Affiliated with:

Challenges of Metadata Silos

Metadata management has become increasingly important for transactional and analytical applications.  Creating metadata silos causes numerous challenges; avoid creating siloed metadata

Metadata management and its use in enterprise information management has become one of the critical information technology (IT) focuses for both global 2000 corporations and large government agencies.  As these entities look to reduce their IT portfolio and control their escalating IT costs, they are turning to the technical functionality that a managed metadata environment (MME) can provide them.  The organizations that have built well-architected enterprise-wide MMEs have achieved a tremendous amount of success.  Unfortunately, like most popular IT trends, companies are making key mistakes in building and implementing their metadata management investments.  One of the chief problems is that they are not building one MME; rather, they are building lots of metadata repositories, none of which speak to each other and do not follow an overall metadata management strategy.  There has been a proliferation of un-architected and disjointed metadata repositories, causing many problems.

The problem of disparate initiatives is not unique to metadata management.  On the contrary, it is a common issue that has plagued many areas of IT.  Technologies like data warehousing, enterprise resource planning (ERP), supply chain management and all flavors of transactional systems have suffered with needless duplication and redundancy.  The four most common problems created by disparate metadata repositories are:

  • Missing Metadata Relationships
  • Solutions Typically Built By Non-Metadata Professionals
  • Costly Implementation and Maintenance
  • Poor Technology Selections

Missing Metadata Relationships

There are different types of metadata (see Table 1: “Types of Metadata”) that must be properly managed and linked.  For example, it is very valuable for an IT developer to have the capability to go to the MME to look at the technical transformation rules (technical metadata) that are being applied to a particular physical field name on a report that is being analyzed.  Once the developer has reviewed this metadata, they can navigate through the MME to find the business rules defined by the business users for that field.  If a discrepancy between the transformation rules and the business rules exists then the developer could use the MME to contact the data steward who defined the specific business rules and resolve this discrepancy.  This is the true power of properly managing metadata, bridging the gap between business and the IT systems, since business operates through IT systems.  When metadata is not managed from an enterprise perspective, this type of click-through analysis is impossible because the relationships between the metadata (both business and technical) are not being captured or maintained.

Table 1: Types of Meta Data

Poor Technology Selections

It is common to find many disparate MME or repository-like initiatives.  For example, one client has over 14 separate MME initiatives (either in production or under development), while another client has over 25 disjointed repositories, most of which have significant monetary expenditures associated with them. 

Disparate MME initiatives can come in many different flavors, sizes and shapes.  There are large repositories which utilize enterprise-level metadata integration tools or are even custom built.  In addition, there are lower technology metadata efforts that come in the form of spreadsheets, which, along with local databases are the most popular form of metadata repository technology.  Is this statement a surprise?  Effective metadata technology solutions should be employed and the selection should be based on an enterprise metadata strategy.

Costly Implementation and Maintenance

Typically, business units and technical groups need metadata and properly functioning managed metadata environments (MME) to operate the business and their IT (information technology) systems.  Therefore, these groups will not be able to wait around for their company to start building their MME, and they may not even have such a project in their current IT plans.  As a result, they quickly build these disparate metadata repositories solutions that are designed to handle only one or two specific problems/challenges.  If executive management knew the cost of these metadata repository point solutions, they would see that it far outweighs the cost of a truly sound enterprise-wide MME. 

This situation closely follows the path that many companies have taken with data warehousing.  For example, the average company needlessly replicates their business intelligence efforts across many lines of business, as opposed to centralizing this function.  Experience shows that this approach can increase the costs of BI / analytics by over 300%, and metadata management has similar siloed initiatives with large cost overruns. 

The types of organizations that build single-point metadata solutions often are concerned by the cost of building a MME.  However, the cost of the MME would pale in comparison to the costs of all of the disjointed metadata initiatives that are currently underway or in production (see Figure 1: “Metadata Management Costs.”  Doing it right the first time is always less costly than doing it wrong and trying to fix it later.

Figure 1: Meta Data Management Costs

Typically Built By Non-Metadata Professionals

When government agencies and corporations have disparate metadata repositories, invariably most of these applications were built by consultants and employees who are not qualified to do so.  This is highlighted by the fact many consulting and software firms are entering the metadata management field even though they have little or no experience in this area.  Metadata professionals study their craft, have made it their full-time focus, and in many instances it has become their life-work.  An operational systems person or a data warehousing person cannot be expected to work in the metadata arena if they lack the proper knowledge, training, and experience.

A MME is not a data warehouse or an operational system.  Building the MME incorrectly will only lead to having to rebuild it again in the future.  In an article written for January 2000, the author stated, “During the 1990s, corporations raced to build their decision support systems as quickly as they could…in their zeal to do this too many of these organizations neglected to build the architecture necessary to grow their systems over time.”  This prediction came true, and it applies to managed metadata environments (MME) as much as it applied to decision support / business intelligence / analytics initiatives.

Conclusion

Companies that build sound enterprise-wide MME will have a distinct advantage over those corporations and agencies that went down the disparate path.  Avoid metadata silos; build a central managed metadata environment based on an enterprise metadata strategy.

LinkedIn
Facebook
Twitter

Dr. David P. Marco, PhD, Fellow IIM, CBIP, CDP

Dr. David P. Marco, PhD, Fellow IIM, CBIP, CDP is an internationally recognized expert in the field of data warehousing, business intelligence, enterprise data management, data governance, and is the industry’s leading authority on metadata. Mr. Marco is founder and President of Enterprise Warehousing Solutions, Inc. (EWS), a Chicago-based enterprise data management consultancy dedicated to providing clients with best-in-class solutions. Author of several books and hundreds of articles and a Certified Data Management Professional, Mr. Marco is also a well-known speaker in his areas of expertise at conferences and symposia.

© Since 1997 to the present – Enterprise Warehousing Solutions, Inc. (EWSolutions). All Rights Reserved

Subscribe To DMU

Be the first to hear about articles, tips, and opportunities for improving your data management career.