Published in DM Review Online in January 2006.|
Printed from DMReview.com
Knowledge: The Essence of Metadata: The First 100 Days of Enterprise Metadataby R. Todd Stephens, Ph.D.
In the past six years, our organization has moved from ground zero to one of the most recognized metadata implementations in the world. In my educational sessions, one question comes up over and over again: "What should you focus on in the early stages of delivery?" Clearly, the answer is, "It depends on the environment and requirements." However, the vast majority of implementations will address and experience 80 percent of the same issues, challenges and obstacles. The components described in this article will include the functions, activities, tasks and deliverables that every implementation will encounter. With the exception of functions, each of the other three should be viewed as a hierarchal requirement of the one above.
Unfortunately, due to the page and time limitations, only the first two layers of the model will be described in detail (functions and activities). The detail tasks and deliverables will need to wait for another column or perhaps a small book. Figure 1 provides a view of the framework functions and associations that will be reviewed here.
Figure 1: High Level View of Functions
The five basic functions include the development of the strategy, delivery functions, metadata production, metadata's support model and the consumption processes of metadata. Each of these map to an associated function related to the business of metadata. These associations include business development, procurement of raw materials, development of products and services and, finally, the delivery of the end-user experience or solutions. Each of these areas must be addressed by the implementation but the sequencing does not indicate a linear progression. Strategy comes first, but once you complete some of the strategic components you can start some of the delivery functions. The order, progression and dependency of the functions will depend on your culture, experience, and staffing.
A strategy is a long-term plan of action designed to scope the metadata effort in order to achieve the ultimate goal of supporting the competitive advantage of the business itself. The process of creating a strategy generally involves setting targets, performing gap analysis, formulating the business strategy and implementation. While some information workers view strategy as the development of goals, objectives and mission statements, the reality is that strategy includes so much more. The following list is a collection of activities that should be addressed early in the development of the enterprise metadata strategy.
These activities of strategy will vary in duration from one organization to the next. However, short changing even just one of these will cause pain in the later months or years of the program. Developing the program strategy is not a one-time event that is never to be reviewed again. In fact, an annual strategic plan should be performed by any information technology program. Annual plans force the organization to continue to evolve the vision and mission of the metadata program. The action plan is the basic deliverable of the strategy function, and the results can come in a variety of flavors. Action plans are the specific activities that you will be using to implement the strategy. Often these are stated as objectives. For instance, our growth goal objective might be to have a 10 percent increase in usage spanning over the repository collection. Another type of delivery is the action plan which identifies the tasks, responsible parties, timelines and financial commitments.
While strategy and action plans are important, the delivery activities are where the rubber meets the road. Product decisions will be made in the business case but the organization will still need to integrate the systems into the lifecycle of the business. The delivery phase is basically the system development lifecycle (SDLC) reviewed from the metadata perspective. The delivery stage begins with a design specification which states a solution to the customer's problem and ends with the delivery of the final solution. The following activities describe the delivery functions:
Every organization will have different requirements and standards on how products are delivered and supported within the technology area. The focus should be to ensure the customer requirements are addressed and delivered on time and within budget. At the end of this effort, the organization should have a fully functional metadata environment that has been architected from all four architecture perspectives: data, technical, functional and application.
Figure 1 combined the strategy and delivery into a single business function called business development. Taking a look at the specific activities of these two areas, you will notice many of the same ones suggested to anyone starting a business. Developing a business model, business case, accounting systems, marketing plan and acquiring funding are elements of starting a business as well as developing a metadata program.
Metadata's Production Functions
With the application implemented into the technology environment, the program should focus on the business processes of getting information loaded into the repository. The activities are categorized as the products and services that are available to the producer of metadata information:
The producer is the group that actually creates the metadata information. They may produce the information directly or indirectly from an application such as a data modeling tool or interactive development environment (IDE). The metadata program must understand that they serve both the consumers of metadata and the producers. The key to getting metadata from the different organizations is to make the collection process as simple and painless as possible. This is where automated extraction applications are critical. If users are forced to key in metadata after the asset development phase, then the quality drops enormously. End users are less likely to update metadata once they have completed the asset creation process. In addition to making the process easy, the metadata program should strive to add value-add services to the metadata information. Informing the asset producer who or how their metadata information is being used can increase the probability that they will continue to provide quality metadata on a timely basis.
