Seeking the straightest path to remaking the Alaska DMV, Mark Boyer discovered that no good deed goes unpunished.
Sometimes mainframes are like kitchen cabinets: They're easier to reface than replace. A Web interface and a layer of middleware glue can yield big paybacks with a small investment. In this article you'll learn: Why new technology may not always be the answer How to reapply one problem's solution to other challenges How relationships can be as important as technology Mark Boyer began the New Year looking for a new job. After four years as commissioner of Alaska's Department of Administration, acting as executive sponsor for some of the most successful IS projects in the state's history, he shouldn't have to look too hard. He can put this on his résumé: Headed a US $300,000, six-month project to implement a Web and telephone interface that allows customers to renew automobile registrations without visiting the Department of Motor Vehicles (DMV). Cut the state's fulfilment cost from US $7.75 to 91 cents while simultaneously slashing the transaction time for the customer from two-and-a-half hours (not including travel time) to just under three minutes. Achieved a significant rise in customer and employee satisfaction at the DMV. Reapplied the process and technology to the state's business-license renewal department. Targeting other administrative niches and anticipating similar results.
If you are a potential employer who calls Boyer for an interview, he'll tell you that three things made those projects work: focusing on the customer, leveraging the department's current core technology in new electronic venues instead of throwing out the old system and developing a general technology solution that could work in other areas.
But he'll also sound a warning about the need to work hard at building relationships across departments, especially when you're the outsider coming in to fix someone else's problem. Boyer burned a few bridges in his zeal to improve the DMV's efficiency, and that's one of the reasons he's updating his résumé (see "On the Firing Line").
Love Your Customer
In the past, when you went to reregister your car at the Anchorage office of the DMV, the clerk would punch your data into a terminal and the database on the mainframe would spit back a record. Owner the same? Yes. Address the same? Yes. Hand over the fee and you'd get your registration tags. Based on a time-and-motion study in the spring of 1996, you'd end up wasting two-and-a-half hours in line for a five-minute transaction. It wasn't much better in Fairbanks, which clocked in at 90 minutes per visit. And this doesn't take into account the fact that fewer than half of the renewal offices (18 out of 44) are outside the major metropolitan areas of Juneau, Anchorage and Fairbanks. In remote parts of the largest state in the union, visiting the nearest office to renew auto registration might mean travelling hundreds of miles.
The citizens of Alaska had two problems with the DMV: time and distance.
The DMV faced a time problem itself: Its employees had to spend lots of time handling renewals, at the expense of some other tasks that actually needed a human hand. At US $4.50, the cost of processing a mail-in renewal was lower than the US $7.75 it took to serve the customer in person, but the six- to eight-week turnaround time to get the registration tags back to the vehicle owner was nothing to brag about. Multiply those expenses by the 220,000 Alaskans who need to renew their automobile registration each year and it was pretty clear where the state was spending a multimillion-dollar chunk of its tax revenues.
The citizens of Alaska didn't know those cost figures and didn't much care.
What raised their ire was the wait they had to endure to perform a relatively simple task. So loud was this grumbling that Gov. Tony Knowles slipped in a promise to fix the DMV during his 1995 State of the State speech.
The responsibility fell to Boyer. The commissioner of the Department of Administration is an appointed position, but Boyer isn't some know-nothing political crony. As the former general manager of a Fairbanks utility and former city manager for Fairbanks, he knew the adage "the personal is political" was true-that is, that the best politics helped citizens right where they can see it. Something as simple as erasing the lines at the DMV would make a big impact statewide. Unfortunately, the situation at the DMV was far from simple.
"The DMV didn't have any business plan whatsoever," Boyer told CIO from his Juneau office before he resigned. "They were just into survival." So, despite the spotlight the governor shone on the DMV, the problem wasn't being fixed.
