The critical decision that must be made in the planning phase is whether you will centralize the data consolidation effort or build a distributed solution that leverages the data owners for the various data sets that you wish to consolidate. As you identify the business drivers for data consolidation and create a vision for the agency’s long-term data integration capabilities, a set of competing imperatives surfaces. Decision parameters like the timeframe in which you need to create the consolidated data set, the degree of influence you wield over impacted stakeholders, data users, and subject matter experts (SMEs), or the resources available for the effort start shape the project plan and direction.

A closer look at these imperatives, however, likely reveals short-term issues that can create risk for a viable long-term data capacity within your agency. For example, project leaders may be tempted to meet urgent requirements by building a centralized IT solution with the capacity to apply data standards and store data from a broad range of systems, which gives them control and removes many external dependencies. The alternative is collaborating with data owners to implement new data standards within the source systems, performing validation locally, which takes time, training, patience, and is subject to the limitations within legacy systems.

Placing responsibility for data standardization, quality, and validity in the hands of data owners and source systems is a key step forward in creating a data-centric agency culture. Data owners develop competency around inventorying their data, defining and implementing new standards within their systems, developing and efficiently executing robust test plans, building remediation processes to address errors, and a host of other skills that result in expanded capacity and data sophistication for your agency.

It goes without saying that before you can standardize and consolidate data, you must have a complete picture of the data within your agency, where it lives, the processes that create it and consume it, how and where it interconnects, what the current data definitions are, and who the stakeholders are for any given system.

One way to build your data map and increase data acumen across the agency is to engage stakeholders in creating a standardized, templated view of existing data across your agency. The more common approach is to do a data call and then normalize the input to create an apples to apples picture of the agency’s data assets. There are some unintended consequences with this approach, however, not the least of which is that the central group that grooms the data call results will have to do it every time there is a new data integration requirement.

Building a set of tools and templates that educates system owners on data definitions and the ways that the consolidated data will be used, along with training, brown bags, and working group sessions builds agency capacity to adopt data standards. Data standards are the foundation for actionable data—you know what it means, you can trend it, compare it to results from disparate processes and external sources, etc. This requires collaboration among stakeholders and creates relationships that will increase efficiency as the integration effort continues.

With a successful pilot phase under your belt, it is tempting to assume that full implementation will be likewise successful. The significant difference is the expanded user population and number of systems and datasets that are all trying to get along for the first time. It is important to maintain engagement with testers, data users, and solution implementers throughout this step. Tap into their expertise and lessons learned from piloting to conduct “mock” production runs of your agency’s data.

Here, you want to ensure that the processes tested during piloting are now scalable to your agency’s broader stakeholder group, end users, and consumers of your data. The focus remains on building a data-centric agency culture, which takes time. Identify data stewards and super users that can become change agents in their business unit or agency component. Create a mechanism for capturing user feedback and respond to it. Test, test, and test again. Implementation does not mean error-free completion. It means that your agency can start using the data. It does not mean that the data is perfect, which is important for agency leaders to understand. Perfecting the data happens only when it becomes available for use.

Achieving data quality is not a one-and-done effort. For large data integration efforts, it is a constant struggle informed by the feedback of the people using the data for analysis and decision making.

Data consolidation solutions generally encompass both automated and manual checks for accuracy, completeness, and logical consistency. We have found that systems that rely solely on automated quality checks often miss the mark. Why? They allow valid entries to slip through that don’t make sense contextually. Having valid numbers doesn’t mean the numbers are correct, and often human eyes are needed to discern practicality—even on a small sample of entries.

Agencies need to establish tools and processes to manage data sets, data quality monitoring, data validation, and security/quality controls. Formally documenting processes for examining validity, robustness, and defensibility drives accountability with data owners, which in turn, increases adoption and engagement with the new solution.

Vetting data and implementing feedback loops for suppliers to cleanse inputs as needed is an inherent requirement for usable data. After data quality reaches acceptable levels, agencies are then able to start realizing the value of the data by processing and reporting on the collective data set in complex and useful ways.

The ultimate outcome of this step is data usability. Is the data in a consumable format, i.e., open source? Is data structured in a way that all stakeholders can use? Can it be used to drive effective decision-making in your agency? Does your data help your agency fulfill its mission? The goal is to be able to answer “yes” to any of these questions.

How much is data worth? Facebook is now worth about $200 billion. United Airlines, a company that actually owns things like airplanes and has licenses to carry passengers on routes all over the world, is worth $34 billion.

Federal agencies are not in the money-making business. The value of data for government is conveyed by adding value to people’s lives. Agencies can better serve their constituents when they understand them, when they can assess their operations, and when they can measure their impact. Data, combined with the way that it is used, delivers lasting value that makes a difference. How will your agency get there?