I’m thinking a lot about an article penned by guest editor Mary Kurkjian, published in the Albuquerque Journal (http://www.abqjournal.com/opinion/guest_columns/052250582652opinionguestcolumns11-05-10.htm), titled “New Governor Needs New Management Principles.”
Kurkjian has worked in state management and has been a consultant in 38 states, including NM, so I’m willing to grant points for her authority to speak on the matter. She works her way through several very good points, including the need to appoint people qualified to manage the area of their appointment (something we’ve seen both governors and presidents fail to do), the need to establish and enforce strict professional standards (obviously NM has been aching for this for generations), and the critical need for transparency (which our new governor Susana Martinez has already addressed by her second day in office).
The one that caught my eye was this: Kurkjian suggests reviewing the state’s IT projects, and bringing down the accountability hammer on agency heads responsible to make them work. She mentions NM’s notoriety for failed IT projects, and the catastrophically-failed SHARE initiative. But I’m fascinated at the reasons she sees: under-resourcing of the projects themselves, and under-training of staff users. Why am I interested? Because that’s exactly what I’ve seen in some cases.
Yes, I’ve learned, you can indeed build your own software. If you’ve got a talented team, and management that understands that defining needs clearly and conducting frequent usability tests are critical, this is a great path. But if users aren’t involved, if people can’t understand the software on their screen, ANY project is doomed.
Since a clear understanding of what it takes to develop software in-house is not usually part of an organizational manager’s repertoire of skills, agencies are likely to be more successful at contracting software development to businesses that specialize in it. This brings other problems: some organizations can’t clearly spell out what they need, in some cases not even with crayons. They learn the hard way, by getting software that doesn’t even resemble what they really need, but nicely fits the specifications they themselves wrote.
And yes, you can buy some great software too, though that brings its own risks. Either it can’t be modified, so users have to change their business practices to fit the software, or it has to be modified, creating a long a buggy process that locks you into a single aging version of a product. Ask anyone who has implemented Indus Passport or Banner how painful this is. But it can be done.
What all this means is that NM needs more involvement from state IT departments in clearly defining needs and requirements, performing constant user testing, deciding on change orders if they are (inevitably) necessary, and managing the features and security of the software long before it’s deployed to users.
At least when it comes to the importance of this process, Kurkjian gets it:
“IT is no longer an appendage to government managemen; it is at the core of it.” – Mary Kurkjian