Despite Texas' well-deserved reputation as a hotbed for high technology, the image inside state government in Austin was once a bit more mixed. For a long time, the Texas Legislature was known for having a "healthy skepticism" about computers, and keeping a close eye on IT spending to corral unbridled enthusiasm about using high-tech solutions in government operations.
Not surprisingly, the job of state CIO is hardly a perk. The position has little authority to drive statewide standards and infrastructure in a state known for being highly decentralized. Yet Carolyn Purcell, who held the top IT position in the state for nine years, managed to carry out a number of significant initiatives that helped Texas evolve and adapt to the IT world for the better. During her tenure, Purcell helped launch the state's Web portal, TexasOnline; oversaw the data center consolidation; nurtured a volume purchasing program for IT products and services that has become nationally renowned; and made significant improvements in IT project performance, to name just a few of her achievements.
Three years ago, Texas Monthly magazine named Purcell among "The Most Powerful Texans in High Tech." Last year, Governing magazine chose Purcell as one of its "Public Officials of the Year," and in 2002, the Progressive Policy Institute ranked Texas third among state governments that use digital technologies.
Because she lacked the power many other public-sector CIOs have to get their jobs done, Purcell has relied on her skills and her talents for collaborating and building relations with disparate stakeholders to achieve her goals. Former Georgia CIO Larry Singer, who is now an executive with Sun Microsystems, called Purcell the "embodiment of class, the model for exercising leadership without clear authoritative power. Purcell has helped transform IT in Texas."
Purcell's reputation as a public-sector IT leader extends beyond Texas' borders, and has been nationally recognized through her work with the National Association of State Chief Information Officers and as the former chair of the National Electronic Commerce Coordinating Council.
Shortly after resigning from her position in state government last year, Purcell wrote about her experiences as a CIO and what she learned.
In 1994, the Texas State Legislature focused on deregulating telecommunications. At the same time, the Department of Information Resources (DIR) was looking for its second CEO, who also serves as CIO for the state. All kinds of people applied -- circus performers, gas station attendants, business executives down on their luck -- and me.
I was a 20-year veteran of public- and private-sector IT at that time, with a fresh M.B.A. from the University of Texas at Austin's executive program and a reputation for collaboration and cooperation. At the time, I was the deputy commissioner for IT at the Health and Human Services Commission, an umbrella agency for a dozen health and human services agencies, where my job was to rationalize the agencies' IT departments. The satisfaction I received from that work encouraged me to seek such a job at the enterprise level.
My intention was not only to accomplish the agency's mission of rationalizing IT expenditures, but to moderate the strategy from one of pure control to one balanced by promoting IT opportunities to strengthen government. I wanted to systemically improve IT development and operational performance, and streamline the agency by jettisoning what we didn't need.
I got the job.
The ABC of IT
The past nine years as CIO of Texas -- the second largest state in the union in size and population -- have reaffirmed my opinion that there are three important components for deploying IT that add value to an organization:
I can best illustrate the value of these components by telling three tales that involve technology, money, state politicians, IT managers and myself.
Alignment: Keeping an Eye on Legislative Goals
To advise the Legislature on funding, the DIR collected and analyzed project proposals from state agencies and presented the data, which included alternatives and expected benefits. Over time, the data was refined to permit an enterprise view of expenditures -- all available and searchable online. The supplier community loved it because they could easily find out where the market was. We loved it because we could compare various ratios (e.g. agency IT budgets as a percent of revenue) to similar industries in the private sector. The DIR assigned in-house analysts to each of the agencies to monitor plans and performance, and validate budget information
But I didn't keep my eye on the ball.
Yes, we were making progress, but what was the Legislature getting out of all of this? As in many other states, Texas has a sunset clause in nearly every agency's authorizing legislation. Agencies are rarely discontinued, but generally some changes are made. In 1997, the DIR went through such a review. Questions arose about the appropriateness of legislators serving on the board of an executive branch agency. The DIR's authorizing statute was changed to eliminate legislators from the board.
The presence of legislative leaders like Rep. Scott Hochberg of Houston and Sen. Ken Armbrister of Victoria kept our operations informed from a legislative perspective. When they were removed, I lost sight of the fact that the Legislature had created the DIR to scrutinize IT proposals and give them an opinionated thumbs up or down. Of course, it was their prerogative to override that opinion, but they wanted to focus their attention on policy decisions, not on whether agency X should be funded to buy 13 PCs. So we were telling them if a project had a chance of success, but not if it was the right project. They wanted both.
