The Case for Analytics

I just read a great book, Competing on Analytics, co-authored by a man who should be familiar to anyone with an interest in managing public-sector technology. Thomas Davenport, along with Jeanne Harris, got me thinking that we are ready for "what's next" in leading large-scale public enterprises.

In his seminal work, Process Innovation: Reengineering Work Through Information Technology, written way back in the dark ages of the early 1990s when we were just paving the cow paths, Davenport called on us to move beyond mere automation of traditional transactions. Today much of that work has been accomplished, and government is zooming forward with new citizen access and services.

Public-sector leadership heeded Davenport's message in the '90s because they had to. People were cruising to ATMs to bypass teller lines, and buying airline tickets and books online, yet they had to line up for hours at the department of motor vehicles, and battle for other basic services, such as getting a copy of a birth certificate. Government came under pressure to compete with the private sector in the name of service, or politicians would feel the heat. Many governments have more than met the challenge but others still lag.

Yet, public confidence in government is at an all-time low. Consider how decisions were made going into Iraq and dealing with Hurricane Katrina. Real justifications for policy and administrative actions are hard to come by in the first place, but when our federal government proves to be just flat-out wrong in its predictions one wonders: Does decision-making have to be done haphazardly? Certainly not.

The answer, according to Davenport, is for the public sector to support better policies, create more data-driven administrative work, reduce risk and take advantage of huge data assets by making better use of analytics. Analytics is the high-end portion of business intelligence (BI) that allows government to use data to predict outcomes of new or changing policies, programs, regulations, enforcement patterns or whatever task is at hand. These tools can boost public confidence and lead to better decision-making.

Some government agencies are leading the way. The U.S. Department of Veterans Affairs created an electronic medical records system that is now used to model the impact of aging demographics on chronic disease rates for veterans, to make evidence-based decisions on what medical protocols to favor, and to manage lots of other health services that depend on that department's ability and reputation for serving its clients.

A May 2007 New York Times article credits Rodney Monroe, chief of the Richmond, Va., Police Department, with the innovative use of analytics to reduce local crime by 20 percent, at a time when crime is on the rise nationally. The article explains that the real value of analytics comes from looking at data sources from outside law enforcement, from elsewhere in government or sometimes from other sources outside government entirely, and then integrating it with crime data.

Many governments already own SAP, Oracle or some other enterprise software program. These business applications can mine data with BI extensions that enable analysis of information collected as a byproduct of service delivery. Tools from SAS Institute, Cognos, Business Intelligence and other vendors are also in most government portfolios. However, the obstacles to using these tools are numerous, especially when analyzing data beyond individual transactional systems and even beyond the agency's own internal systems.

Other obstacles include irrational privacy fears and bad data. But the biggest obstacle is making managers and policymakers aware that CIOs can help them do their jobs more effectively.

It's time to go beyond building data marts and data warehouses, to revisit our business process improvement goals, and for CIOs to regain their seat at the table by not only keeping the enterprise up and running, but by helping government make better decisions through analytics. Whether it's No Child Left Behind or where to deploy our patrol cars, the public expects us to make informed decisions, and CIOs need access to the tools and data necessary to lead the revolution. It's time to answer Davenport's clarion call again.