View Point
Dealing with Challenges of Product,
Category and Channel Growth in Retail
using Oracle PIM
Manikandan Munikrishnan, Nandakumar Balasubramanian, Vivek Kale
Abstract
What is your view of a product data hub? Is it that of a database which consolidates
product data from various applications to provide a single view of the product that
you sell? Is it a single place that helps you describe your products and their features?
Is it a repository that helps you identify a product with its global identifiers and
helps you effectively communicate with your suppliers and customers?
What are the benefits of the product data hub? Does it help in minimizing data
interfacing needs across your application stack? Does it reduce the number of
errors due to wrongly identifying a product? Does it provide enriched business
definition and a single source of truth for your products?
These benefits are just the tip of the iceberg. This paper deals with business benefits
and how to navigate the challenges faced while achieving the benefits.
Infosys – ViewSep
Point |1
2011
Business Benefits
The benefits of a Product Hub are typically seen from a technology perspective. The common perspective is that over the past
few decades, organizations have invested heavily in IT which has helped them handle their day to day transactions efficiently.
But this has resulted in master data spread across disparate systems which is in a constant state of flux. This is a major challenge
and organizations today are countering it by consolidating master data in a single repository, since, ignoring data management
may impact the ability to deliver the key enterprise functions and processes.
A product hub is viewed as a similar implementation to provide a consolidated and unique product data, complete and
accurate item information across the enterprise. These benefits are only the tip of the iceberg. The retail environment faces
many pressure points today. Developing multi-channel initiatives, new product or private label development and product
range or category expansion are seen as critical action items in the near future. In this rapid changing retail environment a
central product hub could serve as a back-bone of the organization to improve speed to market thereby increasing the pace
of product growth, accelerate category expansion and provide a vehicle for enabling multi-channel commerce, cross-border
sales and internationalization
Six Critical Features Required
To get the most out of a Product Hub and realize the business benefits, Infosys proposes six critical features that are required
as part of the complete solution set.
Oracle PIM for Retail as a product hub, along with Oracle Product Data Quality Server provides a powerful tool for achieving
this. Infosys Retail Lab has experience in providing such solutions by enhancing the Oracle PIM for Retail platform as shown
in the Feature Map below.
Oracle Product Oracle PIM
Data Quality Server Flexible Data Model Internationalization
Suppliers Application 1
Common Inbound Common
Publish
Sellers
Application 2
PIM Inbound Oracle PIM
Outbound
Staging
Content PIM DB Staging
Providers
Application N
Business Rules Engine
XML
Internal Metadata External
Applications Management Systems
1. Flexible Data Model
The primary purpose of the Oracle PIM is to provide a clear and accurate definition of the product. Product definition has
various dimensions such as product type (clothing, electronics etc.), information type (structured product features, unstructured
reviews, and rich media content), and associations (product to product relationship, bundles, and categorization).
The product definition also varies depending on the country (such as age restrictions), warehouse (operation attributes such as
back-orderable) or supplier (such as supplier product identifier, lead time). Oracle PIM provides a host of entities and flexibility
to map them as per the business needs.
2 | Infosys – View Point
Oracle PIM
Hierarchy Attachments
Attribution based on ICC
Structure
Cross References
Structures and Bill of
Attributes Materials
• Oracle Seeded
Relationships
Attributes
• User Defined
Attributes (UDA) Association to Alternate
• Multi Row UDA Catalog Categories
• Variant Attributes
Supplier and Organization
Intersection
2. Business Rules Engine
Product data can be received into the Product Hub from various sources for a wide range of product categories, hence there is
a need for standardization of product data, values of product attributes etc. these can be achieved by using a powerful ‘Business
Rules Engine’ provided by Oracle Product Data Quality Server (OPDQ). The metadata of the product defined in Oracle PIM
can be synchronized with the OPDQ where the rules can be built for this purpose.
Data Sources and Oracle Product Data Quality Oracle PIM
Upstream applications Server
Business Rules
Engine Metadata
Management
Suppliers Data Service Application Metadata
for Workflow/Dataflow
Management
Sellers
Content
Providers Product Data
Data Lens for Oracle
Data Cleansing
Cleansed
PIM DB
Product Data
Internal
Applications
Product information is required in some downstream systems to perform their respective business operations for e.g. Order
Management System, Warehouse Management systems etc. Most of this information is common or standardized based on
product categories, hence can be managed by defaulting. This can be either managed by the business rules engine in Oracle
PIM or OPDQ server respectively. Basic defaulting can be addressed in Oracle PIM whereas OPDQ Server can handle complex
rules based on product category and product attribution etc.
