Sie sind auf Seite 1von 8

An Introduction to Enterprise Data Strategy

By Richard Fook
PepsiCo International

May 2006

Copyright PepsiCo International 2006


All other company and product names are trademarks of their respective owners

An Introduction to Enterprise Data Strategy

May 2006

Introduction
Today, many organizations are beginning to understand the importance of data to their
business. They know that access to good quality data in a timely fashion can enhance the
performance of their entire business, and can be a significant competitive advantage in
the marketplace. With better data, an organization can fully understand its customers,
deliver improved products and services, and seek new opportunities for growth.
At the same time, improvements in technology infrastructure along with truly enterprisewide computer systems, such as SAP, are finally in a position to be able to deliver on the
capability of capturing the right data and using it appropriately across the organization.
With this increasing awareness and capability we are able to recognize, as never before,
that data has become a company asset. Like any other kind of asset, how it is treated has
a considerable impact on the financial health of the company, depending on if it is used
wisely or poorly.

What is an Enterprise Data Strategy?


An Enterprise Data Strategy is the set of plans and programs that an organization
develops to manage its data assets. The development and implementation of an
Enterprise Data Strategy is a business-driven initiative, which requires leadership and
participation at the highest levels. To do this, an organization must understand its own
unique requirements, the strategic direction of the business, the potential impact of
regulatory requirements, and create the organizational capability to support all of these
initiatives.

s
in)
ive ha
iat ly c
nit pp
s I su
es ent
sin fici
Bu e, ef
l
m
fro f sca
fi ts o
ne ies
Be nom
o
ec

g.

Be
n
(e. efit
g. s f
int rom
eg
rat De
ion liv
, fa er y
ste Ef
r in fic
for ien
ma cie
t io s
n)

(e .

Fig 1. How an
Enterprise Data
Strategy delivers
benefits to the
organization

Enterprise
Data
Strategy

Benefits from Risk Avoidance


(e.g. SOX, information privacy, data quality)
Source: Developing a Sustainable Enterprise Data Strategy John Ladley

Page 2

An Introduction to Enterprise Data Strategy

May 2006

In the past, data strategies were formed because of the need to manage the abundance of
data that have came from computer systems. However, these strategies predominantly
focused on technology solutions that improved data integration and speed of delivery.
With this new approach of managing data as an asset, broader strategies must be formed
that address the role of data in supporting future business initiatives and managing risk
(see Fig. 1).
The development of an Enterprise Data Strategy should not be regarded as an information
technology initiative, nor should it be viewed as a project that has a clearly-defined
beginning, middle and end. A proper strategy should give rise to a series of continuous,
incremental programs that create a lasting impact in the organization. These programs
may initiate their own projects and tangible deliverables, many of which may be executed
by the information technology function, but there are also many programs that should
focus on integration with business processes and building the proper support
organizations. All of these must be identified, prioritized and implemented in alignment
with the strategic direction of the organization.
For those organizations that have implemented or are in the process of implementing an
ERP software package, it is incorrect to think that an Enterprise Data Strategy is not
required. The principles embodied by an Enterprise Data Strategy are technology-neutral
and can be applied to any business environment, even one that is operated by pen-andpaper. Without doubt, an ERP software package will have a significant effect on the
choice of tools and technologies to implement the strategy, but the principles are
consistent and universal.
There are four main pillars that support an Enterprise Data Strategy.

Understanding the Current Landscape

Managing the Data

Using Data to Make Better Decisions

Ensuring Data Quality and Relevance

These issues are interrelated and interdependent, and in most cases work is done in
alignment with current enterprise business initiatives or business process transformation.

Understanding the Current Landscape


Any organization must first seek to understand what data assets it has, the meaning of
those assets, how they relate to each other, and how they are (or should) be used to
serve the enterprise. The output of this part of the Enterprise Data Strategy is a set of
models that describe the current data landscape, which can then be used as the
foundation for the tools and processes that support the management of and access to the
data.

