3. Getting Started

Overview

This section provides step-by-step guidance for getting started with data governance – how to assess the current situation, establish goals, and create a roadmap tailored to the agency’s needs. There are nine steps, illustrated in Figure 1.

Adapt the process to your situation. This nine-step process can be adapted to situations where data governance is being implemented in a limited fashion. Regardless of the scope, It is still important to identify stakeholders, establish goals, assess current practices to identify gaps, determine a governance approach and make a plan.

Consider creating a data plan. Some agencies choose to launch a data governance effort by creating a data business plan or data strategic plan. The plan development process can be structured to carry out steps 2-8 of the data governance initiation process shown in Figure 1, producing a plan with a roadmap and initial action plan.

This will be an iterative process. You will learn as you undertake these activities, which will lead to course corrections. In addition, there will also inevitably be changes in agency leadership and changes in staff with direct involvement in the data governance effort. These changes will also necessitate re-examination of assumptions and decisions and create the need to modify your plan. That is the reason for Step 9-Implement, Monitor and Adjust.

Step 1: Identify a Lead

Every new initiative needs a leader. Data governance, like any major initiative, requires a lead to steer the effort. Therefore, the first step in launching data governance will be to identify a suitable person to play the lead role. The person responsible for launching a data governance effort is not necessarily the same one who will be designated to manage it once it has been established.

Find the right person. There is no single best organizational home or role for the lead, but there are some qualities that will contribute to their success:

• Management support—close working relationship with a Deputy or Division-level manager who serves as the executive champion for the effort;
• Interpersonal and communication skills—ability to get along with a variety of individuals across the agency: an effective collaborator, communicator, and change agent;
• Data experience—two to five years of experience working with data – as a user, data manager, or database or report developer;
• Project management experience—understanding of how to plan, track, and control a project;
• Logical and analytical mind—ability to formulate logical arguments and approach problems in a systematic fashion; and
• Passion—strong interest and commitment to improving the agency’s use of data.

Failure to assign responsibility to make sure data governance happens is one of the most commonly cited reasons when agencies say they have tried and failed to establish or maintain a data governance effort. Another related reason for failure is putting the wrong person in charge. A person who does not feel supported, empowered, or who does not want to do the job will not succeed.

Step 2: Identify and Assemble Stakeholders

Organize a stakeholder group. Once the lead has been determined, the next step is to identify and assemble an initial group of stakeholders to serve as a steering committee for data governance planning. This group will be responsible for early outreach and assessment activities, formulating goals, selecting a data governance structure and approach, and establishing a roadmap and action plan. Many agencies hire consultants to help them with these early steps; in this case, a strong stakeholder group can review deliverables and connect the consultant team to different parts of the organization.

Select the right collection of members. The data governance effort will eventually benefit from the broad participation of stakeholders who collect, manage, and use data. However, starting a data governance effort with the large superset of all who have a stake in the data can overwhelm the early efforts. The initial stakeholder group should consist of 5-10 people who have the knowledge, positional authority, organizational savvy, and interest to set the right strategic direction for data governance and help to move it forward to implementation. Therefore, placement in the organization as well as personal qualities are important factors to consider in identifying members of this group.

Include both supporters and skeptics. Each member of the stakeholder group should be motivated to participate because they see the potential for data governance to benefit their area of responsibility as well as the organization. However, it can be helpful to include a few skeptics in the group who will ask hard questions about costs and value to be added. Their motivation for participating may be to prevent the organization from adopting new policies and procedures that impede flexibility, which is an important consideration to keep in mind.

Consider including content and records management representatives. There may be opportunities for synergies and resource pooling across data, content and records management. For example, consistent metadata for data and content enables users to search across different types of information. A coordinated approach to retention and archiving for datasets and documents avoids establishing separate and potentially duplicative or inconsistent processes. Involving individuals responsible for content and records management functions allows for early discussion of collaboration opportunities and consideration of whether to adopt a broad information governance scope or keep the data governance scope limited to structured data.

Consider engaging field office representatives. If the agency has strong and semi-autonomous field offices (e.g., Districts or Regions), it is a good idea to include field representation in the stakeholder group. This will help the group discuss the impact and benefits of data governance from the field office perspective. Ideally, the field representatives will buy in to the decisions about data governance that the group reaches and can serve as effective communicators and ambassadors to other field offices once data governance is being rolled out.

