DMV Probe of Computer Deal Released
An internal investigation into the Department of Motor Vehicles’ $44-million computer debacle has found that DMV officials were swayed by overly optimistic sales pitches from manufacturers, ultimately selecting a computer technology without properly evaluating whether it could handle the agency’s needs.
Compounding the problem was the decision by high-level DMV management to essentially relinquish its authority over the project once it was under way, according to a special report made public this week.
The management staff deferred to the agency’s technical staff and unquestioningly accepted their explanations for an array of problems that delayed the multimillion-dollar project, drove up the costs and eventually led to a decision to abandon it, the report said.
“A cultural condition has developed over time within the department, namely that technical projects . . . are owned, controlled and operated by the technical staff,” the report said. “No one was questioning how long the project was taking, how much it was costing, what were the benefits to be derived, (or) when it would be completed.”
Del Pierce, who headed the department when it launched the project and now serves as state lottery director, said he has not seen the DMV report and could not second-guess its findings.
The internal assessment of the computer project failure appears in a special report to the Office of Information Technology, the arm of the Department of Finance that oversees state computer projects.
Written in part to justify a decision to abandon the project, the report is highly critical of the way the new computer system was designed and managed by the department.
Glenn Wilson, manager of DMV’s information system, said the report is much more detailed than it normally would have been because “we wanted to close the financial books on the old project and start a new set of books.”
Although some of the department’s criticisms are similar to those contained in legislative analysts’ reports on the debacle, its assessment provides the first detailed examination of key internal decisions which DMV officials now say led to the project’s failure.
The Department of Motor Vehicles had embarked on the computer project in the mid-1980s, during the Administration of Gov. George Deukmejian. The prime goal was to provide a “one-stop shopping center for information” by merging into a single system the giant databases that provide driver’s license and motor vehicle registration information.
The major contracts for the project were awarded to Tandem Computers Inc., a computer hardware company based in Cupertino, and Ernst & Young, a software vendor. After one year, Ernst & Young left the project by mutual agreement with the department.
The project was envisioned to cost $28 million, but as problems mounted, the costs soared and it fell further and further behind schedule. In 1991, when Frank Zolin was appointed by Gov. Pete Wilson to head the DMV, he ordered a review of the project and an assessment of the best way to salvage it.
But when he was presented with rescue proposals that carried price tags of more than $100 million, Zolin told a legislative committee that he “decided to pull the plug on it.” At that point, $44 million had been spent.
The internal investigation cited two major factors that caused the failure: a critical technological decision that was made in the initial stages of the project and an upper-level management system that generally maintained a hands-off philosophy toward technical projects.
In 1987, when the project was being designed, the report said, the “latest, hottest technological advance” was something called relational technology. Believing that this technology would provide a much more flexible way to retrieve data, the department early on locked itself into using relational technology.
“It now appears that DMV management succumbed too readily to industry hype,” the report said. “The practicalities of what was being promoted by the technical industry was overly optimistic, which is often the case. Management should have taken a more cautious approach.”
Although relational technology had been used successfully in other states, there was little experience in using it in a system as large and complex as the DMV’s. The driver’s license database contained more than 30 million records; the vehicle registration more than 40 million. The system had to process at least 1 million transactions in a typical workday and remain in operation 24 hours a day without breakdown.
In deciding on the relational technology, the report said the department never properly considered whether it would work with its old system for retrieving data, which it did not want to change. As it turned out, the report said, that retrieval system did not lend itself easily to relational technology.
At the same time, the report said, the DMV grossly underestimated how much work it would take to rewrite programs for the new system and clean up data that was in the old system. The old system, for example, was cluttered with data collected for laws and regulations that no longer existed.
“The fundamental technical strategy developed for the project back in 1987, set (it) down an impractical path, which doomed the project,” the report concluded.
It said a steering committee, which might have questioned the direction of the project, was dominated by the technical staff and representatives of Tandem management. “The technical management was too closely involved in the success of the project to take a critical view of it,” the report said. Likewise, it said Tandem officials “were not about to make a hard-line assessment which might jeopardize continuation of the project original technical strategy.”
A spokesman for Tandem was unavailable for comment.
Although the DMV had to submit an annual report to the Office of Information Technology, that also was prepared by the technical staff. “Most any story presented by the project team seemed to be readily accepted,” the report said.
In its examination, the report focused only on the department’s actions and did not touch on any of the conflicts of interest that have been alleged by Assemblyman Richard Katz (D-Sylmar).
Katz has criticized Office of Information Technology Director Steve Kolodney and a former DMV project director, Dennis Walker for relationships they had with the Tandem Corp.
In late 1989, Kolodney left his job as head of the Office of Information Technology and became a partner in a consulting firm that later did business with Tandem. Kolodeny left the consulting firm after a year and went back to his old job at the Office of Information Technology, which had responsibility for monitoring the DMV project. Walker served as director of the project and then left DMV to to work for Tandem. He later returned to state service.
