I am not among those who think government agencies never do anything right. In fact, I think most government agencies do a pretty darn good job. Let's get that clear from the get-go.
But I also used to work for state government, and I've covered both state and local governments in action-so I hold no illusions about how hard it really is to get something done-particularly something with technology. In fact, I would say it's much harder to create change in government agencies than the private sector, and perhaps with good reason. (Do we really want a government that can change course on a dime?)
So, from my personal and observed experiences, I can confidently say that the SOA solution built by IT departments at the New York State Department of Taxation and Finance is pretty darn impressive, in large part because of its efficiencies.
The SOA case study was the winner for Government/Public Sector in the SOA Consortium/CIO magazine case study contest. The overall winner was Cisco's Commerce Transformation SOA project, the details of which SOA Consortium member, consultant and ebizQ blogger Brenda Michelson shared last week. During that post, she mentioned the New York project in passing-much to my chagrin-but I'm pleased to say she returned this week with an in-depth look at the New York Department of Taxation and Finance's e-MPIRE SOA project.
In short, the state's Department of Taxation and Finance wanted to bring its tax system-including filing-into this century. Of course, they faced the usual legacy challenges-IT was supporting several different legacy technologies, including a legacy system for processing personal income returns that dated back to the 1970s. In fact, that system relied on a homemade database system developed by employees who had already retired.
So, goal one was to move the tax system off all those unsupported legacy systems. That's tricky enough-but you have to add in the burden of supporting changes to the tax code every legislative session.
Interestingly enough, the department tried to solve the problem by hiring it out. The winning vendor suggested a system that was essentially a black box-with security, business rules and work management precoded.
The state, to its credit, rejected this approach and decided to design its own system-one that was more open and could use some existing systems.
Unfortunately, the case study doesn't include the level of detail found in the Cisco case study-which may be one reason why Cisco took the overall prize. But it does include the results and four SOA success lessons the department learned from the project.
Among the results: The work backlog was reduced by 80 percent and tax law changes that previously took six weeks to code and two months to test now took a mere month for the entire coding/testing cycle. Plus the processing of tax filings went from a rate of nearly 150,000 returns a night, with a 24-hour delay for fraud detection to a high of 390,000 returns in a night and near real-time fraud evaluation.
But perhaps what's most impressive is that the department was able to obtain executive support for a project that took the long view-the system should be able to grow with the department for the next 20 years. It also managed to build consensus on what that project should look like AND get IT and the business units to work together.
If you work in government, you should find inspiration in their case study. And if you work in the private sector, you may find there are lessons you can learn from your government counterparts.