Faculty Blast $112M Computer Systems

Project ADAPT was supposed to make everyone's life simpler.

But nine months after it replaced the University's financial computer systems--and the way staff use them--administrators are angry, accounts go unpaid and everyone is frustrated.

In the post-ADAPT world, nothing is simple.

ADAPT, the effort to centralize Harvard's aging business databases, has been in the works for more than five years. It is now projected to cost $112 million.

When ADAPT's first phase went online July 1, it instantaneously and dramatically transformed jobs across the University. At last count, 4,193 people in every laboratory and every educational department at Harvard use the system.

ADAPT is about more than just computers. It is about the way Harvard's administration works--or does not. The implementation has been plagued with problems.

Many users say the new financial systems are painfully slow. They say it creates more work. They say bills don't get paid, and professors don't get reimbursed.

Officials from Harvard's 10 faculties, most vocally in the Faculty of Arts and Sciences (FAS), have butted heads repeatedly with the University administrators who oversee the system.

Now, as central administrators struggle to fix the system, they must work with the faculties to mend broken lines of communication. And they must convince faculty and staff that the $112 million being spent on Project ADAPT is not in vain.

The Trouble With ADAPT

Users of the new system almost universally agree it is flawed.

Steven C. Wofsy, Rotch professor of atmospheric and environmental science, calls it "user-hostile."

"It has been very difficult to get the required data in electronic form," he writes in an e-mail message. "We have a large number of grants to manage, for which we have our own system to merge data on salaries, travel, other expenses, and on income."

The more than 40 interviews conducted for this coverage show that Wofsy is not the only one complaining.

In one faculty, a senior administrator got so fed up that he circulated a memo proposing to send a dozen roses a day to central administrators until the systems were finally fixed.

Users complain that the new systems force them to regularly use dozens of 33-digit financial codes. The same codes used to be only 14 digits long.

They also say that with so many users accessing the system at once, it has often been slow and it sometimes crashes.

The new System for Travel and Reimbursement (STAR), which business travelers must use if they want their money back, is a common sore spot. Travelers, or their assistants, must answer dozens of sometimes unrelated questions just to log a simple expense.

STAR was so problematic that it is now being completely rebuilt--from scratch.

While users run the gamut in enthusiasm for the new systems, some schools, like FAS, have been harder hit and are more vocal in their complaints.

"Of over 100 departments, I would be very surprised if there are 10 that are having an easy time," said Geoffrey M. Peters, FAS associate dean of administrative resources, following a presentation to the Faculty Council last fall.

The Faculty Council devoted several meetings over the fall and winter to discuss what is wrong with the system. FAS even hired its own consultant to independently confirm its assessment.

But in his annual February letter to the Faculty, Dean of the Faculty Jeremy R. Knowles reiterated FAS' commitment to making the project work.

"The rewards will, I trust, be worth the patience that will surely be demanded of us all, as bugs are eliminated and as we master the new systems," Knowles wrote.

A Project To Please Everyone

When Harvard administrators first considered replacing the University's aging financial databases in 1993, everyone thought the project was a good idea.

Based on designs from the 1940s, the systems then in place were old and unwieldy, unreliable and uncoordinated.

They were "teetering," says Provost Harvey V. Fineberg '67.

They were "tottering," says Assistant Provost Daniel D. Moriarty.

So Harvard administrators dreamed up a solution they thought would catapult Harvard into the information age.

Over three years, the Administrative Data Project, as it was then called, consulted with more than 400 members of Harvard's faculties and staff in a review of administrative systems. When then-Provost Albert Carnesale formally announced the newly christened Project ADAPT in April 1996, he called for the faculties to join him and the central administration in developing the system.

"[This project] will require cooperation and adjustment throughout our community," Carnesale then said. "[It] is not just about technology, it's about the way we conduct the business of the University."

And officials from the faculties lined up to provide that cooperation.

From FAS, Knowles provided early praise for the project.

"A major upgrade of our systems is more than overdue," he said then.

From Harvard Medical School, Associate Dean for Finance Cynthia Walker anticipated the project would result in simpler and faster administration.

"Project ADAPT will give us back the time we need," she said. "Our faculty are as eager for these improvements as are our administrative staff."

With everyone at Harvard on board for the ride and promising a highly consultative process, it seemed that ADAPT would be smooth sailing.

Revolving Doors and Sinking Ships

But ADAPT staff turnover and a push to get the system online cut down on consultation between University officials and the faculties that would use ADAPT programs.

Since 1996, the leadership of Project ADAPT has been in constant flux.

Carnesale's pick to lead the project, Deb Mauger, left in 1997. Her successor departed this past summer. Carnesale himself was replaced by Provost Fineberg in 1997.

Today, University Vice President for Finance Elizabeth C. "Beppie" Huidekoper, Fineberg and Moriarty, who serves as the University's chief information officer, control the financial systems that ADAPT built.

With the sole exception of President Neil L. Rudenstine, there remains no senior administrator who has overseen ADAPT since its inception.

Communication between project leaders and the faculties was at times tenuous. A rotating leadership did nothing to help.

