Data-driven decisions require data-first culture
Connecting state and local government leaders
To create an environment in which data-driven decision-making can thrive, agencies need to provide analytic capabilities that require no coding and that are intuitive to non-experts.
The concept of data-driven decision-making, one of the linchpins of the President’s Management Agenda, is more complex than it might seem at first.
The current generation of analytic capabilities can help agencies improve the accountability and transparency of government operations and programs, as the PMA envisions. But the PMA also recognizes better decision-making requires more than better analytics. It requires a data-first culture -- one in which people at all levels of an organization recognize the value of data and leverage analytics to cultivate debate and collaboration as part of their everyday decision-making processes.
This definition assumes the development of basic data literacy -- an understanding of how to interpret data, how translate it into insights and how to use those insights to educate others and inform decisions.
These themes are reflected in the administration’s approach to developing a Federal Data Strategy. The key components of that strategy, as defined in the PMA, emphasize the need to develop policies and procedures to improve the access, use and augmentation of data and to integrate it more effectively into decision-making.
It’s easy to see how agencies might find this a daunting task. In the past, the technical complexity of analytic tools presented a barrier to entry to anyone lacking technical or data expertise. Even with good policies and procedures in place, the challenge is to convince agency leaders and employees outside the IT department that they can be effective in a data-driven culture.
One basic problem is that, aside from the data experts and perhaps a small group of spreadsheet power-users, many people are not adept at reading traditional data reports. Given the varying levels of expertise in a given group, it is difficult to get a consensus on what the data means, much less to foster a collaborative approach on how to leverage it.
Additionally, traditional data reports typically are static, often delivered in PDF format. If the data warrants further investigation, users frequently must request another report. Practically speaking, this makes real-time collaboration nearly impossible. But just as important, it tends to create an environment in which employees are less likely to ask probing questions, knowing that answers might be a long time in coming.
Beyond the technical barriers, many people do not understand how to use data to improve the management of a program. Given the right tool, they can track the performance of an initiative by countless metrics, but they might not have any idea which metrics really matter.
All these issues are intertwined. First and foremost, to create an environment in which data-driven decision-making can thrive, agencies need to provide analytic capabilities that require no coding and that are intuitive to non-experts.
Typically, that begins with visualization, whether in a dashboard or another format that helps people to grasp the meaning of data at a glance. Visualization ensures that everyone involved in a program is seeing the same picture -- what in data circles is known as working with a single version of truth -- which is a prerequisite of collaboration.
People also need the ability to interact with that data dynamically. Decision-making rarely follows a linear process, so an analytics platform should make it easy for people to ask question after question, exploring ambiguous results or testing hypotheses. The platform should conform to the way people work, not the other way around.
The easier it is for people to work with data, the more likely they are to see the value of incorporating it into their daily processes -- and to learn how to leverage it more effectively to drive value into programs. People want to do good work, so if they see data as making them more effective, they will buy into it.
This is just as true for top leadership as it is for employees working behind the scenes or delivering services to the public. That is why it is important to showcase the results of data initiatives, not just for the people working on the program involved but for others in the organization to see. Success leads to more success.
Ultimately, people will leverage data as a strategic asset when they see it as a strategic asset. The PMA and the Federal Data Strategy are a critical part of that. But for this vision to become a reality, agencies need to begin cultivating a data-first culture.
NEXT STORY: DARPA looks to procurement bots