change in the business arena, including deregulation, has also made rapid adoption of new
technologies and flexible business strategies basic requirements for businesses. Organizations
have reengineered many aspects of their businesses, automated their back-office procedures,
streamlined their organizations, revised their product or service offerings, and invested in
marketing activities. At the same time, they also face an increasingly complex marketplace with
a high degree of competition and new entrants challenging for market share. With new channels
(the use of Internet) and online and offline markets becoming increasingly available,
technological advances have also opened up a new world of e-business opportunities. As a result
of these changes occurring, customers are better informed, more demanding, and likely to be less
loyal as their expectations are increasing faster than businesses (traditional as well Internet-
based) can deliver.
Customers of e-businesses are making the most impact as they are given more product or service
options while the cost of switching has been reduced drastically with competitors only a mouse-
click away. With the availability of the Internet, unprecedented opportunities are now available
for building sales and increasing revenue streams by expanding geographic scope, reducing
operating costs, improving procurement, productivity, and supply chain efficiency.
The final driver is the application of real-time and interactive customer interaction channels such
as the Web, email, ATMs, call centers, and wireless devices to the customers‟ non electronic
activity in today‟s fast-changing business environment. In particular, wireless technology has
emerged as a new channel for accessing the Internet and will have a large effect on customer
interaction.
3.4.2. Key Applications of e-CRM
Companies understand that e-CRM has significant potential, but they face the challenge of
building the required technology infrastructure quickly and cost-effectively. An easily
predictable reaction is to buy off-the-shelf applications, cobble together a database of Web traffic
and online purchase information, and launch an e-CRM initiative.
One of the fundamental requirements of a successful e-CRM solution is the challenge of
consolidating all customer-related information into a single view. In order to achieve this, it is
necessary to create a multichannel input stream that can take information from any of the
recognized customer interfaces and use it to populate the single view. It could then facilitate the
sharing of information between channels and meaningful cross-channel dialogue with customers.
This forms the basis for intelligent handling of future customer interactions and enables the
creation of personalized service offerings.
29
Moreover, e-CRM can enable companies of all sizes and across all industries to offer one-to-one
relationships to customers. E-CRM applications have the power to create an enormous amount of
value by allowing companies to collect, organize, and disseminate a wealth of customer
information.
The information integration application consolidates customer data and information from
different sources - transactional systems, call centers, Web sites, and ERP systems - into
integrated customer - centric information. This application allows companies to identify and
respond in a timely manner and accurately to their customers whether customers purchase
products through a physical store, a call center, or a Web site. For the most part, an incomplete
view of customers reduces their loyalty and trust. Developing an information integration
application requires multiple data models and database architectures for integration with other
back-end information systems. Since the application is dynamic, producing entities that have to
keep up with every customer‟s interaction with the company, the speed and accuracy is crucial
for enabling a true value exchange with customers.
The customer analysis application measures, predicts, and interprets customer behaviors,
enabling companies to understand the effectiveness of e-CRM efforts across both inbound and
outbound channels. The integrated customer information is used to build a business campaign
strategy and assess results. It also builds predictive models to identify the customers most likely
to perform a particular activity such as purchase an upgrade from the company. This segment
selection process improves response rates and campaign effectiveness and lowers campaign costs
by reducing the size of the original target segment. Generally, there are three major types of
customer analysis applications: online analytical processing, data mining, and statistics.
The campaign management application uses the data warehouse to plan and execute multiple,
highly targeted campaigns over time, using triggers that respond to timed events and customer
behavior. For example, a retail campaign may present a high-profit customer with a birthday gift
or send an email message promoting various specials if the customer has been silent for several
months. Furthermore, because customers are increasingly reachable through diverse
communication channels, successful e-CRM requires an application that reaches customers
wherever they are located: at home, at work, or while traveling. Hence, this application enables
the integration of multichannel communications with individual customers, and in turn increases
the likelihood of customer retention with higher customer switching costs.
The real-time decision application coordinates and synchronizes communications across
disparate customer touch-point systems. It contains business intelligence to determine and
communicate the most appropriate message, offer, and channel delivery in real time, and support
two-way dialogue with the customer. Hence, an effective real-time decision application promotes
information exchange between the company and every customer. Generally, to gain confidence
in their product purchases, customers will interact with several vendors to get relevant
information, conduct comparative analysis, and then decide which products to buy. In this case, a
real-time decision application effectively provides appropriate value-added features and
functionality. It is achieved by integrating with the four other applications.