Over time, general competence improved, and most agencies knew what was required to convince the DIR they could do the project right. By 1999, there were very few projects that failed our review. We should have raised the bar for that legislative session, but we didn't. We should have used enterprisewide capital budget evaluations to select the right projects, but we didn't. We required ROI analysis, but the analysis was not rigorous. The budget committees spent a lot of time figuring out which IT projects to fund rather than discussing policy.
As a result, the Legislature transferred our budget review responsibilities to the Legislative Budget Board. Today, tools are available that would have facilitated better decisions on our part about enterprise projects and relative value. IT portfolio management has matured and is used by many successful companies. It requires discipline and an enterprise view, but it can help identify opportunities for consortia projects and greatly enhance the simplicity of presentations to executive management on competing projects. I regret that we lost the opportunity to get the budget stuff right.
Lesson Learned: Closer Relationships Can Help
The DIR board of directors was unhappy with me because the budget oversight role was lost. My chairman, Harry Richardson, correctly diagnosed the problem as a lack of alignment with the Legislature's goals. He challenged me to develop a closer relationship with legislators to improve the DIR's understanding of what was expected, and ensure elected officials knew what we were doing and contributing.
I developed a plan for communicating regularly with elected officials to see how we could help them and report on our progress. If the DIR had a program that helped state government agencies, we tried to extend it to local governments and school districts, and we made sure legislators knew about it. I told them about our volume purchase program, and how much money was saved by local government and school districts in their individual districts. With then-Gov. George W. Bush's help, we created Y2K remediation tools for local governments. The House Y2K oversight committee chaired by Rep. Jim Pitts of Waxahachie was indispensable to our success.
In 1999, we had a slim authority to create the state's Web portal thanks to SB 974 by Sen. Eliot Shapleigh of El Paso. I repeatedly visited many elected officials during 2000 to ensure they were informed about the initiative and understood the benefits, including the opportunity for local governments to connect.
It paid off. When the Legislature met in 2001, we got more specific authority, including a requirement that agencies use the TexasOnline infrastructure for online service delivery.
Bearing: Travails of the Data Center
In 1993, the Legislature authorized creation of a statewide data center. The legislation (part of an appropriations bill) did not provide for a comprehensive look at all agencies to determine the value of such an enterprise. The case for consolidation should be based on economies of scale and world class service; the mandate I was confronted with promised neither. The DIR was supposed to create a data center from scratch and persuade agencies to participate by providing a superior business case. The best chance of success was to partner with a private firm to market and operate the center.
The DIR negotiated a contract with a nationally known outsourcing firm for operations. The first customer came on board in 1996 to escape an outdated, barely functioning computing environment. The agency had no money to upgrade. Other agencies followed when the Legislature required them to consider a business case, but there was no stampede. Mostly small agencies used the center to bridge their migration to a client/server environment.
At first, the center's performance was poor, so even when the business case closed, agencies were reluctant to use it. Client agencies constantly groused about the service and asked for waivers to avoid participating. Initially the only way customers could talk about the data center was through waivers. There was no organized input from customers to the data center operator. Although the DIR identified the data center's growth as a critical success factor, success eluded us.
Lessons Learned: Stakeholders and Timing Count
At my insistence, an advisory board was formed, which gave stakeholders a say in how the data center operated and where it would go. Participating agency CIOs -- like Sally Anderson at the Department of Mental Health and Mental Retardation, and Brian Rawson at the Texas Education Agency -- had an opportunity to contribute their considerable knowledge and experience to a plan for the center's growth. They also had an appropriate forum to express their concerns. I began discussing agencies' problems and poor performance with data center operations staff weekly via video conference to monitor remediation and improvement.
The DIR engaged a contractor to benchmark the data center against best in class. The data center operator markedly improved performance in the following two years, meeting and exceeding service-level agreements. Once performance issues were addressed, participation in the center became a financial decision, and that case could be made. As more customers joined, economies of scale were realized. Both the customers and the operator were more satisfied with the result.
In the past nine years, I've championed a lot of enterprise causes, but not all have been successful. One consistent factor in successful enterprise projects is timing -- getting out in front of a technology before it becomes ingrained in silos. Even with the right combination of vision, leadership and governance, if the technology is entrenched in silos, the battle for mind share is formidable.
Competence: Overcoming Financial Black Holes
In 1994, a handful of legislators appreciated IT's potential to improve government performance. What most of them saw, however, were huge multiyear projects that seemed endless and never had enough money. They were held hostage by promises of benefits that never materialized. Legacy systems languished for lack of confidence that they could be capably replaced. There was a sense of IT as a financial black hole.