Consider engaging external stakeholders. In some situations, it may be beneficial to include representatives of partner agencies, either as full working members of the initial data governance stakeholder group or in an advisory capacity. They are very often partners at the Tribal, local or regional level in asset management, planning, design, and safety projects. They may supply

• Option 1 - Selected data program(s). In this option, data governance is applied at the data program level – for example, road inventory data, traffic monitoring data, or data to be brought into a data warehouse, enterprise GIS database or other central data repository.
• Option 2- Selected business units. In this option, the scope of data governance is limited to one or more related business units – for example, asset management, performance management or safety management.
• Option 3 - Agency-wide, selected data. In this option, the program is agency-wide, but only certain types of data are considered in scope - for example, data stored in enterprise systems, data pertaining to the transportation system, data that are shared widely, data that are reported externally, data that are private or sensitive. You may also want to be specific about whether you will be governing just structured data or documents as well.
• Option 4 - Agency-wide, all data. In this option, the initial data governance program is comprehensive in scope, covering all business units and all types of data.

The selection of an initial scope will depend on your motivations for pursuing data governance and the level of management buy in or commitment that exists. An agency-wide scope for data governance (options 3 or 4) will be best for agencies that need to address state-level data governance requirements, and those committed to a long term, sustained effort to improve the quality and value of data for decision-making.

Some agencies may choose to begin with options 1 or 2 to address priority pain points or opportunities. These options may also be pursued initially because they require less resourcing and provide an opportunity to test approaches and demonstrate value prior to an agency-wide rollout.

Business Goals for Data Governance

Once the scope has been established, the stakeholder group can initiate a discussion of what you intend to accomplish through data governance and how this aligns with the agency’s strategic directions. Key questions are:

• What problems are we trying to solve?
• What is our vision for how things will be different in the future?
• How can data governance help us to accomplish our current priorities?
• What is our data and analytics strategy – and how can data governance support it?
• What is the value proposition for data governance at our agency?

The product of this step is a set of goal statements that can be used to communicate to agency leaders and others.

At this stage, focus on identifying data-related pain points that:

• pose risks;
• create inefficiencies;
• prevent the agency from meeting state or federal requirements; and/or
• hold back progress on key initiatives for improving safety, project delivery, asset state of good repair, mobility, and other key agency program areas.

In Step 5, you will be gathering more detailed information about current practices and identifying specific gaps to be addressed.

Table 1 presents typical DOT problems or issues and associated goals for data governance that can be used as a starting point for discussion.

Table 1. Typical DOT Data Governance Goals

Stakeholder Engagement

To inform the process of formulating goals, seek input from agency leaders, business line managers, data managers and data analysts. Make sure to include the largest data producers and consumers.

Plan and organize outreach activities – including surveys, interviews, webmeetings, and workshops to learn about stakeholder concerns, pain-points and perspectives on data governance. These activities are time-consuming but are important to shape an effective data governance program and obtain the buy-in necessary for successful implementation.

A strong stakeholder engagement effort will provide a solid foundation for the goals that you establish. It will make people in the agency feel heard and help to build awareness of what data governance is and how it can help.

Step 5: Conduct an Assessment

Purpose of an Assessment

A structured data governance assessment can provide an opportunity to examine your current practices and identify the gaps between where you are and where you want to be. This provides the basis for discussing what actions your agency can take to close these gaps. Actions can be prioritized and become the basis for a data governance roadmap (Step 7) and action plan (Step 8).

The assessment can be repeated periodically and used to track progress and revise priorities based on what has been accomplished and learned.

An assessment produces a score on a maturity scale (e.g., from 1 to 5) for several different aspects of data governance. The score can be valuable for communicating the need for data governance to agency leadership. However,

Each sub-element is assigned a maturity rating from 1-Initial to 5-Sustained. Figure 2 below shows an example radar chart that is produced from the tool that shows the gap between current and target maturity levels for different sub-elements. The blue area in the middle represents the current state; the gray area represents the gap between the current and target states.