The personalized messaging application delivers either text or HTML pages, scaled to support
very high volumes, using an automated mechanism to answer responses and drive recipients to
Web pages through traceable URLs embedded within messages. For example, a company might
30
include three traceable URLs within a particular email message and be able to monitor success at
driving an individual customer or prospect to one, two, or all three sites, and in which order.
Since personalized attention and service were labor-intensive with high costs, most companies
provide personalized attention to a small number of selected customers. But recent technology
makes it possible to personalize products and services for a large number of customers in a cost-
effective manner. To achieve one-to-one service, it builds customer profiles and enables
customized product and service offerings based on the information integration application. The
approach of this application can be classified into three major categories: rules-based,
collaborative filtering and inference model.
To reap maximum benefits of e-CRM system implementation, integration is needed between
front-office applications such as office productivity applications (including word processors or
spreadsheets) and back-office applications such as database management, ERP (enterprise
resource planning) systems, mail servers, fax servers, help desk systems, and so forth. Other
needs include having e-CRM integrated with the company‟s portal site, intranet, and extranet. In
addition to integration with Web-based technologies, part of the overall e-CRM solution can
include a wide variety of telephony equipment to receive and manage inbound and outbound
calls, automatic call distributors, interactive voice response systems, predictive dialers, fax
machines, and paging systems.
3.4.3. Management Steps for e-CRM Integration
Managerially, if one is not careful, e-CRM projects can encounter one of the following
problems11
:
Strong vendor offerings exist within the broad CRM categories of sales, service, and
marketing. CRM evolved with different vendors carving out their own niches in complete
isolation from each other;
Initial CRM efforts were hampered by the lack of a single view of the customer and have
resulted in a separate and uncoordinated customer-interaction environment. Many CRM
offerings will yield only tactical improvements; and
Lack of a single customer-centric data warehouse has caused any addition of more
customer touch points to worsen the problem.
To overcome the problems identified, we have some recommendations for organizations
considering implementing e-CRM or are already managing e-CRM. The following five
management steps are needed for effective e-CRM implementation (Figure 3.12).
11
Shan L. Pan, Jae-Nam Lee “Using e-CRM for a unified view of the customer”, Communications of the ACM
(April 2003)
31
Figure 3.12: Management steps for e-CRM
Source: Shan L. Pan, Jae-Nam Lee “Using e-CRM for a unified view of the customer”, Communications of the
ACM (April 2003), pp. 97
Identifying the existing CRM processes within the organization, both online and offline.
Knowledge of these detailed business processes is important, as it will provide answers to what
specific business benefits are sought from the customer relationship management strategy. When
conducting an audit to understand some of the existing CRM processes, it is crucial that the
implementing organization takes a customer‟s view rather than a marketer‟s perspective.
Formulating an e-CRM vision and strategy.
The second step is to formulate an overall e-CRM vision. To do this, it is important to establish
an e-CRM strategy and its specific objectives. These objectives are best generated and built upon
the existing CRM processes. A well-articulated strategy provides unequivocal direction to
employees selecting and deploying e-CRM applications.
Securing top management support.
After existing CRM processes are identified and an e-CRM vision and strategy is formulated, the
next step is to secure top management support for this project. Executive sponsorship helps the
project to have higher visibility and buy-in across all departments and functionalities. While
most recognized the importance of securing top management‟s support in major information
technology projects including e-CRM, there is no single, effective approach. Every organization
has its own IT culture and must custom tailor a strategy for sponsorship from top management
according to its own circumstances. Top management can act as a project sponsor as well as a
project champion in the implementation process of e-CRM.
Choosing appropriate technology partners.
32
With so many e-CRM vendors in the marketplace offering various capabilities in their products
and services, choosing the best technology partner becomes an important challenge for
organizations implementing e-CRM. This is difficult as some of the vendors provide excellent
product capabilities including flexibility in their applications, customizability, and scalability.
Thus, to select the right e-CRM vendor requires staying focused on the company‟s business
objectives and carefully screening product offerings using those criteria that best match the
business processes and overall e-CRM vision of the organization.
Evaluating current information systems and creating new mechanisms and metrics to monitor
and improve the process.
Once the vendor has been determined, the organization needs to evaluate current information
systems to decide whether a new system is required. Some of the questions to ask include: do
they each fit in with your overall e-CRM strategy? Identify critical areas that require immediate
attention and plan to replace any systems that don‟t fit. Furthermore, to more completely
understand current systems the information flows between front-office and back-office
applications should be assessed. On the other hand, developing new performance measures is a
necessary condition not only to speed adoption and increase overall return on investment, but
also to check the performance of the customer relationship management and improve it.
