Central vs distributed IT support

Next session was on centralised vs decentralised IT support from the University of Mitchigan. 58000 students. 3 campuses. 19 schools and colleges.

Very decentralized organisation. It's not an accident, it's a strategy. Decisions distributed out to deans etc. Budget distributed out to schools and colleges. Deans set strategy for their units. They have their own admissions staff, comms staff etc. IT also distributed. Each has own IT group reporting to academic leadership. Even professional services have own IT groups.
Some groups run very general stuff including networks, wireless, backups, email, calendar. They've gone so far down this route that they've lost many opportunities to get benefits from centralization. Eg wireless across campus run by different units and isn't interoperable. Same wheels being invented over and over again.

So, they've recognised that they need to change. Big funding challenge means they have to drive down costs. Set up an IT rationalisation project. Sponsored by provost and CFO and Accenture brought in to advise. The goals are to reduce IT operating costs without reducing quality, to have smarter sourcing policies, reduce redundancy, and have a sensible transparent cost model.

First stage was to assess the state of IT at UM. Looked like a nightmare to me! Loads of examples of different groups doing the same thing in different ways, lots of duplication.

Second stage was to define the different layers of services:
Public good. IT services can be used for all units, everyone pays whether they use it or not.
Toll services, most users can use
Community services, used by segment of common users
Unique, IT services used by one unit

At moment too much is unique. Working on defining public good. Interesting that they don't appear to be saying that you have to use them, but you have to pay for them. Eg engineering will probably carry on running their own network.

Too much of school IT staff time is spent running back end services, could be done better by someone else. Not providing value added. So, looking to move more to central. But service has to be as good as being provided.

Some schools and colleges engaging with process more than others. Most keen to keep a local face. Eg first and second line tech support. Keep strong ownership at faculty level.

In schools that have engaged, they have taken on more emerging needs, and help more, and have increased customer focus because central IT taking on back end stuff. Have also reduced staff in IT. Have decommissioned server room which has been given over to research.

But cannot continue with incremental pace of change. Budget demands and competitive demands too great. Now have to seriously rationalise

Some issues:
Without the right funding model progress will be limited. If school has to pay central IT for full cost of services, no incentive to change.

Local IT is great at personal communciation and engagement. Don't want to lose this good thing. Can drive costs down, but if personal engagement is lost, will be a bad thing.

Central IT has to get really good at sourcing. Don't try to run everything themselves. If everything moves to centre won't succeed, need to get rid of stuff eg commodity stuff. Pick the right things to run.

IT offerings have to meet reasonable quality expectations, true for central and unit offerings. Reasonable is the important word. Schools have given central IT their requirements, central IT design services, but then it's too expensive. When look close, the school often doesn't meet these standards. Like gardening, if pay someone, has to be perfect. If do it yourself, don't care as much.

IT has to be responsive and agile. Central IT very project focussed. Finish project then move on to next big thing. Need to get things out quickly and incrementally improve them.

Central IT needs to reach out to edge of organisation and talk to these rogues out there doing their own things. Same true for edge. They have to understand central IT. It will take both sides to make the relationship work.