White Paper

Decision Tree Methodology A solution for Local Vs. Global Master Data conflict
- Rajendra Kumar Tamboli, Promodh Narayan Ravichandran

Abstract
Most master data management (MDM) implementations have conflicts in identifying the local vs. global master entities and their
attributes. Many such implementations face challenge in proving intangible benefits time and again, change management and
having a section of powerful business community driving enterprise MDM strategy, data governance. While most implementations
need tactical measures, some can be handled scientifically through a defined methodology. Another challenge is a situation where
each business community want its convoluted definition and acceptable sources of master data. Although such definitions and
sources of data are not interoperable across the enterprise, yet such trade-offs affect data integrity, data quality and purpose of
consolidating islands of information. We propose a decision tree methodology that attempts to address and help enterprises adopt
a strategy to evaluate and induct true master data into their operations unequivocally.

www.infosys.com

The New Normal for Master Data in Digital Age

The economic downturn may have slowed down many manufacturing
enterprises, but it has not impacted their zest of inorganic growth
which needs not only financial consolidation of acquired companies
but also product or customer data consolidation. Even manufacturing
enterprises growing organically and expanding into multiple channels
have relentless pursuit of operational excellence. In either of these
situations, enterprises have felt the need to get their arms around a
gigantic problem called master data. Any program across the value
chain, whether it is SCM, CRM, Big Data, Analytics, or WMS, can easily
get derailed without proper MDM investment and guiding principles.
In this digital age, the master data information plays an important
role in ensuring its consistent behaviour across enterprise through
various touch points or channels. This is the new normal required to
succeed in digital age.

Why New Normal becomes a Battleground for

Master Data?
The aspect of deciding what product or customer entity qualifies
as master data becomes an issue especially in a multi-business unit,
multi-geography, multi-location and multi-channel enterprise. Each
local division or even business unit wants its own version of a key
master data set. For example, manufacturing business unit may define
the master data set in terms of size, specification, materials and so on,
while marketing may have responsibility for package design and size.
Similarly, for different business units, market segments may be divided
differently, which may lead to conflicts while defining customer master
attributes. In a bid to keep all stakeholders (so called data guardians
of the enterprise) aligned to a common definition, the enterprise acts
on the nearest and not the newest master data definitions. As a result
such enterprises suffer from:
Inconsistent semantics of product definition or customer
attributes
Invalid and inconsistent values of master data

2 | Infosys

Why Traditional ERP Methods for governing

Global and Local Master Data fail?
Many ERP packages have in-built mechanism to implement global
and local master data with set of restrictions, security and flags to
enable/ disable mastering attributes as Global and Local. But they
lack the qualification process or principle which forms the basis to
mark a particular attribute or characteristic as Global Master Data
or Local Master Data. This is because the negotiations with business
community on global vs. local master data happen outside the
system which later gets converted into configurations. While this
argument may be extended to MDM packages also, the impact of
implementing a guiding principle or techniques through MDM is far
and wide reaching as it becomes the channel for data distribution.
Many order management applications and call centre or support
applications manage customer master data locally. However, they
lack standard definitions and consistent information for various
customer master entities and attributes.

Is there an alternative to induct Global vs.

Local Master Data?
Definition of Global vs. Local Master Data
Global master data attributes are standard across systems,
divisions, and common to the global enterprise. In some cases,
the global attributes for a business unit, may appear local from the
enterprise perspective. They can be referred to as Glocal (global
+ local) attributes and may end up as local master data as far as
configuration is concerned. Finally attributes which are restricted
to a particular function or a department within a business unit
whose values may vary from one department to another within the
same department or function or even business units are referred
to as local master data.

Figures and tables given further in the article illustrate the decision tree methodology. Figure 1 and figure 2 illustrate the product domain
entries and customer domain entries respectively.

Local
Versions

Figure 1

Master Data
Entities

Global and
Integrated
Process

Geographic
Variations

Product domain entities

Local vs. Global

Local
Compliance and
Developments

Organization
Structure
Enterprise
Standards and
Definitions

Sales
Attributes

Figure 2

Global Elements
Local Elements

Customer
Core
Attributes

Global
Hierarchy

Local Hierarchy

Customer domain entities

Local vs. Global

Financial
Attributes
e.g. Currency

Customer
Contact
Details
Contacts

Global Elements
Local Elements

Infosys | 3

Alternative to Induct Master Data: Decision Tree Methodology

Figure 3 illustrates the use of decision tree methodology for the product master data.
Item Master

