Prepared by BPM Partners

P r o f i s e e • w w w . p r o f i s e e . c o m • i n f o @ p r o f i s e e . c o m

The Case forMaster Data Management

WH

ITE

PA

PE

R

Table of Contents

Introduction ....................................................................................

History: Where the MDM Problem Came From ..........................

Master Data—A Problem for the Enterprise, Not Just IT ...............

The Strategic Costs of Keeping the Status Quo ...........................

The Problem for Finance .............................................................

The Problem for IT .......................................................................

Is Master Data Management an Issue for Your Company? .........

MDM: Vision and Practical Solutions .............................................

Sarbanes-Oxley Compliance ........................................................

A Collaborative Process for Master Data Changes ......................

Integration versus Open System Interface ..................................

MDM Supports Enterprise Growth ..............................................

Centralization ...............................................................................

Conclusion ......................................................................................

3

3

4

4

4

4

5

6

6

7

7

7

8

9

The Case for Master Data Management

Introduction

According to the research and advisory firm TowerGroup, fifty percent of enterprises maintain masterdata separately in eleven or more source systems.Eighty percent of enterprises plan on centralizingmaster data. Because of this, large organizations facea growing problem that already undermines the accu-racy of the reporting and analysis of their operations.Master data is being categorized, rolled up, and fil-tered inconsistently and erroneously across the enter-prise.

Inconsistencies arise between business units thatview the enterprise in different ways. Discrepanciesare also found between different databases, datawarehouses, and BI (business intelligence) or BPM(business performance management) applicationsand tools. According to Forrester Research, mostGlobal 2,000 companies use between five and fifteenreporting and analysis applications; making a single BIplatform “an IT fantasy”.

The results of poor master data management can beembarrassing, as when managers enter a meeting topresent reports that show conflicting totals. The prob-lem can also be costly—when reports are wrong,incorrect decisions are taken, and the company fallsout of compliance with Sarbanes-Oxley requirements.Add to this the toll in employee time to find and cor-rect mistakes. It's very common for sophisticatedmultinational companies to have ongoing, expensivedifficulties in reporting that stem directly from con-fused master data.

The traditional approach—telling IT to fix the prob-lem—does not work because accurate enterprisereporting is a complicated business issue thatrequires ongoing, detailed collaboration with the ITside. IT and Finance—and areas that often use indi-vidualized reports, such as marketing and sales—arenot organized or designed to tackle this problem in ajoint manner.

A logical suggestion is that the company's main trans-action or ERP system is the place to store and protectcommonly used master data pertaining to customers,products, accounts and other business dimensions,but those systems have no "home" or facility for thisinformation. They do not define and store a completeset of master data.

The answer is to proactively manage master data—the report and analysis structures used by employeesin every functional area—with a software solutioncalled an MDM system. The MDM (master data man-agement) system:

• Handles the necessary integration of transaction-al storage, and analytic systems

• Supports compliance with Sarbanes-Oxley

• Enables an auditable, repeatable process

• Captures and protects your company's humanintelligence about report structures—this isknowledge which otherwise lacks a home.

MDM enables appropriate employees to change mas-ter data, does the hard work of implementing changesacross the associated data systems, and keeps track ofthem. This means that the right business users will, toa much larger extent, control and have responsibilityfor the way business results are reported.

History: Where the MDM Problem Came FromOver the years, companies assembled their own port-folios of systems as they took a best-of-breedapproach to software applications. They also inherit-ed a mix of different systems through M & A. Today,it's rare to find a large company that uses a pure sin-gle-vendor approach. With large collections of differ-ent systems, it's well-understood that integration atthe data interchange level is key. Now it is becomingapparent that another common thread must existbetween these disparate systems: master data mustbe trans-system and cross-enterprise.

Historically, while corporations were adding the oper-ational and analytic systems they needed, master datadiscrepancies came with those systems and now are aconstant companion. The variety of systems bought,or inherited through M&A, was the Trojan horse, so tospeak, and master data was waiting inside.

MDM can also be described as a way for some aspectsof managing crucial, shared corporate data to becomelargely a business function, as opposed to being pure-ly an IT function.

This white paper focuses on why organizations needMDM, and how MDM solves this data managementproblem.

