While master data management (MDM) may be a newer term to some IT departments, there is no disputing the value of better managing shared master data between departments, business functions and applications. While new MDM tools promise to create a consistent set of master data that can be leveraged across the enterprise for transactions, analysis and business intelligence, the potentially vast scope of executing against an all-encompassing vision of enterprise-wide master data may be daunting indeed. Here's an example of an evolutionary approach designed to show clear, incremental value.
The need for a unified data "hub" such as the newly announced SAP MDM as a means to more precisely use interrelated master data for better business intelligence cannot be disputed. Even for companies that have embraced a full suite of business applications, many IT initiatives such as service-oriented architecture (SOA), data warehousing, complex demand planning, enterprise resource planning (ERP) consolidation and industry standardization all point to the need for managing master data domains.
Every major technology analyst firm is working overtime to help articulate the value, cost and players in the MDM space. And while some are warning of so-called MDM hype, others such as Ventana Research show that MDM is indeed gaining momentum, with nearly half of the 230 participants surveyed in a May 2007 study, "MDM: Business and Technology Trends," indicating that an MDM project is either planned or under investigation in their organizations, and with 27 percent responding that a master data strategy is already underway.
At its most fundamental level, MDM enables companies to create a "reference book" for unifying master data across all enterprise applications and analysis. This master data can comprise almost anything — customer, product, asset, employee or financial data — and can be in the form of bar codes, part number schemes, vendors by SIC code, customer names or job roles. The key is in delivering a framework for consistent terminology and nomenclature so that apples-to-apples evaluations can be made in creating one true view of the data.
Or course, the idea itself isn't new, and it has existed as a core concept for solutions such as product information management (PIM) and customer data integration (CDI). The challenge is one that will sound very familiar to anyone who has ever survived an ERP implementation: creating and executing on an enterprise-wide vision is a road fraught with peril and one that may be hard to justify in the final analysis. Today, while the market may be marching forward and visionary IT teams may be eager to sign on to a solution that can make their lives easier in the long run, management is rightfully hesitant about embarking on a large-scale odyssey.
A strong MDM strategy touches so many parts of the enterprise that it may take years to define, evangelize and implement, but that doesn't mean that MDM needs to sit quietly on the sidelines until that time. Rather, an evolutionary approach to moving forward with MDM could in fact unlock the door to broader acceptance of an enterprise-wide MDM strategy.
An evolutionary approach means selecting a single project within a contained department or business function, with a readily identifiable business champion and well-defined success metrics. For example, trying to ramp up 50 or 80 manufacturing facilities in preparation for a global product launch might simply be too large and unwieldy a project for an MDM pilot. While visibility may be high and the resulting success could be a triumph for IT, there may be too many teams involved, and project slippages will be extremely visible, making the stakes rather high for an initial MDM project. On the other hand, applying a pilot MDM project to a program like spend analysis represents a more contained and risk-averse endeavor, with one clear champion — the chief procurement officer — and clearly quantifiable business value.
First, let's define spend analysis and why it's important.
Spend Analysis: Taking the Pain out of Procurement
A report by industry research firm Aberdeen noted that CFOs believe corporate procurement suffers from a lack of clear visibility into spend across various commodities, factors that hurt corporate competitiveness by preventing procurement from driving strategic initiatives such as supply base rationalization or commodity normalization. Spend analysis can be the remedy. Put simply: Spend analysis entails taking the existing information that companies track in the form of invoices, purchase orders and receipts and "cleansing, normalizing, categorizing and enriching the data," as marketing research firm Forrester Research defines the process. After doing so, companies will be better able to find sourcing opportunities that reduce cost, track non-compliance with purchasing policies and improve data management. With this approach, businesses could save 2-6 percent of their total spend by improving the way their procurement organizations secure products and services, according to both the Aberdeen Group and AMR Research.
The reasons why businesses fail to collect accurate, useful procurement information in the first place is not a lack of tools and applications to manage spend. The heart of the problem is data complexity, and that is exactly what an MDM strategy sets out to solve. For example:
- Different codes are often inadvertently used to describe the same supplier or item across divisions or even within the same division, making the broad set of data seem inaccurate when it is analyzed. For example, one plant might code HP as HP, while another might record it as Hewlett-Packard. Without a standard way to name a company, aggregate totals can be off, weakening a company's leverage with its suppliers.
- Item codes are used to define products, but these codes don't always connect an item to an industry-standard classification. This makes it difficult to aggregate similar types of data and combine spending across commodities, locations, suppliers and programs. If an item master record contains industry-standard classification, aggregation of spend at various levels of category will become possible.