Figure 3
Decision Tree Methodology
for Product Master

MDM Global Data

Standards Team

Issue

Analyze for Master

Data Qualification

Qualified as Master

tion
spe
cifi
ca

Yes

Yes

Document
Type

and

ial

Global Master Data

Yes

aw
ing
s

m Mater

s, D
r

Yes

Do

cum
ent

utes rela
ted to Ite

d to
la te
t

Da

te

ela

e
ut

ib

ttr

ta

Da

rce

U
to

d
late

rial

ate

to M

ou
Res

ng

MRP Group

nni

pla

es r

but

ttri

A
and

re

su

ea

fM

s re
nd
Att
ribu
te

Data an
d Attrib

nM
ast
er

Raise

n
aa

Data and Attributes related to

Shop floor, Manufacturing Process and Plant specific

Segregate
Master Data Type

Data
Da

Data

tion
Change in Informa
t
ues
q
e
rR
n
Use
o
i
d
t
En
ca
pli
Ap
w
Ne

Yes

Standard UOM

o
nit

sr

dA

aa

Dat

atio

No

form

as

of In

fied

rce

Rejected

al i

Sou

Qu

New

Iss
ue
n

ot
R

ais

ed

Resend Data to
Originator

Data and Attributes related to Item Definition

Item Material

and A

ta

ttribu

dA

tes re

ttr

lated

tes re

ttribu

and A

an

lated

Manufacturing
Plant

to Sto

rage

ibu

Yes

Storage
Location

te
Pro s re
cu late
rem d
en to
t

Yes
Yes

tribu

to Dis

Purchasing
Group

Local Master Data

tion,

Yes

ing

Shipp

Warehouse

Figure 4 illustrates the use of decision tree methodology for customer master data.
Figure 4
Customer
Master

Decision Tree Methodology

for Customer Master

Yes

of

Inf
orm
ge in
at
Inform
ation ion
est

qu
r Re
Use
End
on
i
t
lica
pp
wA
Ne

Resend Data
to originator

MDM Global Data

Standards Team

Rejected

er

ast

Con
fl

ict e

M
on

xist

s?

Analyze for
Master Data

me

alifi

ed

Ma

ste

t
nS

Shippin

Cus
t

g relate

om

Sales Hierarchy etc.

Other
Hierarchy

d attrib

er

m
to

s
Cu

er C
ore

Yes
Contact
Person

Custom Hierarchy e.g.

Master Data
Local / Global ?

Customer
Organization
Structure

e
tur
ruc

acts
sto
ont
Cu
er C
tom
Cus

as

utes

a
Fin
ial

nc

Fin
anc
ial
Att
ribu
tes

Yes
Yes

Shipping

ts
un
co
Ac
Customer
Financial Info

Financial
Accounts

4 | Infosys

Global
Master Data

Yes

ati

niz

ga

r
rO

Qu

Cor
eC
ust
Cu
om
sto
er A
me
ttrib
rD
ute
s
em
og
rap
hic
de
tai
ls

rce

flic

ou

Co
n

Chan

wS

No

Ne

Demographic

Yes

Yes
Yes

Local Master
Data

Decision Tree Guiding Principles and Criteria

Product Domain
Master Data Type

Definition

Global

Local

Recommendation

Pros

Cons

Item Master

An item record comprising

information required
to manage aspects of
Procurement, Sales, Costing,
Storage and commercial
trade

Yes

Yes

All basic, user defined master

attributes will be maintained
globally for organization ,
business unit in Product
Hub and all application
specific, localized master and
operational attribute s to
be maintained in local item
master of ERP

Globally maintained
items avoid duplicity
Easier movement
of stocks across
intercompany plants
for some item

None

Item Material type

All materials like raw

materials , semi finished or
finished goods sharing some
basic attributes

Yes

No

Same item type could be

followed across all locations

Easier management,
control, training,
reporting, knowledge
sharing and future
improvements

Local
requirements
may get
ignored

Manufacturing
plant

A plant is an organizational
unit for dividing the
enterprise according to
production , procurement,
maintenance and material
planning . It is a location
where raw materials are
produced or converted to
finished goods

No

Yes

For a business unit plants

should remain the same with
exceptions. Business logic
for creating a plant should
be same or similar across
locations within the BU.