In the mid-1990s, I met with a group of the most influential state agency CIOs to discuss the sorry state of IT funding in Texas government. We decried the lack of faith, the failure to understand the potential, and the perception that technology projects were disasters.
"Is it possible the perceptions are accurate?" I finally dared ask. A long silence was followed by a slow realization that projects were consistently late and over budget; that results were undocumented and benefits unrealized.
Lesson Learned: Smaller Chunks, Bigger Value
To realize the potential for IT to contribute value to state government, we had to be more accountable for project success and benefit realization. The creation of a Quality Assurance Team (QAT) -- composed of the state auditor, the Legislative Budget Board and the DIR -- forced greater discipline on agency project administration and made project status more visible to executive management. Although the quality team was created through legislation in 1993, we implemented the concept after I became CIO. We got in people's faces about project planning, risk analysis and management, and monitored progress.
The QAT improved project performance by requiring solid planning and risk management. The process also engaged C-level executives as executive sponsors based on the understanding that without good project and risk management, problems aren't relayed to the C-level executives who can re-allocate resources to fix them, send a corporate message for cooperation and compliance, or pull the plug. Over the years, I had some very hard conversations with agency CEOs and legislators about troubled projects.
The QAT and the DIR also supported "chunking" -- defining a project by deliverables that added value to the enterprise in less than one year. Former Pennsylvania CIO Charlie Gerhards used this technique to deploy what he called "Pennsylvania power bursts." The value of chunking was that the Legislature was no longer held hostage by promises. It could evaluate deliverables every session, make a rational choice about whether to continue and still have valuable products to show for its investment. For example, portal development is well adapted to this approach since most new online service delivery products can be produced quickly to add immediate value.
We also improved our ability to report the Legislature's return on investments.
The QAT required agencies to submit reports on selected projects that identify how the project measures up against expectations identified during funding considerations. To demonstrate the value of TexasOnline to the Legislature and the agencies, we developed a template for agencies to benchmark performance and costs before and after going online. TexasOnline had a good Legislative session after this material was provided to the Legislature.
The only way for us to have credibility with the Legislature was to succeed at doing what they asked. To that end, we worked hard to keep our agency in order. The state spent about $1 billion per year on IT, and had about 8,000 IT workers when I assumed the job in 1994. The DIR had 100 staff members and spent about $3 million in general appropriations to macromanage IT.
The agency had a small data center for small agencies that operated on a cost recovery basis, i.e. the dollars didn't come from general appropriations but from agencies that bought services. In 1995, we outsourced the data center and put those jobs back into the private sector. The DIR had a fledgling volume purchase operation (about $50 million in 1994) that also operated on a cost recovery basis. In 1998, we re-engineered our volume purchase program to streamline operations. By 2003, we did $500 million worth of business with a fraction of the original staff.
In 2001, the Legislature moved the state's telecommunications operations from the General Services Commission to the DIR, which increased our staff by 100. It too was a cost recovery operation -- for each of the two years before it came to the DIR, telecommunications lost about $10 million on roughly $70 million in revenue. By the end of the first year, we had it back in the black, without raising rates.
When it came time to build the state's portal in 1999, we turned to the private sector, not only for the investment dollars, but also for the expertise and experience to get the job done. We did the same thing with the data center in west Texas. In almost every case, we had to earn our business; there were very few mandates. I think that kept us sharp. If we couldn't attract a customer, we tried to find out why so we could get them next time.
As I leave the DIR, I'm proud it relies on fewer general appropriations dollars than when I started. The state's IT budget has grown to about $1.5 billion, but not as a percent of budget -- it's still around 3 percent. I think we use the money more wisely and with much greater effect.
Today, CIOs -- and IT in general -- are criticized for failing to add value to the enterprise. I believe a measure of that deteriorating confidence is manifest in survey results that indicate twice the percentage of CIOs report to the CFO in 2003 than in 2002. After years of watching CIOs elevating to the executive staff level, there is a retreat to a secondary role. In the May 2003 Harvard Business Review, issue editor-at-large Nicholas G. Carr said, "IT doesn't matter."
It is tempting to gnash our teeth, beat our chests and decry the unenlightened hoards, but perception is reality. There is enormous opportunity for IT to continue to improve performance and transform government. As CIOs, we must be sure we accurately and convincingly demonstrate business value to government executives, and harvest that value for our citizens.
In every case, these lessons and changes that resulted were brought by the DIR board of directors, the incredibly talented staff at DIR, and Texas state agency CIOs who are committed to excellence in public service. There are many stories of IT successes in Texas, and it is they who are responsible.