3.4.4. Summary
New strategies linking data-mining models and e-CRM lead us to effective customer response
rate. Besides it is important part of modeling which adds effective customer service techniques.
The e-CRM concept is designed to understand who the customers are and the products that are of
interest to them - only then is it possible to provide them with the products and services they
want. A more sound approach is to install a comprehensive software platform of the following
five applications that together enable the e-CRM business process. Equipped with such
infrastructure, companies can continually create significant customer value, automating the
“who, what, when, where, and how” of sales and marketing.
3.5. Data mining and e-CRM
3.5.1. Data mining in e-commerce
E-commerce is the killer-domain for data mining. It is ideal because many of the ingredients
required for successful data mining are easily satisfied: data records are plentiful, electronic
collection provides reliable data, insight can easily be turned into action, and return on
investment can be measured. To really take advantage of this domain, however, data mining
must be integrated into the e-commerce systems with the appropriate data transformation bridges
from the transaction processing system to the data warehouse and vice-versa. Such integration
can dramatically reduce the data preparation time, known to take about 80%12
of the time to
complete an analysis. An integrated solution can also provide users with a uniform user interface
and seamless access to metadata.
3.5.2. Integrated architecture
12
Suhail Ansari, Ron Kohavi, Llew Mason, Zijian Zheng “Integrating E-Commerce and Data Mining: Architecture
and Challenges”, Blue Martini Software (2000)
33
In our proposed architecture there are three main components, Business Data Definition,
Customer Interaction, and Analysis. Connecting these components are three data transfer
bridges, Stage Data, Build Data Warehouse, and Deploy Results. The relationship between the
components and the data transfer bridges is illustrated in Figure 3.13. Next we describe each
component in the architecture and then the bridges that connect these components.
Figure 3.13 Proposed high-level system architecture
Suhail Ansari, Ron Kohavi, Llew Mason, and Zijian Zheng “Integrating E-Commerce and Data Mining:
Architecture and Challenges”, Blue Martini Software (2000), pp. 2
In the Business Data Definition component the e-commerce business user defines the data and
metadata associated with their business. This data includes merchandising information (e.g.,
products, assortments, and price lists), content information (e.g., web page templates, articles,
images, and multimedia) and business rules (e.g., personalized content rules, promotion rules,
and rules for cross-sells and up-sells). From a data mining perspective the key to the Business
Data Definition component is the ability to define a rich set of attributes (metadata) for any type
of data. For example, products can have attributes like size, color, and targeted age group, and
can be arranged in a hierarchy representing categories like men‟s and women‟s, and
subcategories like shoes and shirts. As another example, web page templates can have attributes
indicating whether they show products, search results, or are used as part of the checkout
process. Having a diverse set of available attributes is not only essential for data mining, but
also for personalizing the customer experience.
The Customer Interaction component provides the interface between customers and the e-
commerce business. Although we use the example of a web site throughout this paper, the term
customer interaction applies more generally to any sort of interaction with customers. This
interaction could take place through a web site (e.g., a marketing site or a web store), customer
service (via telephony or email), wireless application, or even a bricks-and-mortar point of sale
system. For effective analysis of all of these data sources, a data collector needs to be an
integrated part of the Customer Interaction component. To provide maximum utility, the data
collector should not only log sale transactions, but it should also log other types of customer
interactions, such as web page views for a web site. To illustrate the utility of this integrated
data collection let us consider the example of an e-commerce company measuring the
effectiveness of its web banner advertisements on other sites geared at attracting customers to its
own site. A similar analysis can be applied when measuring the effectiveness of advertising or
different personalizations on its own site.
34
The cost of a web banner advertisement is typically based on the number of “click-throughs.”
That is, there is a fee paid for each visitor who clicks on the banner advertisement. Many e-
commerce companies measure the effectiveness of their web banner advertisements using the
same metric, the number of click-throughs, and thus fail to take into account the sales generated
by each referred visitor. If the goal is to sell more products then the site needs to attract buyers
rather than browsers. The ratio of generated sales to click-throughs varies by as much as a factor
of 20 across a company‟s web banner advertisements. One advertisement generated five times
as much in sales as another advertisement, even though clickthroughs from the former
advertisement were one quarter of the click-streams from the latter. The ability to measure this
sort of relationship requires conflation of multiple data sources.