While the assessment tool described above was created specifically for DOTs and serves as a supplemental tool to this Guide, there are a variety of other available data governance maturity models and assessment tools. Agencies can review what is available and choose one that suits them best.

Supplementing the Assessment

In addition to conducting a broad assessment of data governance and management maturity, agencies can consider gathering more detailed information for particular categories of data, including:

• An inventory of datasets – contents, size, where stored
• Costs of data gathering/collection, processing, and management
• Data customers – internal and external, how they use the data
• Short- and long-term data improvement needs - considering data quality, availability, and accessibility
• Anticipated changes to data needs based on new legislation, reporting requirements, or agency initiatives.

This more granular level of assessment can be time consuming. It is best conducted once data stewardship roles have been defined, so that stewards

can be engaged in the process. However, if the data governance effort is scoped at the data program or unit level, detailed information gathering is feasible and can provide a solid anchor for developing a data governance roadmap and action plan. Information on data use and costs can also be valuable for prioritizing activities and identifying opportunities for cost reductions.

Refine Your Goals

Based on the results of your assessment, take each gap that was identified and try to assign it to one of the goals established in Step 4. For example, a gap might be the lack of an agency data catalog. This gap would fit within a goal related to data sharing or ability of employees to find what data is available (findability).

Update or augment the high level goals as needed to make sure that all of the identified gaps are accounted for. Create a matrix that maps the gaps you intend to address with the final set of goals. This matrix can be used to guide the roadmap and action plan activities in Steps 7 and 8.

Step 6: Select a Governance Approach

Once there is agreement on goals for data governance and the gaps in practices to be addressed, the next step is to consider alternative approaches to data governance and decide on an initial approach. Key questions to ask are:

• What is our initial data governance operating model?
• What level of change should we pursue?

Operating Model

As used in this Guide, a “data governance operating model” refers to where and how data policies, standards and practices are created and implemented. An operating model for data governance is characterized by two dimensions: (1) the degree of centralization and (2) the degree of compliance focus for implementation. Three basic options can be distinguished:

Option 1 – Centralized-Command and Control. In this option, central governance bodies develop policies and standards and set up mechanisms to make sure that they are followed. There is a strong compliance focus.

Option 2 – Centralized-Facilitated. As in the first option, central governance bodies develop policies and standards, but they act as conveners and facilitators rather than enforcers. There is an emphasis on communication, consensus building and issue resolution

• improvement projects. These might include efforts to bring new data into the agency, create new reporting data sources, or get data ready for system modernization projects. Embed data governance staff and integrate data governance activities and artifact creation within these projects.
• Feedback and Lessons. Following each business engagement, seek feedback from participants and identify lessons about what was most effective and what needs improvement.
• Updated Guidance, Standards and Processes. Use what was produced as part of the business engagements and the lessons learned to update and augment the data governance guidance, standards, processes and other support materials that you have created. Create a library of example artifacts that others can build on.
• Dissemination/Education. Conduct outreach and training activities to make people aware of any updated guidance, standards, and processes.
• Identify Opportunities and Document Successes. Based on the business engagements, identify opportunities to apply what was created and learned. Document accomplishments and value added.
• Update the Backlog. Update the backlog based on accomplishments, re-prioritize remaining items based on what was learned, and seek out the next set of business engagements.

Option 3 – Decentralized-Coordinated by an Operating Unit above is most compatible with the Agile approach, though elements of Agile data governance can be applied with any of the operating models.

Level of Change

To have an impact, all data governance programs must make some level of changes to the organization. The question to consider is: how much change is the agency prepared to take on – at least initially? Attempting major changes to roles, responsibilities, processes and requirements before there is sufficient management support or appetite for these changes can put the entire effort at risk of failure. On the other hand, being overly cautious about making any changes to the status quo is unlikely to achieve the desired results.

Options to consider for the initial level of change to pursue are:

• Option 1-Low. Minor, incremental changes to existing roles and processes, with an emphasis on education and support to achieve data governance goals.
• Option 2-Medium. Limited, selective changes to existing roles and processes focused primarily on data governance sponsors,

• champions and stewards. Existing management structures and processes are leveraged where possible.
• Option 3-High. Major changes to data management oversight responsibilities, decision-making authority, roles, processes, practices, and tools. Requires changes in behavior for data managers, producers, and consumers across the organization.