Page 3

An Introduction to Enterprise Data Strategy

May 2006

The most common deliverable is to develop an enterprise data model (see Fig. 2 for an
example). This represents the context and content of an organizations data, and is
created from a conceptual understanding of the data used during each step of every
business process. By creating a data model, an organization can define a common
language and terminology for all of the data elements, understand the structures that
surround them, and understand the relationships between them.

Fig. 2 An example
of a Data Model
for a Sales process
that involves
Customers, Stores,
Items, Packaging
and Orders

Customer Group

Location

Packaging

Customer Group
Group Name
Global Customer?

Location Id
Delivery Address
Delivery Notes
GPS
Open Hours

Packaging Code
Packaging Name
Returnable?
Recyclable?

Customer
Customer Number
Customer Name
Billing Address
Contact
Customer Group

Item
Customer X Location
Customer Number
Location Id
Effective Date

Item Code
Item Description
Packaging Code
Standard UoM
GTIN

Order

Order Detail

Order Number
Customer Number
Location Id
Order Date

Order Number
Line Number
Item Code
Ordered Qty

An organization should also develop further models that describe other key aspects of the
data landscape. For example, a measures model is a catalog of metrics that describes
how the business is measured, and what data elements are required to calculate those
metrics. Another type of model is a scenario model or process model, which describes
where and how data is used in the execution of the business processes. These models
should include paper-based and manual systems, as well as computer systems.
These models are used to provide an accurate picture of the current data landscape at
any given moment, but as the business changes through new processes, acquisitions,
or divestitures then the models must be continuously updated and maintained to stay
aligned. They are also used as the starting point for understanding changes to existing
systems and processes and driving those changes across the enterprise.

Managing the Data


Any Enterprise Data Strategy must address the challenges of data management and
integration. Proper data management processes will ultimately lead to better quality data,
which in itself can provide significant cost reductions to an organization. By increasing the
accuracy of data, such as customer addresses and product codes, fewer mistakes are
made, and the overhead and waste caused by the work required to remedy those
mistakes can be minimized or even eliminated.

Page 4

An Introduction to Enterprise Data Strategy

May 2006

The leading principle in data management is to remove as many potential points-of-failure


as possible from data entry, data storage and data transmission.

Avoid repetition of effort to enter data, which is prone to human error

Minimize duplication of data, where possible

Standardize data exchange through consistent processes or technologies

An example of a scenario which might introduce data quality problems is if a customers


address is stored in two different databases, with two separate processes run by two
different people when making any kind of change.
To support this principle, new tools are being developed that provide centralized data
management processes for the most important data elements used by the business, such
as Customer, Product and Supplier. These important data elements are known as Master
Data, and the tools to manage them are known as Master Data Management systems
(see Fig. 3).

Fig. 3 How Master


Data Management
interacts with the
business and
system processes
to maintain data
quality

1. Business Users
submit requests to
add or update
Master Data

2. Data Owners
receive and review
requests and grant
approval

Master Data
Management

4. Audit information
and metrics are
captured to control
and manage risk

ERP

3. Enterprise Systems
are automatically
updated and aligned
with Master Data

A Master Data Management system combines a single point-of-entry for business users,
automated and interactive approval processes, two-way integration with information
systems, and audit tracking capabilities. Nevertheless, the key to the success of any
Master Data Management program is the participation and involvement from the business
to ensure that the correct processes are being followed, and that the correct data is being
entered.
Today, organizations also have to align their internal data with external data standards
and definitions, so that they can participate in a direct exchange of information with their
trading partners. Global standards, such as the Global Trade Identification Number
(GTIN), are defined by a neutral body to describe the critical business elements and are
aligned through a process known as Global Data Synchronization. Any Enterprise Data

Page 5

An Introduction to Enterprise Data Strategy

May 2006

Strategy should also meet the challenge of enabling this process and managing the data
in accordance with the global standards bodies.

Using Data to Make Better Decisions


Ultimately, an organization goes through an Enterprise Data Strategy initiative because of
a desire to access accurate information at the right level in a timely fashion. It is this
capability that will provides the long-term strategic and competitive advantage that

Fig. 4 The
evolution of
Business
Intelligence and
Analytical
capability and
value over time

Value

supports future business growth and development.