The business analogy used for the producer is the acquisition of the raw material. The raw material in the world of metadata is the structured and unstructured information. As in any business, materials need to be procured, processed and staged for the business activities to add value and resell them. While acquiring the raw metadata sounds simple, we need to understand that the specific activities of acquisition, scheduling, automating, inventorying and supplier services are critical to the success of the program. There is a reason that supply chain services are an enormous part of every business and any metadata program would be wise to understand these basic principles. Customers are demanding end-to-end visibility of their metadata. Due to the demands placed on IT with outsourcing and the changing business environment, metadata programs are more dependent on suppliers of the information than ever before. Development organizations are continually raising the bar in service levels, quality and cycle time for IT products. Metadata supply chains today must work faster while coping with increased complexity introduced by proliferating supply and demand of all IT products and governance.
Support Model Functions
The support model is the basic foundation for delivering the products and services to the customer base. The internal processes of the metadata group must support the overall vision and purpose developed in the strategy stage. The basic ingredients to making that happen are ensuring that the support model includes a focus on reliability, subject matter expertise and responsiveness, and ensuring the perceptions of value are managed. The support model is the collection of activities you perform to make sure the products and services are delivered according to your strategy. The activities of the support model include the following:
Each of these activities helps solidify the basic foundation of product or service delivery. The easiest way to think about this group of activities is to think of them as internal components that support the business function of collecting metadata and then delivering value. The end result of the support model is to answer five basic questions:
The support model helps both the producers and consumers get to a point where they can self serve metadata. The more you can do to push the creation and usage of metadata to the end user, the greater your chance of long-term success.
Metadata's Consumption Functions
The metadata information collected must be used in order to create a value-add proposition. The leadership of the organization must ensure that they have strategy in which the usage of metadata continues to grow over time. At GE, Jack Welch introduced a simple strategy that if the company could not be first or second in the market then they would not be in that industry. The "fix it, sell it or close it" strategy forced the organization to focus on the customer, and that's exactly what all metadata implementations need to do. The following activities will provide a solid foundation in creating this value add proposition:
Where should you focus your efforts in order to ensure long-term value-add to the organization? A celebrated business book author once said nothing happens until someone sells something. All of your activities of developing a strategy, deploying systems, collecting the data and providing deep support are done for one reason and one reason only: the consumption and usage of the metadata. The consumption can be a passive usage from a well-designed repository or an active usage from integrating metadata into the SDLC. At the end of the day, value will be defined by the consumption of the metadata you collect, process and then consume.
The first 100 days are very confusing, especially if you haven't had prior experience in developing a metadata program. The amount of input coming at you can seem overwhelming at times. The basic utility of this framework is to provide you a sounding board where you take these ideas and see how they fit into the business of metadata. Michael Porter (1985) developed a value chain framework where he described a sequence of activities that most organizations perform. The core activities included inbound logistics, operations, outbound logistics, marketing and sales, and service. The activities were supported by additional functions that focused on infrastructure, human resources, technology and procurement. These same functions can be found in our metadata framework. Metadata's value chain does not exist in isolation and must be integrated into many other value chains such as data architecture, data applications, data resource management, technical and functional architecture, and many others. The bad news is that all value chains must be destroyed and rebuilt over time. The value you will add to the organization in the first year may not resemble the value that you bring to the table in the fifth year. If you don't change that value chain of metadata, then someone or something else will.
R. Todd Stephens, Ph.D. is the director of Meta Data Services Group for the BellSouth Corporation, located in Atlanta, Georgia. He has more than 20 years of experience in information technology and speaks around the world on meta data, data architecture and information technology. Stephens recently earned his Ph.D. in information systems and has more than 70 publications in the academic, professional and patent arena. You can reach him via e-mail at Todd@rtodd.com or to learn more visit http://www.rtodd.com/.
Copyright 2006, SourceMedia and DM Review.