Automating Classification

Creating a job description for a new employee is, in theory, a simple enough process for a federal manager: Identify the duties that need to be performed and the qualifications applicants need, then write up a job description. Fitting that description into the federal job classification system is much trickier, however.

A manager must go back and forth with classification specialists to determine what grade the new position fits into, comparing it with the almost 500 documents that make up the government's classification standards. Reaching agreement can take months--one manager told Office of Personnel Management classification specialist Helene Rosenheim it took 180 days to go from identifying a position to actually seeking applicants for the job.

Armed with similar stories from throughout the bureaucracy, Rosenheim and an interagency working group set about to automate the classification process. Rosenheim's efforts are taking shape as a software product that aims to reduce the time managers spend on classification to a matter of hours.

The Federal Automated Classification Expert System, or FACES, will allow managers to compare job descriptions they write to a database of positions from across government. The database will include information on the grade level of comparable positions, allowing managers and classification specialists to bypass the reams of paper-based standards they now must sort through. Rosenheim predicts that FACES will eventually be available on the Internet, providing managers with a governmentwide up-to-date database.

The biggest advantage, Rosenheim says, will be the time saved.

"Classification is a very analytical process," she says. FACES will save managers time by putting the analysis behind the scenes.

FACES will be available within two years, Rosenheim predicts.

Rosenheim is presenting the classification automation project at the Reinvention Revolution Conference. Check the Daily Fed this week for more stories of frontline reinventors like Rosenheim.