As of 2015, Reference Data Management has clearly become a master data domain commensurate to CUSTOMER and PRODUCT master for many organisations.
EDITOR’S NOTE: For the past 2+ years, our annual MDM Summit has hosted numerous organisations speaking about their reference data management use cases and best practices including Cargill, DnB NOR, O2, Raiffeisen Bank, Royal Mail, UBS, Visa International and others. Our MDM Summit is the *only* such conference to provide such a commensurate focus on CUSTOMER, PRODUCT and REFERENCE data.
The impact of poor or non-existent reference data management (RDM) is profound. Errors in reference data ripple outwards affecting quality of master data in each domain, which in turn affects quality in all dependent transactional and analytical systems. Because reference data is used to drive key business processes and application logic, errors in reference data can have a major negative and multiplicative business impact. Mismatches in reference data (also called “enterprise dimensions”) affect the integrity of
business intelligence reports and are also a common source of application integration failure. Because poor reference data can have a negative impact on transactional and analytical systems, enterprise IT groups have taken the lead in RDM projects. Additionally, since enterprise dimensions are often used for measuring financial performance, Finance departments sometimes “scale up” their implementations to support the entire enterprise. Due to the strategic nature of and difficulty to build/maintain custom RDM capabilities, savvy IT organizations and Finance departments are increasingly opting to “Buy, not build” RDM solutions.
In addition, RDM capabilities are increasingly called out as a standalone or adjunct/support capability to mainstream Master Data Management (MDM) solutions such as customer data integration (CDI) and product information management (PIM). Clearly, RDM is often a good entry-level project to show success for initial MDM investment which can be built on as a Data Governance model. However, it is important to note that the functionality required for an RDM solution is not aligned with every single MDM use case. While RDM aligns quite nicely with a majority of MDM use cases, the one notable exception is CDI projects (generally volumes of multi-tens of millions) that have a strong focus on profiling and quality.
As with enterprise CRM, ERP and MDM initiatives, organizations need to plan for a 2:1 services:software spend ratio. To make matters even more challenging, the mega vendors’
RDM pricing models tend to frustrate rather than encourage wide-spread deployment of RDM capabilities.For the global 5000 enterprise, multi-domain RDM solutions can provide quicker time-to value and higher probability of success relative to the alternatives of custom RDM frameworks built upon commercial MDM (or re-purposing of the more expensive, real-time RDM solutions).
Per estimates by analysts at the MDM Institute, the market for Reference Data Management software (multi-domain RDM, excluding real-time RDM) reached US$`80 million during 2014 and will exceed US$250 million by YE2015. Clearly, RDM is essential to the success and sustainability of the majority of MDM projects, yet prior to 2014 most MDM software vendors did not recognize this market requirement. Given the substantial investment enterprises undertake with RDM programs, they should seek out multi-domain RDM solution providers that understand and have experience addressing complexity of reference data. In short, ‘Buy, don’t build, RDM.’”
What is Reference Data?
Reference data is the Rosetta Stone-like capability for enterprise and departmental applications to understand and work with one another (see Figure 1 – “Reference Data Working Definition” for the general purpose Wikipedia definition. These look-up or code tables are typically used in a “read only” manner by the referencing applications. Unfortunately, due to the nature of IT application development and the reliance upon off-the shelf application systems, reference data is all too often isolated in silos as such reference data exists in many different systems in the typical corporate enterprise architecture. Moreover, each application typically has a different subset, format and internal representation because there are usually no agreed upon internal standards for specific reference tables (and their semantics or taxonomies).
Why Reference Data is MissionCritical.
Analysts and consultants rarely agree to the degree that they do about the common definition of “reference data”. The industry consensus definition is “reference data is the coded, semantically stable, relatively static data sets shared by multiple constituencies” (people, systems, and other master data domains)”. Simple as that sounds, the impact
of poor or non-existent reference data management (RDM) is:
Figure 1Reference Data – Working Definition
“Reference data is data that defines the set of permissible values to be used by other data fields. Reference data gains in value when it is widely re-used and widely referenced. Typically, it does not change overly much in terms of definition (apart from occasional revisions). Reference data often is defined by standards organizations (such as country codes as defined in ISO 3166-1).