Infosys – View Point | 3
3. Common Induction & De-duplication
The need for category and range expansion can be met by retailers only though sourcing product information from various parties.
In-house Applications: It is normal for retailers to create product information in-house for private labels and create rich content
such as images, web descriptions in-house for distant selling channels. Such information to be mastered will flow from In-
house applications.
Supplier Feeds: Supplier feeds are product details received from the suppliers of the product that typically contain the product
features, trade market specific data and the associated product prices received generally in Excel using a retailer specific
predefined template.
Content Providers: These are organizations that specialize in generating and selling specific categories of product data to
retailers. Data is collated by content providers from the market and other reliable sources and hence is a trusted source for
enriched & accurate product data.
GDSN Certified Data Pools: GDSN certified data pool provides the access to huge volume of product data available across
the globe. The content provided is generally supply chain related and not enriched content required for selling the product.
This enables the business to derive value in the supply chain through consistent, accurate information that helps in product
consolidation and uniquely identifying a product. It also helps in increasing the business opportunity by increasing range.
To create a central repository & single source of truth of product in Oracle PIM from disparate systems, the product feeds
need to go through data standardization, validation and de-duplication process before induction or update in Oracle PIM.
Additionally, increase in SKU count will mandate an automated process with lesser manual touch-points
A common product induction process from multiple sources through a principle of ‘single touch of data’ is projected to
enable increase in SKU count by 100 times in 2 years, for a major retailer based in UK
Duplicate product data is a leading cause of bad user experience, inefficiencies in processes and reporting. A Common Induction
process greatly aids efficient de-duplication and also provides a scalable infrastructure for inducting additional suppliers,
content providers and other sources of data.
Additionally, product data from different source will have infinite variability in terms of format, content, syntax, and translation.
Standardization is necessary before mastering in Oracle PIM and publishing to other downstream systems. Cleansing the product
data variations specific to each category is difficult to achieve using traditional tools, custom code or through manual activity.
OPDQ with its semantic approach makes it a right choice as a tool for achieving single product record with required level of
data quality. In the common induction mechanism, product data from different source are passed through OPDQ to identify
duplicates, cleanse, standardize and execute business rules. OPDQ – Oracle PIM Out-Of-Box integration also helps in periodic
cleansing of data that resides in Oracle PIM based on modified business rules. Positioning OPDQ in the architecture should be
decided based on the business & organizational needs as different architecture helps business in different ways.
Different options are available to business for product approvals while processing the products. It is advisable to decide on the
option based on the business & organizational requirements.
• Product data approval in OPDQ as in workflow without manual intervention.
• New Item request feature in Oracle PIM
• Change Management features of Oracle PIM
Product data received from different sources may involve different set of mapping and validations. Source specific data
mapping can be handled through transformations in Oracle BPEL and through predefined templates based on business needs.
Validations can be handled through a set of common functions such as mandatory conditions, list of values. Each source batch
is configured to invoke specific functions to ensure data completeness and correctness. This provides a scalable framework
where new sources can be enabled by configuring a new batch and validations required for the same.
4 | Infosys – View Point
Error Handling Framework
Source intimated
for Data Related
Errors
Sources for Product Common Inbound
Data
GDSN
Data Pool
Administrator
Oracle Oracle
B 2B BPEL Cleansing
Standardization Notification
De -duplication
PIM Staging Inbound
Content Providers
Database Valiations
Error Logging
OPDQ
Supplier
Open Interface Tables Confirmation
Response to GDSN
Legacy Systems
PIM
Database
Through an error handling framework product administrators will be able to view the errors encountered during induction,
classified by error types. Corrective actions can be taken as per business needs and reprocessed.
4. Common Publish Framework
A product hub needs to provide a common definition of the product across the organization and publish product information
for multiple downstream applications such as merchandising, order management, warehousing applications etc. The Infosys
Oracle PIM solution achieves this by defining a product canonical model.
• Product Operational Attributes
Item User Defined
• Product Features, Benefits and Advantages by Item
Attributes Catalogue Categories
Avoid a Big
• Related items Flat
Item Relationships • Substitutes
• Spares and Replacements
Canonical
Model
• Universal Identifiers (EAN,UPC Codes)
Item Cross References • Internal Identifiers required for internal systems Easier to
(Catalogue Numbers)
Understand
Intersection • Supplier Specific Attributes
Easier to
troubleshoot
• User Guides
Item Attachments • Recipes
Easier to
• Buying Guide
maintain and
• Promotional Packs
make changes
BOMs and Structures • Bundle Products
Alternate Catalogue • Merchandising Hierarchies
• Commercial Hierarchies
Categories
• Catalogue Hierarchies
Infosys – View Point | 5
The canonical model consists of a definition of the product as an XML, mapping from each source system to the model, and
mapping from the model to each destination system. The model contains the elements as shown in the pictorial above
Different product outbound batches are defined in the Common Publish Framework, along with trigger criteria and Oracle
PIM entities/ attributes that are required for each batch. Each destination system subscribes to a specific outbound batch and
any destination system specific mapping or transformation is maintained as configurations in the Integration Layer.
