We got our first look at the new U.S. CIO today and he is an interesting choice. Odds are, we will measure his survival in the job in days and weeks, not months and years. But before he either burns out or makes some massive mistake, he will likely have a much bigger impact than someone more measured might. I think it is worth watching him closely, because if he were taking over a corporate job rather than a government job, his approach would actually be the one I'd recommend. But government is a nasty beast where power is not easily defined and antiquated practices are protected by even more antiquated laws. Even for someone experienced in local government, as he is, federal government can be quite a surprise. He is going to be pissing off a lot of very powerful people over a very short period of time.
U.S. Government Moves to Google
Vivek Kundra loves Google's still immature cloud-based solutions. This will place the government on the bleeding edge of the cloud, which should reduce costs while probably sacrificing reliability, security and, most problematic, legally required document retention. Google's offerings aren't yet truly ready for enterprise use; they have been failing too often. Their acceptance with a broad consumer audience, outside of mail, has been mixed. One of the reasons the U.S. government isn't cutting edge is that it wants technology to work like a utility, always there and always on. When something fails catastrophically, as Google's products have been doing on and off, the government also tends to be really good at finding and shooting both the person who made the decision and the related vendor. If I were Google, unless I was really sure all my problems were behind me, I'd suggest the government not go down this path, if only for my own preservation. At least not yet, though it is clearly making good progress. On the other hand, this stuff is cheap and coupled with Twitter, YouTube, and other Web technologies he also favors, it is much easier to be transparent. Of course, there may be times when you don't want to be transparent, and given that we already had several government idiots using Twitter in a way that put lives at risk, we are likely to see some fun here as well
Custom Software Is Dead
Kundra likes packaged software. Heck, I like packaged software. Custom software is an increasing pain to create and support and it puts you into the software business. The nice thing is that when custom software fails, the blame is shared. When packaged software fails, the blame goes to the packaged software vendor and the poor sap who selected it. Given that government is more about blame than credit, this makes for an economically sound, but potentially catastrophic, career choice. On the other hand, once started this could be incredibly hard to stop. Projects that typically are measured in years could take weeks, or sometimes even days to complete and come in working better at a fraction of the typical budget. He could start one heck of a wave.
The Problem Is the Approach
In a heavy bureaucracy, the first step is to build support and, at least initially, work slowly so you build up a background of success so that a failure is viewed against that backdrop. You also want a clear idea of who has your back and who has a knife ready to put in it. There will be a number of failures in any effort to change a complex structure. Once support that you can rely on is in place, if you move with a measured hand (don't change more than you can contain), you can challenge the status quo. But to jump out early and announce what must look like crazy radical ideas to the career IT folks in government would typically result in a lot of these folks looking for ways to find the new CIO a new career fast. We haven't even gotten into some of the relationships between vendors and the folks who advocate their products. They exist, even though many are likely illegal.
Don't Mess with Communications
There is an old rule that applies to both government and large enterprise: You don't mess with communications products unless you have no choice. I can say categorically that everyone I've seen mess around with e-mail lived to regret it. There appears to be no real upside because, if it does go flawlessly (which it almost never does), that means you have barely met expectations. If it goes badly, everyone from the top to the bottom of the organization knows and most want your head on a pike. So he would be better off focusing initially on lower-profile systems that don't span the entire organization and leave e-mail alone unless it simply wasn't functioning well at the moment. This is generally true of the desktop; government tends to still use a lot of secretarial help and folks, secretaries don't like change. If you mess with them, they can make your life a living nightmare. Messing with a lot of them at once with either e-mail or what they have on their desks is just short of suicidal. WordPerfect actually lost most of its market share forgetting this and bringing out an improved (changed) UI product for this group.
I think what Vivek Kundra as the new CIO is going to try to do is important and that he will have a positive impact. But coming out this soon with an aggressive product statement suggests he doesn't really understand the federal bureaucracy yet and that learning on the job will be a good lesson for us, but an expensive one for him. I think he will have a significant impact and may in fact boot the U.S. government closer to this decade in terms of technology. But I doubt he'll survive the effort. Learning how to do the former without incurring the latter is what we may also be able to learn with him as our proxy. Since he is controversial already, technology aggressive, and will probably leave the job in a spectacular fashion, I'm looking forward to what may be a brief, but very interesting, tenure from our new CIO.