Most agencies will choose the middle ground option – leveraging existing management and decision-making processes to the greatest extent possible, but adding “just enough” new processes and role definition needed to achieve the goals that have been established. Another approach is to begin with Option 1 and pursue incremental changes, assess results, and then move on to more significant changes (Option 2 and then 3).

Step 7: Create a Roadmap

Once the initial approach to data governance has been selected, it is helpful to create a roadmap that lays out the major activities or initiatives to be pursued over the next 3-5 years.

Purpose of a Roadmap. Developing a roadmap provides an opportunity to:

• Set priorities—data governance is a long-term endeavor and there are important choices to be made about what to do first – and what to tackle later. Priorities can be based on qualitative judgement or a more rigorous quantitative Return-on-Investment (ROI) analysis.
• Account for dependencies—some activities need to be completed before others. It is important to plan for a logical sequencing of activities.
• Ensure coordination with other efforts—there may be a need to coordinate timing of data governance implementation activities with other agency initiatives such as strategic planning efforts or major system replacement projects. In addition, there may be a need to meet timelines associated with externally driven requirements associated with state-level data governance rules or new reporting expectations.
• Account for resource constraints and management decision-making—consider the capacity of the data governance lead and the stakeholder group to conduct and oversee activities, and the time required for senior managers to approve new policies or other changes. Even when consultant support is used, there are limits to how much can be accomplished within a given timeframe.

Once complete, a roadmap is a valuable tool for communicating to the full range of stakeholders about the planned implementation approach. It helps to set expectations about what is going to happen and when – and tells people when they might need to get engaged.

Roadmap Content and Organization. The roadmap should include a high-level view that can fit on a single page of the sequence of activities to be conducted. One common format (illustrated in Figure 4) is to show a timeline across the top, and category of activities horizontally.

This high-level view can be supplemented by a further breakdown of what is included in each major category.

The content of the roadmap will depend on the results of your assessment, the gaps to be addressed, and the selected operating model and scope. See the sidebar for one example.

In developing the roadmap, consider the following options:

• An incremental but broad-based approach – moving forward in parallel on multiple initiatives at once;
• An incremental but focused approach – moving forward with a single initiative at time; or
• A “big bang” approach – to make significant progress on multiple fronts within the roadmap time horizon (typically using consultant resources).

Roadmap Updates. The roadmap should be updated regularly (nominally on an annual basis) to reflect the speed of progress, identification of new priority activities and emerging external factors. The executive sponsor should approve the initial roadmap and any changes to the roadmap.

Step 8: Create an Action Plan

The Action Plan is created based on the roadmap. It defines the specific, implementable activities at a sufficient level of granularity to enable you to:

• Estimate resource needs for each action
• Assign the action to a lead person

• Identify the products or outputs of the action
• Identify the predecessor(s) of the action
• Track the status of the action (planned, in progress, complete, deferred/eliminated)

However, avoid being overly prescriptive. Allow for some flexibility on the part of the leads to tailor the actions based on stakeholder input and other factors. Avoid having too many actions in your plan; it can set unrealistic expectations and create a burden for tracking and communicating progress. Figure 5 illustrates a simple action plan format.

Treat the Action Plan as a living document. Update it on at least a monthly basis to record action initiation or completion, changes to the scope or timing of existing actions, and inclusion of new actions to meet priorities and opportunities that arise. Keep notes on the next steps to be taken for each action so that you can keep things moving.

The Action Plan can be used to develop briefing materials for your governance groups and other stakeholders to keep them informed about what is happening and the progress being made.

One initial action to consider in the Action Plan is to establish an initial data governance policy that describes the purpose and scope of data governance and the roles and responsibilities of data governance bodies and stewards. This provides an official endorsement and can serve as a foundation for specific data governance initiatives. Each initiative in the Action Plan should be documented through a procedure, practice document or guidance document.

Step 9: Implement, Monitor and Adjust

Move forward with establishing roles and groups (Chapter 4) and implementing one or more initiatives (Chapter 5). As part of this activity, set aside some time each month to review accomplishments, results and any lessons learned. Keep a chronology of key milestones (such as policy adoption, governance body chartering, data catalog rollout). Keep a list of challenges and