This common publish framework eliminates the need for point to point interfacing and drives other powerful benefits
• Eliminates the need for understanding the internal data formats of every interfacing system
• Schema/Data model changes in a system can be isolated from having impact on others
• New batches, sources and destination systems can be added without any changes in Oracle PIM, creating a scalable model
which helps multi-tenancy and internationalization
• Adding categories and attributes can be handled through simple configurations rather than development, easing the
process of category and range expansion
5. Channel Specific Metadata
Multi-channel commerce requires different types of product data available to surface the product. In distant selling channels
like web, a granular definition of the product and additional metadata of attributes such as search-able, browse-able, key
words, attribute descriptions, sequencing etc. are required to provide world class user experience such as side by side product
comparisons and good search-ability. Other rich content such as professional reviews, awards, digital assets (such as images,
videos, 360 etc.) are required to provide a touch and feel to the product.
Oracle PIM provides functionality to store various assets of a product. Channel specific metadata has been provided in the
Infosys solution by leveraging the common publish framework.
6. Regulated Metadata Management
Product Structure Metadata consists of product data attributes, their data types, acceptable set of values, acceptable set of units
of measures, date formats etc. This metadata is in a constant state of flux, especially in an environment where the SKU count
and number of categories increases dramatically. The metadata also influences the mechanism in which various sources induct
product information into Oracle PIM and various destinations interpret the published batches.
The common publish framework along with a regulated metadata management process have enabled a major retailer to
double their category range
A robust mechanism for regulating these metadata changes in Oracle PIM and synchronizing with interfacing applications/
parties is essential for an organized increase in categories and SKUs. Lack of this will at best mandate frequent development
in PIM & interfacing applications, and at worst result in Oracle PIM not being able to handle the pace of change that the
business may require.
In the Infosys solution, various automation scripts are available for speeding the process of metadata changes in Oracle PIM.
An extension has been created for staging the metadata in Oracle PIM such that changes are released to Oracle PIM users and
interfacing applications only through a staged release process.
6 | Infosys – View Point
Oracle PIM, along with OPDQ and the Infosys solution provides a powerful
platform for the retailer to derive real business benefits and could prove to be
Conclusion the vehicle for driving revenue expansion through category, SKU and channel
expansion.
About the Authors
Mani Munikrishnan is a Principal Consultant with the Oracle Retail practice of Infosys.
He specializes in delivering client value through package enabled re-engineering in the
Financials, Distribution and MDM domains. He has led many end-to-end Oracle e-Business
suite implementations as a solution and program manager. He holds a master’s degree in
Industrial Engineering from State University of New York at Binghamton
Nandakumar Balasubramanian is a Senior Consultant and has around 8.5 years of work
experience with 4 years in Product information management. He has extensively worked
for clients in Retail and Publishing industry domain. He provides process, domain and
technology consulting and has extensive experience in business process analysis, solution
design, and implementation in the area on product information management. He holds
Bachelor’s degree in Engineering with specialization in Mechanical engineering.
Vivek Kale is a Lead Consultant and has over 10 years of work experience with 4 years
in Product information management. He has extensively worked for clients in Retail and
Hi-Tech industry domain. He provides process, domain and technology consulting and has
extensive experience in business process analysis, solution design, and implementation in
the area on product information management. He holds Bachelor’s degree with specialization
in Computer Science & Engineering.
Infosys – View Point | 7
About Infosys
Many of the world's most successful organizations rely on
Infosys to deliver measurable business value. Infosys
provides business consulting, technology, engineering and
outsourcing services to help clients in over 30 countries
build tomorrow's enterprise.
For more information about Infosys (NASDAQ:INFY),
For more information, contact
[email protected] visit www.infosys.com.
© 2011 Infosys Limited, Bangalore, India. Infosys believes the information in this publication is accurate as of its publication date; suchinformation is subject to change without notice. Infosys acknowledges the proprietary rights of the
trademarks and product names of other companies mentioned in this document.
8 | Infosys – View Point
