DATA WAREHOUSING AND CRM

A producer wants to know.

Which are our lowest/highest margin customers ? What is the most effective distribution channel? Who are my customers and what products are they buying?

What product prom-otions have the biggest impact on revenue? What impact will new products/services have on revenue and margins?

Which customers are most likely to go to the competition ?

Data, Data everywhere yet ...

I cant find the data I need
data is scattered over the network many versions, subtle differences

I cant get the data I need

need an expert to get the data

I cant understand the data I found

available data poorly documented

I cant use the data I found

results are unexpected data needs to be transformed from one form to other

What is a Data Warehouse?

A single, complete and consistent store of data obtained from a variety of different sources made available to end users in a what they can understand and use in a business context.

[Barry Devlin]

What are the users saying...

Data should be integrated

across the enterprise Summary data has a real value to the organization Historical data holds the key to understanding data over time What-if capabilities are required

What is Data Warehousing?

A process of transforming data into information and making it available to users in a timely enough manner to make a difference
[Forrester Research, April 1996]

Information

Data

Data: Good for Business

Data is a fact or multiple fact or a set of

values that is raw material stored in a structured manner. With interpretation and human intelligence applied it becomes useful information. A database is a set or collection of these structured files and managed by a Database Management System. With DBMS one can access data in a variety of ways.

Data structure
DS important when deciding what is important to

you in the data nugget. Example online surveys Entities and attributes: Entities:

Data defined by a common group of characteristics that are

of interest to the business. It could be a person, place, thing, concept, event.

Attributes
Are descriptors attached to entities.

In RDBMs entity is a table and the attributes as a

column in a table.

CRUD matrix
Developing a CRUD matrix , once the entities

and attributes are defined , is the first step in linking business and data. Create, Retrieve, Update and Delete Creates 360 degree view of customer It plays two vital roles:
When functions are mapped to the data useful

relationship show on the matrix Missing business processes or data entities are uncovered as are data and process redundancies.

Metadata
Data about data Primary purpose is to be able to describe and communicate business and technical information to persons within the organization. These are classes that are used to group the data

in organized and understandable fashion. For example Last name, first name phone no. is metadata. Greenberg, Paul,571-213-6988 is data.

Data Quality
Identify good data, bad data Good data means accurate and non

redundant data. Bad data is inaccurate and repetitive data. Can clog up your system, waste time and slow down system. Maintaining quality of data not easy enough especially when you have hundreds of gigabytes or even terabytes of data.

Seven characteristics of Data Quality

Accuracy: Data represent reality or a verifiable

source Integrity: Is the structure of the data and relationships among entities and attributes maintained consistently. Consistency: Are data elements consistently defined and understood. Completeness: Is all the necessary data present Validity: Do all values fall within acceptable ranges defined. Timeliness: Is data available when needed. Accessibility: Is data easily accessible

Customer satisfaction is ensured if quality data is available.

Data Models
Conceptual data models:
Is the first thinking about the data in the early phases of

system development. Data requirement are scoped from a business standpoint. This stage the CRUD matrix is developed.

Logical data model:

Follows the conceptual data model Technical theories of data architecture are used here. Normalization is an example to prevent redundant data. No database is created

Physical data models:

Is the mapping of database design data groupings, into

physical database areas, files, records, elements, fields, and key adhering to constraints of the software.

Datawarehouse
An enterprise wide data collection that is organized

around subjects , collected from multiple sources and centrally merged into a coherent body over time. Datawarehouse have time period associated with it. Various functions include:
Data can be extracted from operational database. It is processed and cleaned to eliminate incorrect and

redundant data or add missing data. Then loaded on RDBMS like oracle 9i and DB2 and analytical operations are run using analytical tools like SAS or OLAP

Possible Problems

Architectural and human issues can turn a database into a place for repossession and dismantling. 80% of time is spent on extracting, cleaning, loading no time for application. Incompatibilities in the system that are feeding the datawarehouses. Data not being captured turns out to be important Query and reporting tools that are easy to use everyone actually use and reports request overload. Conflicting business rules among the users-same calculations performed differently. Data homogenization Heavy overhead Security not assignable without process driven approach Lack of customer management against over concerns with resource optimization. High maintenance system.