And the automobile owners weren't the only ones upset at the situation. "DMV employees were demoralised," Boyer says. "They had counselling sessions on violence in the workplace and dealing with difficult customers." Plus, as the ugly stepchild of the public safety department, the DMV got very few budget dollars. "Cops put all their money into guns, cars and jails," Boyer explains.
The DMV was using old desktop computers backed by an MVS-based mainframe holding the license and auto registration databases.
The easy answer to the DMV's problems was to throw out the mainframe and reengineer the whole DMV. But that wasn't necessarily the right answer. In fact, according to George Lindamood, vice president and program director in GartnerGroup Inc.'s IT Executive Program and former CIO of the state of Washington, it rarely is. A veteran of a failed DMV conversion himself (see "To Hell and Back," CIO Section 1, Dec. 1, 1998), Lindamood says that DMV projects that overhaul the mainframe are mistakes for two reasons. First, you wind up with something that's obsolete as soon as it's complete and second, these types of projects tend to run out of control. Several states, such as Washington, Oregon and California, took the overhaul route (see "When Bad Things Happen To Good Projects," CIO Section 1, Oct. 15, 1997) and found themselves fighting spiralling budgets and paltry results.
At Alaska's DMV, the existing technology worked, Boyer emphasises; it was the long lines that were unacceptable. His analysis of the situation got him pointed in the right direction. It was a customer service problem, not a computer problem. Getting the computer to spit out a database record two seconds faster wasn't going to shorten the lines. Keeping people away from the DMV offices was. What if, Boyer thought, we put in a Web front end and let people renew over the Internet? Gartner's Lindamood likes that approach. "We have to learn, first of all, what is possible with the Internet and what the public really wants," he says.
"Not everyone has Internet access, but the growth is there." In the face of this growth, Lindamood says, organisations can incorporate the Internet into their IT projects and give employees experience that anticipates that growth. A small project like Alaska's served that purpose.
Clean Up in a Jiffy
Boiled down, the DMV's online renewal system is basically a new way to pay for car registration. The Web interface and the interactive voice response (IVR) phone system that was added to serve citizens without Internet access are no more than credit card terminals tacked on to the existing back end. What the state of Alaska did not do was build a new application; rather, it put a shiny new face on an old one. "We made the decision early on to leverage the substantial fixed costs in the mainframe environment," Boyer says.
Another key to Boyer's success was picking those two processes to start with. There are enough of these routine types of transactions to make a significant dent in the line at the DMV. Last November only about 9 percent of renewals and vanity plate requests occurred online or over the telephone. But with a little judicious advertising, the usage rose. The figure jumped to 12 percent after the state ran a series of brief television commercials over three weeks later that fall. (Boyer would like to see an all-out campaign to promote the program, and he's already devised the commercial: It shows happy people having fun in America's last frontier while a voice-over intones, "We know you have more important things to do than stand in line.") The only facet of the project that was a problem was-no surprise here-a budget. That didn't bother Boyer. "You gamble on the savings accruing to you." He bet that if he could get the online transactions running by December, halfway through the state's July-to-July fiscal year, the savings in the latter half of the year would fund the project. He won the bet. His crew started work in October 1997. A test system went online Dec. 15 of that year. The next month, the project went live. The US $300,000 Boyer spent to bring the system online was paid back within four months. Not only that, but he saved enough money to give DMV employees a raise.
The state of Alaska has yet to repeat the time-and-motion studies that measured the horrific waits in 1996, so no hard numbers exist to document the line shrinkage at the DMV. But Boyer says that he not only sees a difference at the Anchorage office, he's been able to reassign personnel to other duties, such as administering driving tests.
Those DMV lines could get even shorter if and when the next stage of the renewal project gets finished. Boyer says the same technology can be used to renew driver's licenses and to schedule a time to take a driving test. With Boyer gone, though, will Alaska's state government continue its efforts to improve its customer service via the Web and IVR? Absolutely, says Dennis Hoffman, a data processing manager for the Department of Administration who believes in what Boyer started. The technology integrator for the online renewal project, Hoffman developed the credit card authorisation scheme underpinning the online renewal process, and he made sure that it was a generic solution.
