Quick links


Common Metadata Framework

Metadata Case Studies


 All METIS pages (click arrow to expand)
Skip to end of metadata
Go to start of metadata

General considerations

It is not sensible to prescribe an ideal model for corporate governance of metadata. Every statistical organization works under different legislation, organizational arrangements, organization culture, business rules and levels of autonomy with respect to other public sector agencies.

This section looks at good practices for governance. Each statistical organization implementing a metadata management strategy will evaluate its own objectives, strategies and organizational arrangements. There is value in first considering the experience generated by organizations that have already done this.

The MetaNet project, sponsored by the European Commission, included a working group on adoption issues for statistical metadata systems. This group conducted a survey of statistical organizations in 2003, the results of which are referred to in this section, particularly those concerning barriers and organizational issues - both of which are relevant to governance.

Lessons for good corporate governance of metadata

The following are examples of the lessons for corporate governance of data and metadata management based on experiences of statistical organizations in the implementation of a metadata management strategy:

i.     Senior managers, including the Chief Statistician, should be closely involved in developing the vision, formulating policy, approving SMS development plans and evaluating progress.
ii.    The roles and accountability of all organizational units with respect to metadata should be clear. Subject-matter areas are responsible for the creation, maintenance, re-use, and approval for dissemination of all data and metadata content for their statistical domain. A 'corporate data management unit' could be responsible for providing client support, developing and maintaining infrastructure and providing training.
iii.   The organization should adopt an information management culture. All staff must understand their responsibility to work towards achieving statistical integration, comparability of statistics across surveys and time, and to reuse statistical metadata as appropriate. These goals are achieved by adherence to the metadata management principles.
iv.    Specialists, such as business and systems analysts, information technology architects and statistical standards experts, are more likely than others to come across new opportunities for advancing better metadata integration, so a particular focus is needed on working with these groups.
v.     Make sure that your organization has an endorsed metadata strategy, including a global architecture and implementation plan, and that this strategy is integrated into broader corporate plans and strategies.
vi.    Either commit yourself to a metadata project - or don't let it happen. Lukewarm enthusiasm is the last thing a metadata project needs.
vii.   There is often scepticism in the organization against metadata projects. Metadata projects are usually strategic projects for the organization. If they should be carried out at all, managers across different levels and parts of the organization must all be committed to the project.
viii.  Metadata projects are often more abstract, complex and difficult to manage than other types of projects. These characteristics need to be recognized in project plans and the importance of communication with the rest of the organization about the project cannot be overstated.
ix.    Learn from failures and successes in other statistical organizations. Benchmarking and international cooperation are always useful. The case studies referred to in Section 8 are very valuable in this respect.
x.     Systematically use metadata systems for capturing and organizing tacit knowledge of
individuals in order to make it available to the organization as a whole and to external users of statistics.

Back to top

Risk management

Realizing there are potential barriers is an important part of the management and governance of metadata projects. Appropriate risk mitigation actions are a significant component of project governance.

This subsection explores some of the potential barriers to the adoption of metadata solutions - technical, organizational and human - that were identified by the MetaNet survey referred to above. This survey included questions to identify and assess the importance of potential problems, as well as to go into more detail concerning the different aspects of human related issues.

Back to top

Important challenges to introducing metadata systems

The respondents were asked to answer the following question: "For each aspect of metadata please indicate what in your view poses the greatest challenge to the introduction or use of statistical metadata systems in your organization".

The result of this for all organizations was the following.

Figure 5: Perceived challenges to introducing or using SMS (click to enlarge)

The greatest challenges in relation to documentation and retrieval of data are considered to be partly organizational and partly human. However, technical challenges are the most important in relation to documenting data for exchange and retrieval by IT systems.

Back to top

Human issues in relation to adoption of metadata systems

The MetaNet working group reported that the human factor is fundamental to the successful adoption of metadata systems, and a number of challenges were identified.

There might be a substantial gap between some of the more theoretical and abstract contributions on metadata versus what are considered applicable by practitioners within statistical organizations. Some subject-matter specialists tend to delve into one specific area and not take into account the long term perspective for documentation. Motivation for general metadata solutions might therefore be low. There is a need to acquire input and feedback from subject-matter specialists from different areas regarding metadata/data concepts and methods in order to come up with viable common standards and methods for metadata. However, the viability of metadata solutions requires the motivation and commitment of metadata providers. Given that an inability to engage this community constitutes a major concern for many statistical agencies, it is important that underlying human barriers are fully understood if they are to be addressed in future.