3

The Case for Master Data Management

Master Data̶A Problem for theEnterprise, Not Just IT

Large organizations waste considerable time andmoney figuring out who has the right view of theirinternal world. They expend more time and moneyattempting to apply management's preferred viewacross the many different systems used in a company,and repairing mistakes that sneak into reports.

The Strategic Costs of Keeping the Status QuoThe cost of not undertaking master data managementcan be significant:

• Mistaken understanding of which products areprofitable and growing

• Mistaken belief that reports are accurate, leadingto incorrect decisions such as closure of thewrong retail branches or funding of the wrongproducts

• Misalignment of management reports versusstatutory reports

• Missed opportunities, through a systemic inabili-ty to track a particular attribute or activity at agranular level or, conversely, at the enterprise-wide level

• Vulnerability to security and data integrity prob-lems that arise from the lack of a defined processto control important aspects of the analytic toolsused in the enterprise.

The benefits of managing master data proactively(covered in the next section) include accurate report-ing, easier integration of mergers and acquisitions,and better compliance with Sarbanes-Oxley and inter-national accounting standards. MDM brings the flexi-bility to proactively drive more report options, toexamine business activity by slicing it in differentways, and to more easily compare year-over-yearresults even if the master data has been altered.Finally, it keeps track of when master data is changedand by whom.

The problem that MDM solves has its own 'iceberg'analogy. You might see only the tip of that iceberg:reports that are delivered late or with errors, or aninability to mesh financial statements after an acqui-sition or restructuring. Less visible is the mixture oftechnical and business issues that underlie enterprisereporting.

The Problem for FinanceThe pain of reporting is not always evident to seniormanagement. By virtue of their position, they areinsulated from the difficulties. In general, the problemof master data is not well understood. That makes itembarrassing for mid-level managers to mention it totheir superiors. The difficulty behind the scenes goesbeyond discomfort; it breeds data and report errors.

As an executive, you have requested a new report,based on a different rollup of business operations.You notice anxiety and consternation among yourpeople. You may have just initiated the following typeof scenario:

• The finance manager who received your requestknows enough about the systems to realize thatmost of the information you need lives in a datawarehouse. She'll have to talk to the data ware-house technical administrator, however, to con-firm if and how it can be extracted in the formyou want. Quite likely she'll ask an analyst orOLAP system administrator to build a new ‘cube'of information sourced from the warehouse.There's a time element, and the anxiety of meet-ing your deadline. Finally, she is justifiably con-cerned because many errors are born at theconfusing juncture of systems, particularly whendata is regrouped.

• The same manager has to entrust the challengeof normalizing data from the different source sys-tems to the technical administrators. Simply put,she is now responsible to you for the accuracy ofa task she does not control: merging data fromdifferent systems. New enterprise reports typical-ly require a mix of technical and finance issuesthat are not simple to sort out. With a tight dead-line, the risk of faulty reports will rise sharply.

A marketing manager, receiving the same instructionto make a new report, probably has less experiencedealing with IT and sorting out accuracy problems.

The Problem for ITTechnical administrators of databases and data ware-houses, the people with the hands-on ability tochange report rollups, are also in a difficult position.They receive conflicting instructions from analysts,product managers and sales management on how tocreate and change reports. This frequently forcestechnical administrators to make business decisions.

4

Example: How It Gets Complicated

Let's take a hypothetical multinational consumer goods company to illustrate the complications that can flowfrom a lack of master data management.

Its US division includes all coffee products under Breakfast Foods. That always worked, so the practice has becometradition. But the company's division in France rolls up fresh coffee under Beverages. The French division treatspowdered decaffeinated coffee quite differently; lumping it under Miscellaneous Snacks, a different perspectivethat a decaffeinated product receives in that country. Just with this simple example, a global report on the mar-ket share of Breakfast Foods, or Total Coffee Sales, is difficult and requires significant "Excel time", pulling detailfrom a variety of databases. It can easily become worse. Next, in the US division, the product manager of OrganicFoods goes to management and makes a point: organic and shade-grown coffee—which happens to be explodingin popularity while overall coffee sales are flat—should be under her purview. She asks the technical administra-tor for reports that show the growing clout of Organic Foods in the US and Europe. How difficult will it be to rollup historical and ongoing organic coffee sales under her category? At the technical and accounting levels, thisbecomes much more complex, because different ERP systems and different charts of accounts come into play.

For example:

• Should they revamp the data structures they lookafter, to meet requests from Finance orMarketing?

