Driving Analytics in State Government

Can we afford to wait and see?

Analytics in state government is no longer optional. Legislators will not allocate funds without a strong business case. Everyone from taxpayers, voters, and academic researchers to auditors and federal overseers are looking for the data to reach their own conclusions about public processes. They want the latest information, and they want it now. One IT director quips, “Everyone wants the numbers. Even me.”

What’s an agency to do? Building a sustainable culture and practice of fact-based management—and the data infrastructure this depends on—isn’t easy for any organization.

In state government, the hills are higher than ever. (See the Sidebar: Getting Stuck: Nine Crushing Dilemmas in Leading Analytics in State Government.)

We spoke to seven state agency IT executives to understand how they are responding. Let’s use the simple schematic below to describe the components of the analytics “journey.” To be successful in establishing and using analytics, a state agency will want to put a vision and strategy in place; develop accessible data in which it has confidence; analyze, visualize, and share the insights; and use the results to take action.


analytics stairs

Driving forward

Most agencies have made headway in pockets. The state government reality of feast-and-famile funding as well as come-and-go senior leadership opens windows of opportunity, then slams them shut. The result is an uneven landscape of analytics maturity in state government. Some agencies have developed good, clean data but little real analytics. Others have the ability to answer questions, but lack a way to publish the results. Still others have a collection of potentially interesting dashboards, but no consensus on how to use them. The graph below compares seven different state agencies in terms of their analytics maturity.

IT executives have relentlessly driven ahead when they have had funding, resources, and senior leadership engagement. In a few cases they have been able to establish multi-year initiatives that set strong data platforms in place. In these agencies, the results are tangible and striking. Neil Adcox, CIO of the South Carolina Department of Employment and Workforce, says, “Leadership needs to know this is not easy. It will require more time than they think and a moderate to large resource commitment—time, energy, people, and money. They will ask themselves half way through, ‘Why am I doing this?’ But they have to finish it; there’s no other way. It will be worth it.”

Even when progress has been limited to pockets, IT executives have made the best of their opportunities. They have some lessons to share:

Lead the horse to water. If he’s not thirsty, wait a few months and lead him again. One agency IT executive says, “Our leadership was told by the auditors that they need dashboards, but they could never spell out what they wanted. We went ahead and made some examples on our own to get things rolling.”

Develop a methodology that allows you to ask the right questions. The right questions have to reflect a nuanced mix of mission and practical reality, goals which often stand in direct contradiction. Public sector IT leaders stressed the sensitivity that’s required to incorporate numbers-centric performance measurement into mission-driven activities.

Focus on what matters most. Move the needle for everyday work as well as for management performance reporting. This IT executive teamed up with business analysts. Together they are defining the “next right activity”  for each field worker to tackle based on predicted risk factors and potential consequences.  As they get funding to replace core systems, they are building both types of analytics into the requirements.





Preserve the story. Khush Tata, Vice President for IT and Chief Information Officer at SC Technical College System, reminds us that it’s easier to retain the algorithms and decision models than the context that led to them. He used the example of a funding allocation model. To avoid the big mistake, he suggests digging into the story behind the model before settling on changes to improve it.

Publish your data. Make sure it’s right, then put it out there for others to analyze. Establishing an open data program invites participation, but gives up control over interpretation. One agency IT executive said, “I had to become comfortable with the lack of control. Some people will call to make sure they are drawing sensible conclusions; some won’t. You have to let it go.” Another executive says, “We tell our users we have two types of data—there’s the data we are really sure about and then there’s the invalidated data you get in a hurry. They are beginning to wrap their heads around that.”

Drive use. Whatever data and insights you have, get it to those who can put it to work. Help them bring it into their work processes. One IT executive says, “This has to be a concentration. Otherwise everything you’ve done is just a cool widget.”


The public sector professionals taking their organizations on the journey toward analytics deserve our regard. The hills are high; the trek is long, and the obstacles are everywhere. As they are relentless in their commitment, so should we be relentless in our support of their work and acclaim for their accomplishments.


Getting Stuck: Nine Crushing Dilemmas for Public Sector Leaders in Big Data and Analytics

  1. Implementing a culture and practice of fact-based management needs sustained and engaged leadership from the top. But state government is an environment of institutionalized boss-de-jour turnover.
  2. Wading into new processes and technologies takes experimentation, but the big penalties for failure in the public eye push people to their bunkers.
  3. The same is true of transparency. To build trust, you have to be able to talk about what’s working and what isn’t. In state government, admitting that something isn’t working could land the well-intentioned public servants in an unflattering story on the front page of the newspaper.
  4. Uncertain funding and shifting priorities make it difficult to mount a sensible, phased effort over multiple years.
  5. There is a huge benefit in comparing across similar organizations, but even sharing data, let alone standardizing it across agencies is a monumental challenge.
  6. Tangible problems—rotting bridges, accidents and fatalities, failing students, etc. get attention. But that attention is more likely to a cause a witch hunt rather than the concerted, sustained effort to build a platform, literacy, and analytics culture likely to fix the root causes of the problems.
  7. IT skill sets are essential in a mature analytics practice. But state government salary levels make it difficult to attract talented professionals with the right skills.
  8. Public sector leaders have to retain their commitment to their primary mission amidst an onslaught of numerical analysis, performance-based funding, and resource constraints. Keeping score with numbers creates temptations, some say incentives, to move the numbers if not the outcomes.
  9. Legislatures and constituents are demanding that data be publicly available real-time and highly secure, and that also assumes bullet-proof accuracy. To meet this standard requires a level of funding that is rarely available.

For more information, please contact Dr. Jane Linder at NWN Corporation: jlinder@nwnit.com or 781-472-3498