"The e-commerce piece can be used by anyone in the state government," Hoffman says. It consists of a firewall-protected server that takes any request for a credit card payment, dials out to a verification vendor and passes back the transaction results to the requestor. Nothing fancy; nothing proprietary.
Another reason Alaska will get extra mileage out of the new technology: Hoffman makes an effort to spread it among his programming colleagues. "I had the idea to start a mentoring project to bring together the best programmers from other agencies to do a week-long 'Webette' project-to build a Web front end to an existing legacy database system," he says. It was a proof-of-concept program attended by people from the DMV, the student loan office, the natural resources department and other government offices. Everyone had a different database and a different query in mind, but the basic principles were the same.
The DMV wanted to know if a vanity plate was available, while natural resources wanted to know if a particular forest-service cabin was open for rent.
"Everyone walked away knowing that it didn't matter where the data lived," Hoffman says.
Hoffman's guerrilla marketing efforts among the programmers in the government seems to be paying off. The Division of Occupational Licensing has built a Web-based business license application and renewal interface built on the credit card authorisation technology developed for the auto reregistration process.
Make no mistake, Hoffman has learned from Boyer's mistakes. He's careful when he goes to talk to the department and division managers, the future executive sponsors within state government. "My approach usually is to go in and talk to someone about their business," Hoffman says, "not technology. When it comes to business managers, it's a matter of ensuring them that there's a technology solution that fits within their business goals." If Hoffman needs proof to convince business managers that it's better to reface the mainframe than replace it, he can point to what Boyer engineered with the DMV's online auto registration renewal interface. From day one, Boyer approached it as a business problem. "We didn't come at this as a technology project," he says. "We focused on the customer and worked backward." On the Firing Line Why partnering may be more prudent than pushing in the long run Why has Mark Boyer resigned from his position in Alaska's department of administration? He was too eager. At one point during the DMV project he offended some of his colleagues at other state agencies simply by offering to help upgrade their systems. He now feels that his eagerness compromised his effectiveness to reengineer. "How do you fix something that you don't own and don't control?" Boyer asks. Although the department of administration is responsible for many bureaucratic functions, from payroll to the public defender's office, when Boyer started the job the department did not own the DMV.
So when Boyer set out to fix the DMV's problems, he had two choices. He could partner with the cops in public safety, supplying the expertise they lacked in computer systems and project management, or he could take over the department and, as boss, implement the changes he wanted. At the time, a takeover seemed the logical step. He had everything else he needed to build the application-the programmers, the finance department, the computer equipment-except authority over the DMV. In the spring of 1997, Boyer convinced the legislature to hand over the DMV to his department. "How I sold the legislature was by telling them, 'I've got all the levers you need to pull to make the change,'" Boyer says.
With the clarity of hindsight, Boyer sees that he should have chosen the partnering option over the takeover option. "If I had been a better internal politician," he says, "I would have been more smooth and handled it differently. Instead it was a hostile takeover. It ruffled lots of feathers." As a consequence, he had difficulty reapplying the simple Web- and phone-interface solution in other areas. "We wanted to fully leverage [online access] into things like fish and game [licenses] and online filing of financial disclosure information," he says, but "I couldn't sell more of this.
People see me as a takeover person, because that's what I did." Boyer acknowledges that the public safety department was part of the DMV's problem. "The cops weren't interested [in fixing the long wait times]," he says. "They're not business-focused, they're not technology-focused, they're not customer-focused." But the wiser path, he still believes, would have been to work with them. "I should have [been more] low-key and spent two years selling it to the public safety people," he says. And that's a lesson he thinks is important for every CIO. "It's not any different in the private sector," Boyer says. "You try to avoid making scars." Amy Helen Johnson covers technology from the Silicon Forest in Washington. She can be reached at firstname.lastname@example.org.
Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.