• Should they fend off those requests, or simplyinsist that the request come from a vice-presi-dent?

• Should they accommodate only requests thatconform to a particular chart of accounts?

• Is it the technical administrator's responsibility toensure that a rollup change in one OLAP cube isreplicated throughout other databases? Howdoes that happen—and how should it happen?

In few companies do technical administrators haveguidelines on when to second-guess business users onthe issue of report structures. Yet, usually it is notpossible or sensible to accede to all requests.

With a variety of databases and analytic tools in use at many enterprises, the software vendors who soldthese tools are not putting energy into getting appli-cations to use common master data. While certainOLAP tools come with "dimension editors," they con-form only to that tool, and thus do not help to enforceintegrity or consistency. Some "included editors" donot allow the administrator to revert back to an earli-er structure after changes are made. All the pain ofinterfacing among transactional, reporting, and ana-lytic systems falls to IT personnel, often with Financefinding and trying to fix errors that inevitably result.

MDM would give the Finance manager the ability todeliver the report, independent of the IT personnel.She could create a new rollup of data, according tothe request, and it would be remembered andtracked. Once created, the new rollup can be reused.It can also be discarded completely and cleanly, if thecompany finds it less than useful. MDM makes it pos-sible for the Finance professional to make the neces-sary changes without being entangled in technicalissues.

Is Master Data Management an Issue for YourCompany?One way to size up the master data managementissue at your organization is to gauge it by factors ofcomplexity and rates of change. The complexity fac-tors include:

• The number of separate transactional, opera-tional, and analytic (BI, OLAP, BPM, etc.) systemsyour company uses

• The number of business units with their ownreport structures

• How many charts of accounts are in play

• How many large-scale dimensions are used.

The change factors include the rate of organizationalchange, the rate of master data changes, and thenumber of report requests. If all these factors go up ata linear rate, it tends to complicate the master datamanagement problem at a geometric rate.

The Case for Master Data Management 5

In most companies, analytic independence isinevitable. If reports don't categorize and aggregateinformation the way users want it, they tend to goafter the detail data and work it on their own, build-ing idiosyncratic reports. These independentapproaches have quiet ripple effects that quickly spi-ral out of control. With an MDM solution, analysts canhelp improve business performance and the chaos ofmultiple report hierarchies is controlled.

MDM: Vision and PracticalSolutions

Reporting, business analysis, and performance man-agement will be more accurate and work better if thesame master data is employed enterprise-wide.Business managers hold knowledge of which rollupsaccurately portray business performance. A depend-able process is needed to ensure that the right man-agers have appropriate influence over particularreport hierarchies and master data.

This can be accomplished today with an MDM systemthat resides between the databases and the businessusers. The databases and data-handling applicationssubscribe to the MDM system. The appropriate usersare assigned privileges to manage particular masterdata, or to manage a particular level in the hierarchyof rollups. When an authorized user makes a change,the change will be committed to each subscribingapplication when the technical administrator of thatapplication gives permission. The system keeps a per-

manent audit trail of who changed what data, andwhen. Master data management becomes a repeat-able process that is largely independent of IT.

Sarbanes-Oxley ComplianceThe Sarbanes-Oxley legislation drove many compa-nies to evaluate MDM for the first time, but wheredoes MDM rank on the list of Sarbanes-Oxley priori-ties? That differs from one company to the next, butin nearly every organization a repeatable, self-docu-menting process to manage the structures and consis-tency of reporting is essential to compliance.

In particular, because an MDM system records eachchange to master data and report structures, it sup-ports Sarbanes Oxley compliance. All vendors of busi-ness intelligence and performance management tools

How Is Master Data Handled Today?

The Case for Master Data Management6

It may be illuminating to ask your analysts how muchtime and money is expended to maintain masterdata and rollups or hierarchies. Ask them how likelyit is that, after a change of rollups, the followingissues occur:

• Something ends up in the wrong region

• Data is double-counted

• Data needed for analytic purposes is omitted.

proclaim themselves to be "SOX-compliant," but if themaster data they use isn't consistent then the foun-dation of their reporting is inherently flawed and non-compliant.

A Collaborative Process for Master DataChangesMaster data management at the enterprise level isfundamentally collaborative, with multiple usersacross multiple departments working together toestablish and maintain the most useful data. Theapproval process provided by an MDM solution’sworkflow capability helps balance control of masterdata, easing the burden on IT departments. MDM alsofacilitates corrections when inconsistencies are foundin the report hierarchy.