Easier management,
control, training,
reporting, knowledge
sharing and future
improvements as within
BU processes are same
and similarities can
found with different BUs

Document type

A document that
standardizes creation of
sourcing processes and helps
purchaser predefine features

Yes

No

Same documents process

should be followed across
locations for the same item
category

Easier management,
control, training,
reporting, knowledge
sharing and future
improvements

Local
requirements
may get
ignored

Storage location

An organizational unit
allowing differentiation
between various stocks of
items in a plant

No

Yes

Business logic and naming

taxonomy for creation of
storage location across
business units should be same

Can accommodate local

inventory and freight
requirements

None

Purchasing Group

A key for a buyer or group

responsible for purchasing
activities

No

Yes

Based on local business logic

but should be similar across
locations

Can apply local

purchasing or
organization structure
requirements

None

Customer Domain
Master Data Type

Definition

Global

Local

Recommendation

Pros

Cons

Customer Master

An individual or organization
type Customer who buys
product / avails services

Yes

No

All basic information,

attributes related to
Customer will be maintained
globally in Customer Master

Improved quality
Less duplicates
Consistency

None

Contacts

An individual who is primarily

responsible for the Customer
relationship and manages
payments, shipments etc

Yes

Yes

Primary Contacts are defined

globally in Customer master.
However individual BUs may
have local Contacts

Contacts can be defined

in both Customer
master or in CRM / Order
management applications

Contacts may
be duplicated if
defined locally

Hierarchy

Organizational structure of
the Customer

Yes

Yes

Global hierarchy can be

maintained in Customer
master and can have
secondary hierarchies
defined for individual BU

Easier reporting at
organization or BU
level
Flexibility to define
custom hierarchies

Locally defined
hierarchies could
be duplicates and
required to be
merged

Shipping
Information

Shipment related information

like shipping method,
customer shipping preference
etc.

No

Yes

Definition must be based

on standard set of values,
however individual BUs
to maintain the shipping
information consolidated
information available in Hub

Flexibility to reuse and

define shipment details

None

Financial
Information

Customer financial
information required for order
management applications e.g.
credit limit, financial accounts

No

Yes

This is to be maintained at
BU level, but consolidated in
Customer Hub

Single view of all financial

information available in
customer master

Aggregation
reporting might
be required if
maintained locally

Infosys | 5

Conclusion
While many may argue the global and local
definitions vary based on the enterprises
flavour and unique processes, but that is
exactly the conflict which decision tree
methodology tries to address. Instead of
breakdown in negotiations during the
conceptualization, this methodology
helps establish a scientific benchmark
and informed decision making. The
above shown illustrations are only sample
scenarios but Infosys has full blown
capability to implement a decision tree
methodology to any manufacturing
enterprise or even further extend it to
retail and supply chain processes. Also,
the decision tree illustrated here only
addresses customer and product master
data domain; however the methodology
can be applicable and extended to
any master data domain. In fact we
have already tested and proven this
methodology in some of our projects in the
manufacturing vertical.

6 | Infosys

About the Authors

Rajendra Kumar Tamboli (rajendra_tamboli@infosys.com) is a senior technical
architect in the Consulting & Systems integration unit within manufacturing vertical
at Infosys. He has more than 12 years of experience as consultant and architect
in functional areas like master data management and customer relationship
management. He has delivered MDM and data quality solutions to various clients
in manufacturing space.

Promodh Narayan Ravichandran, now an ex-Infosys employee, has worked in

the capacity of a principal consultant in the Consulting & Systems integration unit
within manufacturing vertical at Infosys. He has spent more than 12 years in solution
consulting and implementation in function areas like MDM, core merchandising,
supply chain and logistics. He has consulted with clients in manufacturing, retail,
and consumer packaged goods (CPG) across USA and Europe.

Infosys | 7

About Infosys
Infosys partners with global enterprises to drive their innovation-led growth.
That's why Forbes ranked Infosys #19 among the top 100 most innovative
companies. As a leading provider of next-generation consulting, technology
and outsourcing solutions, Infosys helps clients in more than 30 countries
realize their goals. Visit www.infosys.com and see how Infosys (NYSE: INFY),
with its 150,000+ people, is Building Tomorrow's Enterprise today.

For more information, contact askus@infosys.com

www.infosys.com

2013 Infosys Limited, Bangalore, India. All Rights Reserved. Infosys believes the information in this document is accurate as of its publication date; such information is subject to change without notice.
Infosys acknowledges the proprietary rights of other companies to the trademarks, product names and such other intellectual property rights mentioned in this document. Except as expressly permitted,
neither this documentation nor any part of it may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, printing, photocopying, recording or
otherwise, without the prior permission of Infosys Limited and/ or any named intellectual property rights holders under this document.