The respondents in the survey were asked to identify the most significant barriers to the provision of effective metadata within their own organization, as far as human issues are concerned.

Figure 6: Perceived barriers to the provision of effective metadata (click to enlarge)

The result indicates that there is no clear consensus on what the main barriers are. However, the majority think that loss of individual power as a result of providing metadata, something frequently perceived as a significant deterrent, is not significant.

Other possible issues that have often been suggested, namely that metadata provision is boring and does not benefit the provider directly were largely identified as being of only medium importance. Of greater significance was the belief that time spent providing metadata detracts from the real job. This suggests that the importance afforded metadata creation is low and that this activity will inevitably suffer at the expense of traditional work aspects. Moreover, over 25% think that management underestimates resources needed for metadata capture. If the barriers to effective metadata provision are to be overcome, the activity must be given greater importance. This demands not only the education and active involvement of would-be providers, but also increased management awareness and support.

Back to top

Organizational issues

The MetaNet report says: "It is often emphasized that it is necessary to ensure the commitment of top management in order to succeed in putting in place metadata solutions". In addition, organizational issues might be critical when planning and implementing metadata strategies and applications, even if these are not often put high on the agenda for meetings discussing metadata. Questions, such as what type of staff should be involved and to what degree there should be a central unit and which tasks such a unit should cover, must be discussed. The survey tried to uncover some of these issues.

However, three fundamental issues need reflection in order to address organizational issues:

i.     It is important to reach a common understanding within the organization of what metadata are and what their functions are. The specification of these functions will have implications on how projects should be designed and organized.
ii.    Organization of tasks related to metadata should be based on the information management strategy for the organization. One reason for the failure of specific metadata projects can be that they are not anchored in a more global view of the information architecture.
iii.   In order to sell the need for basic changes in technology or organization to improve data/metadata management, it is necessary to present the benefits and the proposed solutions in an understandable way, possibly based on practical experiences acquired in other organizations. Management is not prone to take decisions involving risk to continuity of production. Once again, management might not be a barrier, but the limiting factor might be the experts' ability to come up with convincing and practical proposals related to metadata. Proposals that reach too far and have a too long a time perspective will have difficulties as management will normally ask for results within a short time frame.

Back to top

The degree of central coordination

One might assume that a central coordinating unit is a signal that metadata/documentation is taken seriously and that there is a relatively high level of horizontal coordination. According to the responses to the MetaNet survey presented in the table below, only three statistical organizations reported having a strong central coordinating unit. The majority of them had a coordinating unit with limited tasks. Data archives apparently have a stronger central coordination of metadata. Even though data archives are often small organizations will more limited tasks compared to statistical organizations, they are information management experts and recognize the crucial role played by metadata.

Back to top

Tasks of a coordinating unit

The contact persons were asked to indicate the tasks allocated to any coordinating unit. It is interesting to note that for a majority of both the statistical and other organizations having some central coordination, an important task for this unit was to develop common systems and solutions.

These coordinating units also have an important role to play in developing common terminology and standards and to ensure general coordination and information in this field of work. Supervision and training apparently is not an important task of many units.

Back to top

The involvement of different types of specialists

Information technology specialists appear in most organizations to have central positions in relation to planning and development of metadata systems and solutions, which is not surprising due to the traditional importance of metadata in computer based systems. Also, in most organizations, specialists in statistical methodology have a central role in this area. However, it is perhaps somewhat worrying that management and subject-matter specialists are to a lesser degree involved.

Back to top

Cooperation with other organizations

Metadata is a field of work where one should expect a large degree of cooperation with other organizations in order to ensure harmonization and exchange of best practice. The survey confirms this. A large majority of statistical organizations foresee cooperation with colleagues in other countries in this field, whereas many also see a possibility for cooperation with consultants and vendors of information technology systems. Cooperation with international organizations is also foreseen. Many statistical organizations foresee cooperation with data archives/documentation centres.

Many large statistical organizations are searching for efficient models for handling data and metadata in an integrated way throughout the production process. Decentralization of technology, in some cases also leading to loss of central documentation of files and processes, has in many organizations made it even more important to find ways and means for coordinating documentation across the organization.

Thus it is useful to look more into the experiences of different organizational models in order to achieve common and efficient metadata solutions.

Back to top