Reasons for Failure

Design: bad architecture, data driven

methodology rather than business driven plan Technical: Ignoring the obvious issues related to query volumes and network traffic Procedural: using methodology which doesnt have prototypes, proof of concept Sociological: he who controls data controls the world.

Real Time Databases

Customers run their own universe. Hence companies need as current data as

possible to make appropriate decisions. Data is perishable commodity : the older it gets, the less relevant it is.

States of Data
Data marts: focus on a single subset of

enterprise data. Datamarts are smaller and use software to summarize , store and analyze data that may be useful to you someday. Operational Data Store: analogous to short term memory. They are interim areas that is used to store continuously updated recent data gathered through the course of a business day.

States of Data contd

Corporate Information Factory: it combines the

producer and consumers of information into a single architecture. CIF uses datawarehouses or ODS as the assembly point for data captured from operational systems and business processes of the subject company. To maintain CIF effectively four operations and administration function needs to be implemented. System management Data acquisition management Service management Change management

States of Data
Data vaults: are linked directly to the business

processes that move the company. Functional areas are indentified such as financial or marketing, then the unique subject areas within units such as billing or campaign management, then then topic areas such as invoicing or direct mail.

Customer Relationship Management Process

Capture Capture Customer Customer Data Data and Measure and Measure Results Results

The Customer Data and Measure

Results

Capture Customer

The CRM Dynamic

Take Action to Capture Customer Enrich the Customer Data and Measure Relationship Results

Store Data, Mine Capture Customer and Make Data and Measure Information Results Accessible

Capture Manage Build and Customer Data and Measure Customer Value Results

The building blocks of CRM allow an organization to manage this cycle and use the knowledge on customers to enhance the Life Time value of the customer portfolio. No organization has perfect information on its customers. Knowledge of customers is continuously enhanced through the CRM dynamic.

Customer Relationship Management is a ongoing, dynamic learning process for an organization

Implementing CRM must be approached from an Integrated Perspective

Capture Capture Customer Customer Data Data and Measure and Measure Results Results Taking action to improve the relationship without measuring the results provides no evidence of success or failure and limits the opportunity for learning. Capturing gigabytes of customer data in disparate operational systems that are next to impossible to access may render the data useless.

The Customer Data and Measure

Results

Capture Customer

CRM without an Integrated Approach

Store Data, Mine Capture Customer and Make Data and Measure information Results Accessible A data warehouse full of data without the tools to extract knowledge is nothing more than expensive inventory. Sophisticated mining tools only produce results only as good as the data they mine.

Take Action to Capture Customer Enrich the Customer Data and Measure Relationship Results

Capture Manage Build and Customer Data and Measure Customer Value Results

Implementing new technologies without the knowledge on how to enrich the relationship is likely to yield a return below the cost of the capital expenditure.

Developing insights on how to improve the value of the customer relationship without having the infrastructure to take action has no impact on the bottom line. In addition, there is no opportunity to test the theoretical analysis.

All areas must be implemented, to some degree, to effectively manage the customer relationship. When pieces are implemented in isolation, the benefits are less than overwhelming.

The Building Blocks of CRM

Data Capture
Customer Touch Point Integration Market Research External Databases

Data Warehousing

Knowledge Management

Enabling Technologies

EIS

Segmentation Customer Profitability Data Mining Statistical Modeling

Call Centres Sales Process Automation e-Business

OLAP

Data Cleansing

MetaData

Organization People Deployment and Support The building blocks of CRM are the things that need to be in place for an effective Customer Relationship management program

Data Capture and Warehouse

What Data do we capture on Customers?
Derived Data
Segments Profitability Life Time Value Intentions

Customer Behaviour
Usage Profile
Migration in Usage Acquisition Information Loyalty / Switching

Customer Interactions
Inbound Contact Campaign History Outbound Contact

Customer Profile
Product / Service Preferences
Demographics / Firmgraphics Attitudes

External Data
Geodemographics Census

Product Portfolio

Householding Base Data

The Customer Data Model

Data Warehouse for Decision Support & OLAP

Putting Information technology to help the knowledge worker make faster and better

decisions
Which of my customers are most likely to go to the

competition?
What product promotions have the biggest impact on

revenue?
How did the share price of software companies

correlate with profits over last 10 years?