Government agencies should start small and build an "analytics culture" to ensure success when using data to enhance performance, according to a Nov. 30 study (.pdf) from the IBM Center for the Business of Government and the Partnership for Public Service.

IBM, coincidentally, is heavily engaged in the selling of analytic tools to state, local and federal agencies.

The report also mentions myths that impede the analytics culture, such as the notion that analytics leadership must start at the top. An analytics culture sometimes starts when a component in a department takes the lead, report authors note, before the culture spreads throughout the agency.

For example, the Federal Aviation Administration's Safety Management System began in the FAA's Air Traffic Organization. The SMS uses algorithms to point out safety hazards. When ATO employees note a hazard, the information becomes available across the agency's computer systems and can be shared with an airline or airport, the report says.

Another myth noted in the report is that everything needs to be measured. In fact, having fewer measures, but ones that are more useful, helps agencies reach objectives more efficiently, the authors say.

Estelle Richman, acting deputy secretary of the Department of Housing and Urban Development, agreed that what an agency does with data matters more than collecting a lot of it. The report says that after HUD analyzed its program to combat veterans' homelessness, it decided to measure success by the number of veterans who use housing vouchers instead of the number of vouchers the agency issued.

HUD now works to identify "bottlenecks" where homeless veterans who have received vouchers fail to successfully use them to find housing, and HUD employees can then work to fix particular obstacles.

For more:
- go to the IBM report (.pdf)

Related Articles:
Participation lacking from open government plans
Study offers 10 tips for agency telework policies
Report: Government can, should collaborate across boundaries