The Analysis component provides an integrated environment for decision support utilizing data
transformations, reporting, data mining algorithms, visualization, and OLAP (Online analytical
processing) tools. The richness of the available metadata gives the Analysis component
significant advantages over horizontal decision support tools, in both power and ease-of-use. For
instance, the system automatically knows the type of each attribute, including whether a discrete
attribute‟s values are ordered, whether the range of a continuous attribute is bounded, and textual
descriptions. For a web site, the system knows that each customer has web sessions and that
each web session includes page views and orders. This makes it a simple matter to compute
aggregate statistics for combinations of customers, sessions, page views, and orders
automatically.
The Stage Data bridge connects the Business Data Definition component to the Customer
Interaction component. This bridge transfers (or stages) the data and metadata into the Customer
Interaction component. Having a staging process has several advantages, including the ability to
test changes before having them implemented in production, allowing for changes in the data
formats and replication between the two components for efficiency, and enabling e-commerce
businesses to have zero down-time.
The Build Data Warehouse bridge links the Customer Interaction component with the Analysis
component. This bridge transfers the data collected within the Customer Interaction component
to the Analysis component and builds a data warehouse for analysis purposes. The Build Data
Warehouse bridge also transfers all of the business data defined within the Business Data
Definition component (which was transferred to the Customer Interaction component using the
Stage Data bridge). The data collector in the Customer Interaction component is usually
implemented within an On-Line Transaction Processing (OLTP) system typically designed using
entity relation modeling techniques. OLTP systems are geared towards efficient handling of a
large number of small updates and short queries. This is critical for running an e-commerce
business, but is not appropriate for analysis, which usually requires full scans of several very
large tables and a star schema design which business users can understand. For data mining, we
need to build a data warehouse using dimensional modeling techniques. Both the data
warehouse design and the data transfer from the OLTP system to the data warehouse system are
very complex and time-consuming tasks. Making the construction of the data warehouse an
integral part of the architecture significantly reduces the complexity of these tasks. In addition to
typical ETL (Extract, Transform and Load) functionality, the bridge supports import and
integration of data from both external systems and syndicated data providers (e.g., Acxiom).
35
Since the schema in the OLTP system is controlled by the architecture, we can automatically
convert the OLTP schema to a multi-dimensional star schema that is optimized for analysis.
The last bridge, Deploy Results, is the key to “closing the loop” and making analytical results
actionable. It provides the ability to transfer models, scores, results and new attributes
constructed using data transformations back into the Business Data Definition and Customer
Interaction components for use in business rules for personalization. For example, customers can
be scored on their propensity to accept a cross-sell and the site can be personalized based on
these scores. This is arguably the most difficult part of the knowledge discovery process to
implement in a non-integrated system. However, the shared metadata across all three
components means that results can be directly reflected in the data which defines the e-
commerce company‟s business.
3.5.3. Data collection
3.5.3.1 Business event logging
The clickstream data collected from the application server is rich and interesting; however,
significant insight can be gained by looking at subsets of requests as one logical event or
episode. We call these aggregations of requests business events. Business events can also be used
to describe significant user actions like sending an email or searching. Since the application
server has to maintain the context of a user‟s session and related data, the application server is
the logical choice for logging these business events. Business events can be used to track things
like the contents of abandoned shopping carts, which are extremely difficult to track using only
clickstream data. Business events also enable marketers to look beyond page hit rates to micro-
conversion rates. A micro conversion rate is defined for each step of the purchasing process as
the fraction of products that are successfully carried through to the next step of the purchasing
process. Two examples of these are the fraction of product views that resulted in the product
being added to the shopping cart and the fraction of products in the shopping cart that
successfully passed through each phase of the checkout process. Thus the integrated approach
proposed in this architecture gives marketers the ability to look directly at product views, content
views, and product sales, a capability far more powerful than just page views and click-throughs.
Some interesting business events that help with the analysis given above and are supported by
the architecture are13
Add/Remove item to/from shopping cart
Initiate checkout
Finish checkout
Search event
Register event
The search keywords and the number of results for each of these searches that can be logged
with the search events give marketers significant insight into the interests of their visitors and the
effectiveness of the search mechanism.