Collaboration implies more business users acting onreport structures, and this is only possible with ade-quate security. MDM systems provide security in theform of role-based access privileges and a permanentaudit trail of changes.

Collaborative master data management calls forautomation. The MDM system should actually carryout authorized data changes on subscribing systems.In companies with 10 to 20 systems, this automationmeans enacting master data changes once, not 10 or20 times. It also removes the element of error from arepetitive process, opening MDM up to a wider audi-ence of business users.

Integration versus Open System InterfaceMost companies have upwards of ten applicationsrunning, and with just ten, there are already poten-tially 45 point-to-point interfaces to manage. MDMsystems are considerably more useful if they are ven-dor-agnostic; that is, if they handle subscribing sys-tems from multiple vendors. The integration problem,for report structures, is greatly reduced by an MDMsolution that has an open system interface. But organ-izations also need to map out, and in some cases tem-porarily adopt alternate hierarchies. A prime exampleis found in the merger and acquisition process, whereit is necessary to plan and evaluate different organiza-tional scenarios. Pro forma financial reports for differ-ent "possible new entities" are needed. To supportthis activity, MDM software systems need versioncontrol capabilities that support and maintain anunlimited number of different report scenarios.

This product feature is called "parallel branching" andgoes beyond the rudimentary form of version controlwhich tracks incremental changes to the current setof master data. An important core capability, versioncontrol allows the company to cleanly revert to a pre-vious version if the new hierarchy is not working out.

MDM Supports Enterprise GrowthAs companies grow geographically, it is quite impor-tant to ensure that all regional units stay within acommon reporting structure. The issue of coordina-tion, integration, and reconciliation becomes morepressing with each geographic step, and with eachyear of system evolution.

MDM Makes Structural Changes Easier

M & A

The Case for Master Data Management 7

Your company plans a merger. This triggers differentplans across the company. IT's reaction may well be,"We need to acquire more ETL (extract-transfer-load)tools to handle the integration."

The Finance group makes a plan to "modify its OLAPcubes" to incorporate data from the other company.Marketing sets out to create new spreadsheets fit-ting the product lines together.

The more effective response to the merger is a uni-fied approach: a master data model that lets busi-ness users map out alternate structures for joiningfinancial and marketing data while preserving stan-dard structures of the past - and allows going back toan earlier version if a new model isn't working.

A change happens: your company shifts its Phoenixoperation from the Central Region to its WesternRegion. Expenses and revenues related to Phoenixmust be recoded and moved from Central to Westernregion, and that task is probably carried out with suf-ficient accuracy. But an indirect expense is easily over-looked. How do we know that the overhead allocationfor Phoenix also moved from Central to Western? Ifthere are ten such organization chart changes, what isthe best way to ensure that the overhead always trav-els with the reassigned city? MDM is an answer to thischallenging issue.

MDM enables quicker analysis across an enterprisethat is growing by acquisition. This is a key businessbenefit. The acquiring company can quickly map thetarget company's chart of accounts to its own, butdata and analysis on both the new entity, and the sep-arate pre-merger entities, is still available.

CentralizationLarge companies benefit by centralizing most areas oftheir information management, and this is true ofmaster data. MDM provides a central repository for allmaster data in the enterprise, allowing:

• Consistency

• A common analytic language across the company

• Easier administration of each subscribing tool/application.

Unlike in-house systems that usually focus on central-izing transactional data, MDM applies equally to ana-lytic, reporting, and operational systems.

MDM in Banking

The Case for Master Data Management8

The following example is drawn from a large regional bank in the US, which faced the following problems:

• The Cost Center roll-up structure included a "ragged," or unbalanced, management reporting hierarchy up to18 levels deep, and a legal entity external reporting hierarchy.

• Frequent acquisitions required reporting of all levels, showing results with, and without, the acquisitions.

• Many alternate hierarchies were requested for type (branches only versus ATMs only), market areas, and linesof business.

• Center hierarchies were maintained directly in an OLAP tool. All changes depended on one technical OLAPadministrator. Consolidation integrity was not enforced. Hierarchies were not exposed to other systems, suchas the data warehouse, in a usable form.

• Because hierarchies were not derived from an external data model, but were all edited within the OLAP tool, the maintenance effort was growing sharply.

The solution came in the form of an MDM system.

