This article was originally published in the Fall 2012 edition of OnAnalytics, published by the Institute for Business Analytics at Indiana University’s Kelley School of Business.
This article focuses on insights from Tom Preston, who at the time of the article’s publishing, led Booz Allen’s Enterprise Resource Planning (ERP) Advisory Services.
Business analytics has enormous potential to increase efficiencies and improve outcomes in government operations. The federal government alone has a budget of $2 trillion and annual expenditures of approximately $3 trillion; the Department of Defense, even before war supplementals, allocates some $520 billion each year. Federal agencies are always looking for ways to save money by finding more efficient ways to operate. For example, instead of relying on the historical cost of a product or capability, the government is leveraging analytics to model and determine what something should cost and then using real-time dashboards and visualization techniques to assess their performance against industry best practices. In addition to data on these expenses, sensors and radio frequency identification (RFID) are adding date/ time/location stamps to everything from shipped packages to key components of vehicles and aircraft. These many tracking systems offer rich material for business analytics, and government departments are beginning to implement such strategies as:
Stochastic simulations and optimizations to approach what-if scenarios, from assessing financial stimulus and money supply/interest rate impacts on the US economy to modeling the recruiting efficiency of the Armed Forces in meeting future military needs.
Visualization tools that graphically demonstrate relationships between program expenditures and mission outcomes. For example, illustrating how a vaccine program is effectively protecting a city, county, or state by graphically depicting vaccine spending and flu cases with contrasting images depicting the amount of spending by location overlaid with population density. Text data analysis to incorporate public feedback via social media – for example, information on how citizens feel about the vaccine program as well as their ideas for improving disease prevention.
Modeling multi-dimensional problems
The challenge, however, is that the overwhelming amount of information can sometimes paralyze departments instead of guiding them. The complexity of public sector issues and structures makes every problem multi-dimensional, with each decision carrying multiple implications for a variety of programs and stakeholders. Modeling these problems in their entirety can seem impossibly daunting.
Because of this complexity and the massive amount of data these systems generate, the temptation is to fall back on the “salami cut” strategy of making a uniform percentage budget reduction across all departments. This approach has the political advantage of appearing not to favor any one area over another. Unfortunately, it’s almost always a mistake, because it addresses only the cost and not the impact of the various departments and programs. A better strategy would be for agencies to identify ways to reallocate funding and share workload across departments, deploying a sharedservices model to minimize program impacts. In order to do this, however, the various stakeholders must openly share information and perform the multi-dimensional analysis required to find effective solutions.
Business analytics, when utilized to their full potential, can not only reduce waste and conserve resources, but also strengthen an organization’s ability to achieve its mission. The greatest opportunity that business analytics presents in the public sector is the potential to uncover powerful levers for positive change. Several new trends are enhancing this potential such as visualization tools, mobile devices, and crowd sourcing.
The 21st-century business analytics toolkit
The newest visualization software takes advantage of large-scale memory computing to graphically represent large volumes of data. These tools are designed to demonstrate relationships and comparisons in an intuitive manner, so they can be understood without the need for extensive training in analytics. It’s important to keep in mind, however, that additional methods and techniques are required to develop actionable recommendations. For example, visualization software may show you that a certain area of the country has a higher-than normal incidence of traffic delays for truck shipments. This recognition can motivate action, but the graphic won’t tell you what factors are contributing to the problem, or what the widespread and long-term results would be after implementing different solutions.
With respect to mobile devices, their ubiquity creates potentially limitless opportunities for both data collection and analysis. Wherever people go, analytics can go, so long as software development is focused on preserving computational capability both on- and offline. The apps we leverage for analysis, which will include everything from Microsoft Excel to the most advanced business intelligence tools, must provide realtime analytic capabilities and the ability to access large information stores via parallel processing and in-memory databases. These apps could assist with everything from mapping sewer lines to measuring ambient noise levels, but must be designed to function in areas without cellular service, with the capability to sync back to databases when the user goes online.
Crowd sourcing also promises a wealth of actionable information. Because it captures multiple individual viewpoints, this type of data can be a particularly informative resource. For example, I am a frequent business traveler, and I rarely go to a restaurant that does not earn positive visitor reviews. Public agencies should also find ways to leverage social media to identify and address problems, which could include everything from reporting potholes to suggesting new services. In addition to gathering feedback from people in their service area, crowd sourcing can be a means of tapping into the expertise of researchers and professionals who are eager to help but don’t always have a conduit for their ideas. These tools can also contribute to constituent buy-in, so long as the organization actively responds to and makes use of this feedback.
Looking forward, these and other analytic tools offer promising opportunities for the public sector to address the problem of “doing more with less.” In order to fully leverage these capabilities, however, decision makers will need to first of all invest in understanding business analytics enough to trust its conclusions. Secondly, the link must be established between using visualization tools to identify problem areas and following through by implementing the complementary analytic techniques and tools that can deliver solutions. Finally, departments and organizations can start to explore new technologies in the form of mobile apps and crowd sourcing, which will enable them to get more information and innovative solutions directly from the people and areas they serve.