13
Suhail Ansari, Ron Kohavi, Llew Mason, and Zijian Zheng “Integrating E-Commerce and Data Mining:
Architecture and Challenges”, Blue Martini Software (2000), pp. 5
36
3.5.3.2. Measuring personalization success
The architecture also supports a rules engine that runs on the web site for personalization. Rules
can be deployed for offering promotions to visitors, displaying specific products or content to a
specific visitor, etc. After the rules are deployed, business events can be used to track the effect
of deploying these rules. A business event can be collected each time that a rule is used in
personalization and these events, coupled with the shopping-cart/checkout events, can give an
excellent estimate of the effectiveness of each rule. The architecture can also use control groups
so that personalization rules are only activated for a fraction of the target visitors. This enables
analysts to directly look at sales or results for visitors when the rules were and were not
activated.
Similar data collection techniques can be used for all the customer touch points like customer
service representatives, wireless applications, etc. Collecting the right data is critical to effective
analysis of an e-commerce operation.
3.5.4. Analysis
The data warehouse is the source data of analyses in our architecture. Although dimensional
modeling is usually a prerequisite for analysis, our experience shows that many analyses require
additional data transformations that convert the data into forms more amenable to data mining.
The business user can define product, promotion, and assortment hierarchies in the Business
Data Definition component. Hierarchical information is very valuable for analysis, but few
existing data mining algorithms can utilize it directly. Therefore, we need data transformations to
convert this information to a format that can be used by data mining algorithms. One possible
solution is to add a column indicating whether the item falls under a given node of the hierarchy.
For each order line or page request containing a product SKU (Stock Keeping Unit), this
transformation creates a Boolean column corresponding to each selected node in the hierarchy. It
indicates whether this product SKU belongs to the product category represented by the node.
Figure 3.14 shows the enriched row from this operation.
Figure 3.14: data record created by add product hierarchy transformation
Suhail Ansari, Ron Kohavi, Llew Mason, and Zijian Zheng “Integrating E-Commerce and Data Mining:
Architecture and Challenges”, Blue Martini Software (2000), pp. 6
Since customers are the main concern of any e-commerce business, most data mining analyses
are at the customer level. That is, each record of a data set at the final stage of an analysis is a
37
customer signature containing all the information about the customer. However, the majority of
the data in the data warehouse is at other levels such as the order header level, the order line
level, and the page request level. Each customer may have multiple rows at these levels. To
make this detailed information useful for analyses at the customer level, aggregation
transformations are necessary. Here are some examples of attributes we have found useful14
:
What percentage of each customer‟s orders used a VISA credit card?
How much money does each customer spend on books?
How much is each customer‟s average order amount above the mean value of the average
order amount for female customers?
What is the total amount of each customer‟s five most recent purchases over 30€?
What is the frequency of each customer‟s purchases?
What is the recency of each customer‟s purchases (the number of days since the last
purchase)?
These attributes are very hard to construct using standard SQL statements, and need powerful
aggregation transformations. We have found RFM (Recency, Frequency, and Monetary)
attributes particularly useful for the e-commerce domain.
E-commerce data contains many date and time columns. We have found that these date and time
columns convey useful information that can reveal important patterns. However, the common
date and time format containing the year, month, day, hour, minute, and second is not often
supported by data mining algorithms. Most patterns involving date and time cannot be directly
discovered from this format. To make the discovery of patterns involving dates and times easier,
we need transformations which can compute the time difference between dates (e.g., order date
and ship date), and create new attributes representing day-of-week, day-of-month, week, month,
quarter, year, etc. from date and time attributes.
Based on the considerations mentioned above, the architecture is designed to support a rich set of
transformations. We have found that transformations including: create new attributes, add
hierarchy attributes, aggregate, filter, sample, delete columns, and score are useful for making
analyses easier.
With transformations described, let us discuss the analysis tools. Basic reporting is a bare
necessity for e-commerce. Through generated reports, business users can understand how a web
site is working at different levels and from different points of view. Example questions that can
be answered using reporting are:
What are the top selling products?
What are the worst selling products?
What are the top viewed pages?
What are the top failed searches?
What are the conversion rates by brand?
What is the distribution of web browsers?
14
Despite I have used a example from Suhail Ansari, Ron Kohavi, Llew Mason, and Zijian Zheng “Integrating E-
Commerce and Data Mining: Architecture and Challenges” book, these questions should be formulated depending
on specific business sphere
38
What are the top referrers by visit count?
What are the top referrers by sales amount?
What are the top abandoned products?
Our experience shows that some reporting questions such as the last two mentioned above are
very hard to answer without an integrated architecture that records both event streams and sales
data.
Model generation using data mining algorithms is a key component of the architecture. It reveals
patterns about customers, their purchases, page views, etc. By generating models, we can answer
questions like:
What characterizes heavy spenders?
What characterizes customers that prefer promotion X over Y?