Effectiveness
Performance
Management

Efficiency
Information
Access

Reporting

Enterprise
Information

Intelligent
Enterprise

Information on Demand
Embedded Analytics
Information available to all

Business Performance Management


Advanced Analytics and Mining
Enterprise Alignment

Information integration
Enterprise Data Warehouse
Single version of the truth

Improved access and delivery


Departmental Data Marts
Subject area Data Warehouses

Individual or system focused


Spreadmarts in abundance
Management reporting

Time

Historically, most organizations have accessed their data via simple management
reporting in isolated systems or, more commonly, through manually-created spreadsheets.
The problem with this approach is that data becomes distorted and devalued because of
the different interpretations by different people and systems. An Enterprise Data Strategy
will start the journey that ultimately leads to a new level of analytical capability that gives
rise to the Intelligent Enterprise (see Fig. 4).
A critical step on the journey to achieving this goal is the development of an Enterprise
Data Warehouse. This is a repository of the key transactional data that is captured as a
result of executing the business processes over time. Based on the enterprise data
models captured earlier, this repository provides a consistent and thorough view of the
organization at all levels. Patterns can be identified that were not previously visible from a
divisional, functional or departmental viewpoint.
The Enterprise Data Warehouse provides the single version of the truth that is the
starting point for the dissemination of information to the business users. With the right
tools in place, it is now possible for users to query, extract and mine the data to
understand past and present events and ultimately analyze and predict consequences to
drive future business opportunities.

Page 6

An Introduction to Enterprise Data Strategy

May 2006

Ensuring Data Quality and Relevance


Any organization must support their Enterprise Data Strategy initiative with a businessdriven governance program. This program is dedicated to defining and maintaining quality
standards, integrating data management processes with the business and managing data
security (including issues of privacy and regulatory compliance).
A data governance program becomes especially important in the context of enterprisewide system initiatives, such as an SAP implementation. A great deal of effort is spent
cleansing and preparing the data for such an initiative, but without a governance program,
the quality of data rapidly declines after the initiative goes live. This deterioration of data
quality leads to frustration and mistrust among the users, and threatens the sustainability
of the original initiative. It can take a considerable amount of time and effort to return data
quality back to acceptable levels, leading to a phenomenon known as the Valley of

Fig. 5 Maintaining
data quality in a
large Enterprise
Initiative

Data Quality

Despair (see Fig. 5).

Enterprise
Initiative
Go-Live

Maintaining quality through


data governance program
Desired level of quality

Valley of Despair

Data Readiness
And Migration

No
Governance

COE

Reactive
Data
Governance

Proactive
Data
Governance

Time

This support structure for a data governance program is often referred to as the Data
Management Office. It should not be confused with the Program Management Office
function which typically provides project management services and controls to the
technology group of an organization. As new business requirements arise, the Data
Management Office works directly with the business and technology groups to implement
new or updated data management processes. They are the primary agents for the Master
Data Management tools and are the stewards for maintaining and implementing
enterprise-wide data standards.
By putting the proper data governance programs in place at the right time, data quality can
be sustained at a high standard with lower costs from the very beginning.

Page 7

An Introduction to Enterprise Data Strategy

May 2006

Summary
In this white paper, we have discussed the benefits of an Enterprise Data Strategy, and
the four main pillars that support the strategy:

Understanding the Current Landscape (by creating a set of models for the enterprise)

Managing the Data (using proper tools such as Master Data Management)

Using Data to Make Better Decisions (starting with an Enterprise Data Warehouse)

Ensuring Data Quality and Relevance (with the Data Management Office)

In the next white paper, we will review the specifics of the PepsiCo Enterprise Data
Strategy and discuss some of their key choices in support of their strategy, the impact of
Project 1-Up, and how this strategy affects PepsiCo International

For more information about this white paper, please contact:


Richard Fook
Data Architect
PepsiCo International
100 Summit Lake Drive, 4-400
Valhalla, NY 10595 USA
Tel: +1 914 801 1893
Fax: +1 914 801 1995
E-mail: richard.fook@pepsi.com

Page 8

Das könnte Ihnen auch gefallen