- Typical examples of reference data are:
- Units of measure
- Country codes
- Corporate codes
- Conversion rates (currency, weight, temperature, etc.)
- Calendar and calendar constraints.
Reference data should be distinguished from masterdata, which represents key business entities such as customers and materials in all the necessary detail (e.g.,for customers: number, name, address, and date of account creation). In contrast, reference data usually consists only of a list of permissible values and attached textual descriptions.”
In 1999, NASA lost a $125 million Mars orbiter because a US engineering firm’s team used the English units of measurement while the NASA agency’s team used the metric system for a key spacecraft operation.
Errors in reference data will ripple outwards affecting quality of master data in each domain, which in turn affects quality in all dependent transactional systems. Because reference data is used to drive key business processes and application logic, errors in reference data can have a major negative and multiplicative business impact. Mismatches in reference data impact on data quality affect the integrity of BI reports and also are a common source of application integration failure (see Figure 2 “The ‘Tsunami Effect’ of *Bad* Reference Data). Additionally, management of such reference data is needed in both operational and analytical MDM use cases where such capability often used to provide attributes, hierarchies and KPIs.
Buy vs. Build
Commercial RDM solutions are a relatively new offspring of Master Data Management (MDM) functionality. RDM provides the processes and technologies for recognizing, harmonizing and sharing coded, relatively static data sets for “reference” by multiple constituencies (people, systems, and other master data domains). Many contemporary MDM vendors such as IBM, Informatica, and SAP have augmented their MDM hub functionality to manage reference data as a special type of master data. Such a system provides governance, process, security, and audit control around the mastering of reference data. In addition, RDM systems also manage complex mappings between different reference data representations and different data domains across the enterprise. Clearly, governance of reference data is vital. Moreover, manual or custom RDM often lacks change management, audit controls and granular security/permissions. Most contemporary RDM systems also provide a service-oriented architecture (SOA) service layer for the sharing of such reference data.
Figure 2 – The ‘Tsunami Effect’ of *Bad* Reference Data
Prior to the availability of commercial RDM solutions, organizations built custom solutions using existing software such as RDBMS, spreadsheets, workflow software (business
process management or BPM) and other tools. Such systems often lacked change management, audit controls, and granular security/permissions. As a result, these legacy solutions have increasingly become compliance risks.
Because reference data is used to drive key business processes and application logic, errors in reference data can have a major negative and multiplicative business impact. Mismatches in reference data impacts data quality and affects the integrity of BI reports and also are a common source of application integration failure (a.k.a. “systemic failure”)
Within the realm of commercial RDM solutions, there are two main families: “multi-domain RDM,” and “real-time RDM”. “Multi-domain RDM” solutions are non-industry specific solutions that can span functional areas (finance, risk and compliance, human resources) and content types (ISO country codes, and other non-volatile reference data to be mastered and shared). “Real-time RDM” is typically a very high performance solution for use in the capital markets industry (brokers, asset managers, and securities services firms) as well as command and control military/intelligence markets.
Increasingly, many large enterprises have begun to make RDM their initial test case or proof-of-concept for their MDM evaluations. As a consequence, MDM vendors are rushing to market RDM solutions to apply an MDM approach for centralized governance, stewardship and control. Cognizant, iGATE Patni, Kingland Systems, Wipro Technologies, and other systems integrators will move into the “securities master” market (some by OEMing of Informatica and IBM MDM). During 2015-16, pervasive, low cost RDM will be commoditized via the efforts of Microsoft and Oracle as these vendors provide low or no-cost RDM solutions as part of their software families. Moreover, as many large enterprises have begun to make RDM their initial test case or proof-of-concept for their MDM evaluations, the vendor community is responding by providing easier-to-manage entry points into RDM use cases using either existing MDM platforms or purpose-built RDM solutions which use MDM as their foundation. Clearly, managing “simple” reference data will prove to be a key sales entry point for large enterprises and their MDM vendors.