What characterizes customers that accept cross-sells and up-sells?
What characterizes customers that buy quickly?
What characterizes visitors that do not buy?
Based on our experience, in addition to automatic data mining algorithms, it is necessary to
provide interactive model modification tools to support business insight. Models either
automatically generated or created by interactive modifications can then be examined or
evaluated on test data. The purpose is to let business users understand their models before
deploying them. For example, we have found that for rule models, measures such as confidence,
lift, and support at the individual rule level and the individual conjunct level are very useful in
addition to the overall accuracy of the model. In our experience, the following functionality is
useful for interactively modifying a rule model:
Being able to view the segment (e.g., customer segments) defined by a subset of rules or
a subset of conjuncts of a rule.
Being able to manually modify a rule model by deleting, adding, or changing a rule or
individual conjunct.
For example, a rule model predicting heavy spenders contains the rule15
:
IF Income > $80,000 AND
Age <= 31 AND
Average Session Duration is between
10 and 20.1 minutes AND
Account creation date is before
2000-04-01
THEN Heavy spender
It is very likely that you wonder why the split on age occurs at 31 instead of 30 and the split on
average session duration occurs at 20.1 minutes instead of 20 minutes. Why does account
creation date appear in the rule at all? A business user may want to change the rule to16
:
15
Suhail Ansari, Ron Kohavi, Llew Mason, and Zijian Zheng “Integrating E-Commerce and Data Mining:
Architecture and Challenges”, Blue Martini Software (2000), pp. 7
39
IF Income > $80,000 AND
Age <= 30 AND
Average Session Duration is between
10 and 20 minutes
THEN Heavy spender
However, before doing so, it is important to see how this changes the measures (e.g. confidence,
lift, and support) of this rule and the whole rule model.
Given that humans are very good at identifying patterns from visualized data, visualization and
OLAP tools can greatly help business users to gain insight into business problems by
complementing reporting tools and data mining algorithms. Our experience suggests that
visualization tools are very helpful in understanding generated models, web site operations, and
data itself. Figure 5.4 shows an example of a visualization tool, which clearly reveals that
females aged between 30 and 39 years are heavy spenders, closely followed by males aged
between 40 and 49 years.
3.5.5. Summary
We proposed an architecture that successfully integrates data mining with an e-commerce
system. The proposed architecture consists of three main components: Business Data Definition,
Customer Interaction, and Analysis, which are connected using data transfer bridges. This
integration effectively solves several major problems associated with horizontal data mining
tools including the enormous effort required in pre-processing of the data before it can be used
for mining, and making the results of mining actionable. The tight integration between the three
components of the architecture allows for automated construction of a data warehouse within the
Analysis component. The shared metadata across the three components further simplifies this
construction, and, coupled with the rich set of mining algorithms and analysis tools (like
visualization) also increases the efficiency of the knowledge discovery process. The tight
integration and shared metadata also make it easy to deploy results, effectively closing the loop.
16
Suhail Ansari, Ron Kohavi, Llew Mason, and Zijian Zheng “Integrating E-Commerce and Data Mining:
Architecture and Challenges”, Blue Martini Software (2000), pp. 8
40
4. Conceptual decentralized supermarket model
Various, disparate techniques for dealing effectively with customers have given way to the more
methodical approach of customer relationship management (CRM). CRM encompasses all
activities around customer „touch points‟ and aims to identify, attract and retain the most
valuable customers to enhance retention and loyalty and to sustain profitable growth. CRM
integrates all customer touch points (traditional and new) into a comprehensive data warehouse,
providing a continuously updated customer-knowledge database. CRM uses new technology
(Internet, mobile communications, database systems) to create a strong relationship with
customers. As the Internet has evolved over the last few years to become an important tool for
customer interaction, so will the wireless Internet connection evolve as one of the key
components in the CRM process.
Mobile communications is the latest in a long list of new technologies that enables companies to
interact innovatively with their customer base. We will discuss the relevance of mobile
commerce for the management of customers in a traditional supermarket industry.
The one of the most interesting part of this paper is sophisticated conceptual model demonstrates
new company - customer interaction methods using new technologies. The model concentrates
on wireless communication devices - mobile phones or PDA‟s (personal digital assistant) that
enables WAP (wireless application protocol), EDGE (enhanced data rates for GSM evolution),
Bluetooth or simple wireless internet connections. The potential of this concept shows the spread
of mobile devices at present time.
The biggest decentralized supermarkets (Figure 4.1) would be able to deliver information about
all the products in the area that want customer, despite the absence of general database of all the
products.
