Lessons Learned

When Ohio's workforce system caused more headaches then it cured, the state had to face some harsh realities.

by / April 16, 2002
Embracing change is not without risk. Few agencies know this better than the Ohio Department of Job and Family Services (ODJFS).

In 1999, the agency replaced an older, Unisys-based job matching system called Ohio Job Net with a newer system called Ohio Works. Ohio Job Net had been very popular among users, but the job matching system had begun to show its age. Job Net did not comply with the Workforce Investment Act (WIA), which requires states to provide one-stop employment and training services. The time was ripe for a change.

The state went to work developing Ohio Works. The Ohio Works system was written in a proprietary language called ColdFusion, and was designed to allow job seekers to post resumes, employers to post vacancies and all users to access data on salaries and hiring trends.

But Ohio Works was fraught with problems, and the problems released a flurry of accusations and negative publicity. Critics charged that the state did not provide proper oversight of project partner Accenture, that there was a lack of records on employees assigned to the project and rubber-stamp approval of invoices without review of the work.

The Ohio Works system was prone to technical problems from the beginning. Job seekers and employers reportedly found the system complex and difficult to use. They complained of being disconnected from the system after signing in, being required to provide excessive personal data and being unable to browse job listings. The internal segment of Ohio Works, known as ServiceLink, was prone to losing data.

These are the problems Tom Hayes inherited when he took over as director of the newly created ODJFS.

Cleaning Up The Mess

Hayes initiated a competitive bid process to find a consulting firm to analyze the state of Ohio Works. The project was awarded to Compuware Corp. of Columbus, Ohio. Compuware studied the program, and presented the state with three options.

First, the state could continue with the existing Ohio Works system while attempting to modify and stabilize the underlying code. Compuware rejected this option due to the inherent instability and uncertain future of ColdFusion. The company stated in an October 2001 report to the ODJFS: "Not only would ODJFS be building a key enterprise level system on a niche product, but that niche product seems to be in a state of flux." The projected cost of this option was $25.4 million.

Second, ODJFS could rewrite the system to work with the department's existing IBM Websphere/Oracle/Unix computer platform. This option would require Ohio Works to be rebuilt from the ground up, but Compuware anticipated that the rewrite would result in a product that was more stable, more robust and longer lasting. Total anticipated cost to the state: $26.6 million.

Third, the department could undertake a two-phase, hybrid approach that would attempt to stabilize the existing system and then migrate it to the department standard platform. This final option would yield some cost savings in analysis and design required, but the savings would be quickly eaten by the need to fix the problems inherent in early versions of Ohio Works. This hybrid solution would cost the state $36.6 million.

After quickly rejecting the first alternative, Compuware was left with the final two options. Since the hybrid option would cost an estimated 30 percent more than the second alternative, Compuware recommended the second option. Hayes supported the recommendation, identifying ways that a new system could be constructed. In a report to state Sen. Richard Finan, Hayes suggested that the department explore options such as constructing an all-new system, modeling a system on one existing in the public domain, or using a combination of commercial tools and search engines. Ohio agreed, and the rewrite was under way.

In the meantime, the popular Ohio Job Net has been resurrected and pressed into service to fill in the gaps in service. The old system seems to be functioning satisfactorily. Hayes estimates that it is receiving a half million hits per month. "[We wanted to] give something back that people actually like," Hayes said of the decision.

The project will not be without its costs. The original Ohio Works endeavor cost an estimated $60 million, and the anticipated $26.6 million cost of the rewrite will hit the state just as Ohio experiences a statewide $1.5 billion budget shortfall in a recessionary economy.

But Ohio officials are putting the best face on the problems. "It was a pretty cut and dried business decision," Hayes said of the move to rewrite Ohio Works. In the meantime, the state plans to find a proven model of success to follow for their next attempt. "We're looking at existing systems [that are] WIA compliant in other states," Hayes said. Ohio is examining systems in use by Pennsylvania, Illinois and other states. By adopting all or part of a public domain system that has already been tested by real-world use, Ohio will minimize the risk of failure of a project that has already seen more than its share. The department reviewed job-matching systems used by other states through March, and then hoped to begin importing the best model for use with the Ohio Works system.

Pushing Forward

Although Ohio Works was intended to be a model of Gov. Bob Taft's desire to put job seekers "online, not in line," he is taking the crisis in stride. "When you see something not working, you have to make a decision about what is best," said Joe Andrews, spokesperson for Taft.

This time, ODJFS has some clear goals to measure the success of the state's future job-matching system. "[We want to] make sure it works, is easy to use, and is a rewarding experience," Hayes said.

Jennifer Patterson Lorenzetti is a freelance technology writer in Centerville, Ohio.