Additionally, RDM can be expected to become a “ramp up” point of entry for many organizations planning for CUSTOMER, PRODUCT master and other domains, as well
as an entry point into master data governance. During 2015-16, we believe a great amount of current and next-generation commerce will be facilitated by on-premises and cloud-based RDM solutions with support for both “private” and “public” reference data. “Public” reference data is what many people typically think of when they consider reference data. Public reference data is based on standards where overall consistency is a primary goal. Examples of public reference data include industry standards (GS1 GPC), national standards (European Banking Authority’s Legal Entity Identifier (LEI)), International Standards (ISO, ISIC), and data from vendors (Bloomberg, D&B, S&P). “Private” reference data is used to maintain consistency when doing business with external parties. Examples of private reference data include financial and organizational hierarchies (e.g., DUNS hierarchies), and employee organizational structures. Mapping logical connections between the different master data domains and reference data illustrates that both kinds of reference data (public and private) have a large number of connections to every MDM domain. This means that an error in reference data will ripple outwards affecting the quality of the master data in each domain, which in turn affects the quality in all dependent transactional systems. Clearly, the heavily interconnected nature of reference data is why it requires separate management and governance. BOTTOM LINE: Buy, not build RDM. Just as enterprises no longer build their own
CRM, ERP and MDM systems, organizations should acquire commercial RDM solutions which can be easily tailored or configured and have the full ongoing support of a proven RDM software vendor. For the majority of enterprises, multidomain RDM solutions can provide quicker time-to-value and higher probability of success relative to the alternatives of custom RDM frameworks built upon commercial MDM (or re-purposing of the more expensive, real-time RDM solutions).
Aaron Zornes is the founder and chief research officer of the MDM Institute (formerly the CDI-MDM Institute). Mr. Zornes is a noted speaker and author on Global 2000 enterprise IT issues and is the most quoted industry analyst on the topics of master data management (MDM), customer data integration (CDI), and data governance. Mr. Zornes is also the editor and lead contributor to DM Review’s CDI and MDM Newsletters as well as the monthly columnist for both CDI and MDM. Prior to founding the MDM Institute, he was founder and executive VP of META Group’s largest research advisory practice for 15 years. Mr. Zornes also held line and strategic management roles in leading vendor and user organizations, including executive and managerial positions at Ingres Corp., Wang, Software AG of North America, and Cincom Systems. Mr. Zornes received his M.S. in Management Information Systems from the University of Arizona.
Aaron will be making the following presentations at this year’s MDM Symposium Europe 2015, London, 18-21 May:
- PRE-CONFERENCE WORKSHOP: MDM (CDI, PIM, RDM & MDG) Quick Start
- MDM KEYNOTE: “Master Relationship Management” is Coming of Age
- MDM Track 2 – MDM of Product & Reference Data: Expert Testimony: Field Reports for’Top 10′ RDM Solutions
- MDM KEYNOTE: MDM Analyst Keynote: Field Reports for ‘Top 15’ MDM Solutions
- MDM Track 3 – Master Data Governance, Futures: Expert Testimony: Field Reports for ‘Top 10’ MDG Solutions
All articles are © 2015 by the authors.
The views and opinions expressed by our authors are those of our speaker and do not necessarily reflect the official policy or position of IRM UK.
This article was featured in IRM UK’s Monthly E-newsletter. To subscribe please visit http://www.irmuk.co.uk/usefulinfo/enewsletter.cfm Please note we are always on the look-out for new contributors so if you have an article you would like published please forward it to firstname.lastname@example.org for consideration.
IRM UK Strategic IT Training Ltd.
2nd Floor, Monument House, 215 Marsh Road, Pinner, Middlesex, HA5 5NE
T: +44 (0)20 8866 8366 E: email@example.com W: www.irmuk.co.uk
Nice blog about Data Management. Thanks for share with us. If you need Data Management Strategy services in the USA then EWSolutions is the best place for you.