Figure 4.1: decentralized supermarket
Shop 1
Shop 7Shop 6Shop 5
Shop 4Shop 3Shop 2
Shop 12Shop 11Shop 10Shop 9
Shop 8
Shop 16Shop 15Shop 14Shop 13
Product 1
Product 2 Product 1 Product 5 Product 4
Product 2Product 3Product 5
Product 4Product 5Product 1Product 4
Product 3Product 2Product 3
Product 1
Let‟s look at this form from a customer‟s view in greater detail.
41
The key advantage of this concept is empowering every customer to browse for all types of
products in a one space – screen of a mobile phone or PDA – avoiding waste of time looking for
the same sort of items in separate shops. The database placing data and metadata about all
products enables to make structure of „hybrid‟ shop containing features of both ordinary
supermarket and electronic shop. There are some fundamental steps for this concept (Figure 4.2).
Figure 4.2: Steps for concept
1. Access to products
database
2. Request of products
matching fixed parameters
3. Selecting group of most
likable products
4. Getting locations of
selected products
Process of buying in
ordinary supermarket
1. Access to products database.
2. Request of products matching fixed parameters
3. Selecting group of most likable products
4. Getting the locations of selected products
5. Process of buying in ordinary supermarket
The customer is able to search for products selecting the parameters (for example color or size of
shoes) and browse with a help of hierarchy of products (Figure 4.3).
42
Figure 4.3: Hierarchy of products
Products
Tops
Athletic
Shoes Bottoms
Running
SandalsCasual
Football
WomenMen
The second advantage of this type of system is possibility to collect additional customer logs
using the same e-CRM methods (Chapter 3.5). Furthermore data-mining and CRM enables
improving the value of customer suggesting additional items or discounts.
Building customer loyalty today is not just one of many ways to boost profits; it might just
become the only sustainable way. But customer loyalty is not won through technology (even if it
might be wireless Internet connections), but through consistent delivery of a high-quality
customer experience. This can be achieved with the implementation of a comprehensive
customer relationship management strategy comprising two parts: a customer focused approach
and e-enabled processes. M-commerce applications enabled by wireless Internet connections will
become an integral part of a solid CRM strategy and will be added to the existing customer touch
points.
43
5. Conclusions
Risk is a form of behavior that has implications in many industries. I will shape my model to car
insurance industry (because it is simply to visualize) using just prime variables like risk ratio,
driver‟s age, driving experience, gender and some numbers characterizing cars. This type of
analysis is very useful developing marketing strategies.
There is simple to group customers into segments because of small amount of variables. We can
use simple algorithm to the every risk depending variable set a different multiplier and get a risk
ratio in form of numbers (Table 5.1).
Every industry has drivers that can be effectively segmented. This simple exercise can provide
direction and generate ideas for improved customer profitability.
Table: 5.1
Customer profiles
High
risk/Low
revenue
High
risk/High
revenue
Low
risk/Low
revenue
Low
risk/High
revenue
Total/Average
Total customers 7349 34618 17415 42413 101795
Age (year)
mean 23 27 42 45 37
max 90 90 90 90 90
min 18 18 18 18 18
Experience
(year)
mean 3 7 12 17 5
max 72 72 72 72 72
min 0 0 0 0 0
Percent
male (%)
mean 53,12% 55,74% 47,11% 49,38% 51,42%
max 100,00% 100,00% 100,00% 100,00% 100,00%
min 0,00% 0,00% 0,00% 0,00% 0,00%
Car engine
power (HP)
mean 214 132 120 97 121
max 300 300 300 300 300
min 60 60 60 60 60
Car value
(EUR)
mean 1.350 17.020 2.320 23.410 16.036
max 100.000 100.000 100.000 100.000 100.000
min 500 500 500 500 500
Risk score
mean 350 341 312 312 325
max 432 420 415 415 418
min 274 271 254 254 261
Customer
revenue (
EUR per
year)
mean 73 614 69 653 498
max 230 2.015 230 2.015 2.015
min 0 230 0 230 0
44
The next step should be group customers to segments depending on counted indexes (here risk
ratio and customer revenue per year). Let‟s describe every customer segment17
.
Consummate consumers
These are the most profitable customers. They are low risk and generate high revenues. Car
insurance companies can use this knowledge to offer extra services and proactively offer lower
rates in the face of steep competition.
Risky revenue
These are also profitable customers. Their main liability is that they are high risk. With higher
pricing, though, these customers can be the most profitable because they are less likely to attrite.
