I serve on the Information Services Board, (ISB) a group of high-level state and private sector people who review and approve the state’s investment in information technology. Specific projects that are high risk get a lot of attention from the group. It’s interesting to me, and one of the few things I do that’s directly related to my previous life at Microsoft.
I wrote the following text as I sat in the meeting last week.
We are reviewing a number of programs:
FamLink, a replacement of an ancient dysfunctional system to keep track of outcomes for children in the care of the state in some way or another. We track cases of child abuse and other problems, plus placements of foster children, etc. The system handles payments and outcomes. They’ve done a reasonable job replacing the system. The first part went live in February 2009. Phase 2 of the project is a little late, and we’re giving them a few more months. (It’s not like we have any choice – the project is late.)
The old data about outcomes for kids is intractible and hard to translate into new formats. This doesn’t affect payments to vendors, but does affect the amount of information that’s available to people who track placements. They’ve done a reasonable job of releasing functionality one step at a time, and fitting it into the middle of complex web of accounting softare at the state. I’m comfortable with the risk choice here.
Another project we’re looking at is a system the Office of Financial Management (OFM) wanted to use to manage grants, loans and contracts across the whole state. This was a disaster, and we would up not delivering a system. We pulled the plug a while ago and today we’re hearing the post-mortem.
Unfortunately, it seems from the presentation that we’re paying the post-mortem vendor by the word. The short story is that OFM tried to do too many things in one system and didn’t ask enough specific questions in the planning. It’s not a huge system, but a painful lesson that we learn over and over again.
The Court system is the next presenter (this is a half-day meeting.) an aside during their presentation is that they are expecting 6000 new “infractions” a day as a result of the tolling methodology being chosen for the 520 bridge scheme. I had realized that there would be some interaction with the court as a result of all electronic tolling, but this seems over the top. I will investigate this and see if we can come up with a better system. (Later update: this was a big problem and is getting fixed. The interesting question is “How do we most efficiently allow people to pay online?” rather than “How do we give people tickets when they don’t pay in advance?” We are working on changing the approach.)