According to senior faculty officials involved in the project, a number of advisory committees stopped meeting early in 1998.

"I think during that year and a half [before the launch] it is fair to say that the people from the faculties--both the end users and school financial offices--were listened to less," said a senior financial administrator from one of the faculties. "The project changed to be viewed much more as a central project. I think it started out being viewed much more as a University project."

Yet this was the same time span in which some of the project's most crucial decisions were made, including which systems would go online in July.

A Premature Launch?

But central administrators say they involved the faculties along the way.

For instance, faculty officials signed off on so-called "requirements documents," which set out many of the details which are now sources of irritation.

Some of the same faculties that bemoan the problems were among those who agreed to the "requirement documents."

But Moriarty says he is not surprised that faculty officials don't see these documents as signs that they were fully informed.

Moriarty should know. He was the chief information officer at the medical school during the development process.

He acknowledges that it can be difficult to see what is missing from a proposed piece of software when looking at hypothetical specifications on a piece of paper.

"Generically asking for requirements doesn't do it," he says.

While officials in all capacities now say now they wish consultations had gone differently, most agree that the people who would actually use the systems--the "end-users"--did not have enough input.

"It was unbelievably participatory in certain components of the community and not participatory enough in others," Moriarty says. "The amount of end-user involvement in a number of cases was clearly inadequate."

Many users say they wish they had been better consulted especially in the months leading up to July's launch date.

They say that had such conversations taken place, they would have pushed the central administration to hold off on launching the system until all the kinks were ironed out.

Wofsy says more users should have had more chances to evaluate the systems before they were put into use.

"There should have been a staged

cycle of develop-test...and test again," he writes in an e-mail message.

But the administrators running the project say that the date was carefully considered and absolutely necessary.

According to Huidekoper, the University needed the new system before the New Year to be Y2K compliant. July 1 was picked to coincide with the beginning of the University's fiscal year.

Moriarty adds the system was ready to go and was desperately needed. He says they knew a bumpy start was inevitable regardless of when the project was launched.

"There will always be a trade-off between fully addressing the local requirements across the University and deciding to go ahead," he says.

Moriarty likens a major systems implementation to a large building project.

"There aren't many happy users when you renovate a library and the library is kept open," he says. "These are tough, difficult projects everywhere."

Moriarty says it is only reasonable to assume improvements will have to be made on a system as massive as ADAPT.

Fineberg says the University was forced to put some user requests on hold in order to get the system running.

"We made purposeful design compromises taken from endless discussion," he says.

The Next Steps

The problem with ADAPT is larger than flawed bits and bytes--larger, even, than its ballooning budget.

Those involved say it has been the story of what happens when the people who run Harvard do not listen to each other.

Users say to regain their confidence, the project leaders must make this system work and convince them they will have more input in the future.

Project leaders say they are now in fact taking concrete steps to fix the financial system.

Moriarty says that the problem is double-sided: users want solutions that work, but they want them fast. Sometimes those goals just do not mesh.

But Huidekoper, who is now responsible for the systems in place, says her team recognizes the problems and are working to address them.

Starting March 28, she and her team will also be meeting with about 150 users in FAS over a series of brown-bag lunches to get input on what needs to be done, and in what order.

"We need the input from FAS," Huidekoper says. "I can't stress enough that we're going to try and get out there in front of users and get their input-- but they've got to come to the meetings too."

Three hundred fifty people in the

financial administration are currently devoted to fixing the problems.

Huidekoper is spearheading efforts to get financial deans to talk about the problems and solutions they develop at their own schools.

But discontent remains, and Huidekoper says she is concerned.

"I'm most concerned about when morale reaches such a low level that people cease to believe in our ability to get things done. It's a real human

resources issue."

But as Huidekoper tries to defuse tension, some say they are nervous because the ADAPT team has begun work on a second phase of the project that will revamp human resource (HR) information systems. A small group is currently assessing whether to proceed and will at some point produce a formal report.

Many in the faculties say they fear a new flawed system while they are still mastering the first phase.

Faculty members say central administrators could restore confidence in the project by promising that the human resources phase will not get the go-ahead without the full backing of the faculties.

Fineberg says he is hesitant to afford the faculties formal control.

In such a large University, he says, it is virtually impossible to get everyone's support on a project of this size. He says he does support giving system users input in project decisions.

But an outright vote over the future of the project? Fineberg calls this "bad management."

Fineberg says that when the discussions and consultations are over, the decision to move ahead on HR must rest with Rudenstine alone--just as did the decision to go live with the financial system in July.

For Rudenstine's part, he says he will consult with the faculties before moving ahead--but he makes no promise to give them a final vote.

"The process for making decisions about University-wide information systems--including specifically any new systems--is broadly consultative, across all the Faculties and other units," he said yesterday.

"Before any decisions were made, the President, Provost and Deans would discuss the proposals again, taking into account all the issues raised in the course of the entire review process," he said.

Huidekoper says users should not worry about the next phase yet. No one has even officially recommended that the University move ahead with HR. In fact, according to central administrators, the earliest HR would be implemented would be 2002.

But members of FAS say the timetable for HR is a mystery to them: another example of poor communication.