• A centralized model of “Centers” was created in the MDM system.

• OLAP, DW, GL and numerous reporting systems become subscribers to the MDM system.

• Appropriate business users were authorized to change specific components in the model.

The outcome included the following benefits:

• Many more alternate hierarchies for analysis and reporting were made possible, while reducing maintananceworkload by over half.

• Numerous systems are now in sync. Each system uses a portion of the central data model. Not all systems havethe same hierarchy requirements.

• Multiple people, across multiple business functions, contribute to the Center change management process.Accounting acts as the final arbitrator of revisions in the MDM system before publishing to all subscribing sys-tems.

• Completeness and correctness of hierarchies is enforced through the MDM model to eliminate downstreamreporting errors.

The Limits of MDM

Conclusion

With "fifty percent of enterprises maintaining masterdata separately in eleven or more source systems,and eighty percent of enterprises planning on cen-tralizing master data"—as reported by research andadvisory firm Tower Group—the adoption of MDM isgreater than ever. Integration challenges, compliancedemands, the need for a consistent and coherentprocess for data management, and the tendency totake a best-of-breed approach to IT tools drives thenecessity of MDM.

As stated by Forrester Research, "most Global 2,000companies use between five and fifteen separatereporting and analysis applications." Therefore, notagreeing on a company-wide set of reporting andanalytic structures is much like setting up a financialsystem without agreeing on a chart of accounts. Byadopting a vendor-agnostic MDM solution, businessusers are able to handle changes across a wide rangeof systems, from simple tools, to those that affect theentire reporting structure.

MDM solutions can significantly reduce errors inreporting, boost Sarbanes-Oxley compliance, enablecreative analysis of opportunities, and save time andmoney in reporting activities. For instance, a multina-tional consumer goods company who has implement-ed an MDM solution has reduced the time it took toupdate their 23 different customer systems from 161hours to one hour, an exercise performed at leasttwice monthly. This has provided them with a threeyear ROI of $267K.

MDM is a major step to taking control of the criticalissues of organization growth, IT change, the adop-tion of new business practices, and software evolu-tion within the enterprise. A major US financialinstitution upon implementing an MDM solution wasable to centrally manage their thousands of cost cen-ters and charts of accounts. This effort has enabledthem, while managing the same amount of masterdata, to reduce operating costs by two full-timeequivalents (FTE). This financial institution now hasmaster data updated within minutes, compared totwo to five days per month.

The value of MDM adoption is undeniable—it reducesthe reliance on IT, simplifies report management,eases the challenges of M&A, and dramatically lowersthe high probability of errors when replicatingchanges across systems and report groups.

Master data management is the only comprehensive,effective means to address the challenge of keepingreports accurate and consistent in large organiza-tions.

The Case for Master Data Management 9

What are the limits of MDM, in terms of the master data covered? Generally speaking, some lower-level attrib-utes do not need to be handled by the MDM system, as they will not find their way into higher-level analysis.Here's an example: product color. A toy manufacturer will track sales of blue trucks and blue doll dresses, but itdoes not need to analyze across product lines on the basis of color. Of course, at a different company, color wouldbe a key dimension: garment retailing, for instance, where color trends are critical.

Some solutions integrate only with a single set of master data (e.g. product information). This is a serious limita-tion; such MDM systems are not practical for working with the current reality of best-of-breed IT portfolios.

© 2008 Profisee Group, Inc. All rights reserved. No part of this publication may be copied, modified or transmittedfor any purpose without the express written permission of Profisee. All brand and product names are the trade-marks of their respective companies.

ABOUT PROFISEE • Profisee is a specialized services company with expertise infinancial analytics, operational analytics and master data management.

We assist large and mid-sized organizations envision, design, implement andoperate advanced financial systems tailored to meet the challenging demands oftheir complex and dynamic business.

Profisee also specializes in helping organizations implement operational appli-cations as key enablers for decision support, performance measurement, prof-itability improvement and process improvement initiatives.

Profisee provides world-class master data management services for large organ-izations looking to leverage MDM processes and technology to improve dataaccuracy, systems integration, and transparency. Formed from the managementteam of Stratature, the leading MDM software provider acquired by Microsoftin 2007, Profisee employs some of the world’s most experienced MDM consult-ants and the first dedicated MDM project and design methodology.

For more information, visit Profisee online at www.profisee.com.