Business builders
These are the most challenging customers. The CRM supports some proper strategies like cross-
selling. For this segment it could be to sell additional insurance package or to tax greater amount
in a most accidental season.
Balance bombs
No one wants these customers. They are risky and do not bring desirable profits.
Companies should invest on finding more profitable customers and increasing their loyalty
(Figure 5.2). It is much cheaper to cross-sell to current customers than to acquire new ones.
Figure 5.2: Customers profiles by segment
Rev
enue
Risk
Consummate
customersRisky revenue
Business builders Balance bombs
Customer loyalty can be achieved in some cases by offering a quality product with a firm
guarantee. Customer loyalty is also achieved through free offers, coupons, low interest rates on
financing, high value trade-ins, extended warranties, rebates, and other rewards and incentive
programs. The ultimate goal of customer loyalty programs is happy customers who will return to
17
As a background information for customer segmentation and data mining methods is used theory from Olivia
Parr Rud “Data Mining Cookbook” John Wiley & Sons, Inc. (2001), chapter 8 - Understanding your customer:
profiling and segmentation
45
purchase again and persuade others to use that company's products or services. This equates to
profitability, as well as happy stakeholders.
Customer loyalty may be a one-time program or incentive, or an ongoing group of programs to
entice consumers. Buy-one-get-one-free programs are very popular, as are purchases that come
with rebates or free gifts. Another good incentive for achieving customer loyalty is offering a
risk free trial period for a product or service. Also known as brand name loyalty, these types of
incentives are meant to ensure that customers will return, not only to buy the same product again
and again, but also to try other products or services offered by the company.
Excellent customer service is another key element in gaining customer loyalty. If a client has a
problem, the company should do whatever it takes to make things right. If a product is faulty, it
should be replaced or the customer's money should be refunded. This should be standard
procedure for any reputable business, but those who wish to develop customer loyalty on a large-
scale basis may also go above and beyond the standard. They may offer even more by way of
free gifts or discounts to appease the customer.
In any industry, the first step to finding and creating profitable customers is determining what
drives profitability. This leads to better prospecting and more successful customer relationship
management. You can segment and profile your customer base to uncover those profit drivers
using your knowledge of your customers, products, and markets. Or you can use data-driven
techniques to find natural clusters in your customer or prospect base. Whatever the method, the
process will lead to knowledge and understanding that is critical to maintaining a competitive
edge.
CRM enable to make strategic plans on every customer segment, therefore using the separate
actions for each client brings destructive result.
46
Bibliography
Lawrence Ang and Francis A. Buttle “ROI on CRM: a customer-journey approach”, Macquarie Graduate
School of Management
Bearing point “Wake-up call: to fix CRM, fix the customer experience now!” (2004)
Tom Breur “Integration of Analytical CRM in Business Processes: An Application”, ING Bank
(Netherlands) (2001)
Gareth Herschel “Introduction to CRM Analytics”, U.S. Symposium/ITxpo (2002)
Gartner Group “Market Analysis. B2C Internet Business Models” (2000)
Jan Johansson, Jörgen Sparredal “CRM in e-Business”, Master‟s thesis, Luleå University of Technology
(2004)
Shobha Ganapathy, C. Ranganathan, Balaji Sankaranarayanan “Visualization strategies and tools for
enhancing customer relationship management”, Communications of the ACM (November 2004)
Herb Edelstein “Building profitable customer relationships with data mining”, Two Crows Corporation
Shan L. Pan, Jae-Nam Lee “Using e-CRM for a unified view of the customer”, Communications of the
ACM (April 2003)
Steve Gallant, Gregory Piatesky-Shapiro, Dorian Pyle “Successful customer relationship management in
financial applications”, Xchange, Inc. (2000)
Michael E. Porter “The five competitive forces that shape strategy”, Harvard business review (January
2008)
Suhail Ansari, Ron Kohavi, Llew Mason, Zijian Zheng “Integrating E-Commerce and Data Mining:
Architecture and Challenges”, Blue Martini Software (2000)
Olivia Parr Rud “Data Mining Cookbook” John Wiley & Sons, Inc. (2001)
Tricia Fox, Steve Stead “Customer relationship management: delivering the benefits”, A White Paper by
CRM (UK) Ltd and SECOR Consulting Ltd (2001)
Web pages:
http://www.SalesAgility.com
http://www.bearingpoint.com
http://www.gartner.com
http://www.e-global.es
http://portal.acm.org