- Relationships between suppliers aren't defined or are hard to decipher in a business application such as ERP. Understanding these relationships — such as the fact that Lab Safety Inc. is a subsidiary of W.W. Grainger — can help provide leverage when negotiating bulk deals or rationalizing per-vendor spend.
- Many crucial bits of information, such as the minority status of a supplier, might not exist within an ERP system. This information can help a company take advantage of tax breaks or be used for regulatory compliance. Without it, supplier rationalization initiatives can fall short of company-wide goals.
These pain points are leading the smartest procurement executives to initiate spend analysis initiatives, the key to which will be identifying, normalizing and connecting data. While the data may come from multiple catalogs or business applications, this is a very contained, finite set of data that can be readily cleansed by automated tools and subject matter experts. Once the initial cleansing and framework is in place, ongoing data management becomes simply a maintenance issue.
Implementing a Successful Spend Analysis Campaign
Whether an organization accomplishes this internally or works with a service provider, three key phases are required to implement a successful spend analysis campaign.
Phase I — Extract and cleanse vendor and spend data
- Combine spend data that may be stored in multiple databases. These data often exist as individual transactions in ERP systems, e-procurement applications such as Ariba, or American Express travel and expense reports.
- Map each product to popular standards-based classification taxonomies using the Universal Standard Products and Services Classification (UNSPSC) and eCl@ss. Mapping allows users to aggregate spend information by commodity type — such as five of the same motors coming from two different suppliers — despite the fact that these motors are labeled with different item codes in the ERP system. Classification data can be added to the item data within an MDM hub so the information is accessible for current and future initiatives that require this information.
- Gather all useful information about a company's suppliers, including parent company name, revenues, credit rating, standard industry code and diversity status and incorporate it into the system. These data give buyers the weapons they need to negotiate better deals across their subsidiaries. All the updates are applied to the data within the MDM hub.
Phase II — Analyze spending and identify potential savings
- Use automated spend analysis charts — either custom-developed or from a business intelligence application — to pinpoint the areas of greatest potential cost savings and estimate the cultural changes required to achieve targeted results.
- Use analysis techniques to determine savings potential, the amount of disruption expected and how receptive the organization will be to the change. So-called "quick wins," changes that will prove easiest, most beneficial or quickest to make, can be prioritized and implemented.
- If spend analysis shows that one department pays significantly more for a motor than another, for example, the company can move to consolidate motor purchases. Also consider implementing the processes needed to ensure all departments comply with new contract rules regarding purchasing the less expensive motor. This elimination of duplicate, equivalent or similar items allows companies to achieve significant cost savings.
Phase III — Expand into supplier rationalization and other spend optimization initiatives
- Once a common data framework is in place, companies can use the data gathered during the analysis phase to refine their sourcing strategies. Key performance indicators can also be established and used as a way to analyze suppliers.
- Data also can be used to identify issues that cause costs or the amount of time it takes to complete a transaction to spike when working with suppliers during the design, manufacturing and procurement process.
Since the updated master data are managed within an MDM framework, this can ensure that any changes to a specific master data can be propagated to other occurrences of that data. As a result, Phase I does not have to be repeated whenever spend analysis is initiated again, which for most organizations is once a quarter or even once every six months.
Through adopting the three phases of spend analysis, a large manufacturing company we worked with recently was able to reduce its overall purchasing spend by about 5 percent. Throughout the process, the company analyzed procurement rules within 16 systems, including SAP. The company extracted and consolidated a total of 22 million records and standardized more than three million item/service records. It also created a five-level classification scheme for items and services. After finishing the spend analysis initiative, the company was able to reduce its rate of purchase of similar commodities from multiple suppliers at varying contractual prices. The CPO also eliminated maverick buying. After clearing up price variance issues — a prevalent problem throughout the company — the organization reported significant savings, for example a $42,000 annual savings on the purchase of just one item.
Moreover, the value of spend analysis goes far beyond a one-off project. A long-term program will help a company continue its success by preventing maverick buying, encouraging contract compliance and reducing the practice of buying from multiple suppliers at different prices. This kind of a project — with clear and achievable goals, one business champion, data contained within a single business function — can then be used to present a solid use-case for how MDM can deliver tangible business value in other areas of the enterprise. Instead of evangelizing an MDM vision, an evolutionary approach such as this can put IT in the position of prioritizing which line-of-business executive's project could be next in line for success.