Home / Guidance By Gwen / Choosing Governance Models

Choosing Governance Models

It’s important to define the organizational structure of your Data Governance program. But before you can do that you have to define your governance model at a higher level. You need to consider what types of decisions your governance bodies will be called upon to make, as well as the policies and standards they’ll be establishing. Consider the “direction” that your governance decisions will follow through your organization:

      • Top-down
      • Bottom-up
      • Center-out, or
      • Silo-in

Your program will probably employ all four of these models at different times. Understanding them will help you design structures that have appropriate linkages.


Top-down governance and decision flows

Will your leaders be making decisions that affect part or all of the organization? If so, you’ll need that decision to be communicated down through governance groups, much as executive decisions flow down through the org chart via management channels, with each sub-layer of management adding details as appropriate.

Top-down governance and decision models are based upon authority patterns. Once made, decisions are generally not open to debate, and compliance is not optional. Workers are expected to do as they are told.

For this to be successful within a Data Governance framework, you’ll need a clear linkage between your executive-level Data Governance Council and other elements of your program. Consider whether your program’s sponsor will have the time to be that link, or whether a member of the DGO and/or tactical stewardship groups should be at council meetings to scribe.

This model – and all other models – are also influenced by a condition that auditors call “Tone from the Top.” If you have an official policy, but leadership speaks or behaves as if the matter is not important, then middle management and line staff are less likely to think it is important to adhere to the policy.


Bottom-up governance and decision flows

In contrast, some data-related decisions are made by individuals as a part of their everyday work, and the results bubble up through an organization.

Examples are naming standards for data. Sometimes these originate with a single, dedicated team. Eventually they become broader standards, and then perhaps later enterprise standards.

For this to be successful, individuals with stewardship responsibilities must know who their point of contact is within the stewardship hierarchy and/or Data Governance Office. Suggestions and issues and requirements must be able to “bubble up” to higher levels.


Center-out governance and decision flows

Sometimes there’s no substitute for experts telling us all what we should do.

Your Data Governance Office (or CIO or other leader) may ask experts to specify data models, to interpret compliance requirements, to design controls, or to set protocols. Center-out decisions are made by one or more centralized resources who consider options and then decide what is best for the enterprise. Leadership may ultimately issue a mandate (make a top-down decision), but before that happens it is the centralized group that considers options and their impact on stakeholders.

For center-out governance and decision flows to be successful, you can’t rely upon the authority of the source. Rather, you should set up multiple paths of persuasion to convince stakeholders to abide by a center-out decision. Consider:

      • Asking leadership to set a strong “Tone from the Top” regarding compliance
      • Employing multiple communication channels to send clear messages about the importance of the decision and the need for compliance
      • Educating stakeholders about why the decision was made, and what alternatives were considered

Be sure to set up a clear exception process, issue escalation path, and issue resolution process


Silo-in governance and decision flows

The silo-in decision flow brings together representatives from multiple groups to collectively agree on a course of action. Data Governance and Stewardship Councils are examples of this model. With objectives of balancing the needs of individual silos against the needs of the entire organization, such groups meet to .issue policy, set standards for the enterprise, and decide how to resolve data-related issues.

In making silo-in decisions, the council is expected to provide multiple perspectives. Its members are expected to consider impacts to stakeholder groups across the enterprise.

Sometimes this model is followed even though a central group of experts has agreed upon a recommendation for a course of action. Why? Often, members of a federated organization will accept a tough decision made by a group better than an edict from a centralized source.

Silo-in decision flow models can be very effective at making “decisions that stick.” They eliminate the argument that stakeholders were left out of a decision-making process and therefore should not be forced to abide by the decision. Participants become invested in the process and are more likely to promote compliance.

For this model to work, the group must be granted authority by leadership to make decisions on behalf of the enterprise. Representatives must be authorized to act on behalf of their groups, and they must be respected by those they represent so their decisions will be accepted.


Hybrid models

Some decisions need to be made at multiple levels. One group may craft a recommendation, and then a higher-level group will make the final decision. When politics or budget/resource impacts are involved, a hybrid decision flow model may called for: a center-out or silo-in recommendation will be followed by a top-down decision.


Next:  Defining Organizational Structures


Image courtesy of Stuart Miles at FreeDigitalPhotos.net

About Gwen Thomas

Currently the Corporate Data Advocate at the World Bank Group's private sector arm (IFC, The International Finance Corporation), Gwen Thomas is the Founder of The Data Governance Institute and primary author of the DGI Data Governance Framework. Gwen has personally helped build Data Governance programs at the Federal Reserve System, Sallie Mae, Disney World, NDCHealth/Wolters Kluwer, American Express, Washington Mutual Bank (WaMu), Minnesota Pollution Control Agency, Wachovia Bank, Coors, and others. Gwen frequently presents at industry events and contributes to IT and business publications. She is the author of the book Alpha Males and Data Disasters: The Case for Data Governance.