24 October 2006

The drunken man

By Andrew Clifford

Walking home one evening, I found a man crawling about under a street lamp. "Are you OK," I asked. "I've losht my key," he replied, obviously worse for a few drinks. "Where did you drop it?" "Over there," he replied, pointing to a dark patch of bushes nearby. "Then why are you looking over here?" "I can't see over there. It's too dark."

I have been developing a model of IT costs, to throw some light on how we can save money, or divert more money into supporting business change.

It is a simple model, but not simplistic. It looks at the whole of IT spend, not just spend in one area. It recognises that some costs, like having a PC on every desk, are hard to reduce. It distinguishes between new IT work and changes to existing systems.

I put in some figures based on my experience of large UK companies. The model showed that there was a potential to save 41% of annual IT budget every year. This could mean running three times as many business change projects, without increasing the IT budget.

Of course I didn't believe the figures. So I checked the model. It seemed OK. Perhaps it was my assumptions. So I tried different assumptions. Every time I ran the model, it showed a potential saving of between 20% and 60% of annual IT budget.

To see why the potential for savings was so high, I looked in more detail at the figures. The key seemed to be the total volume of IT. Data centre costs increase with the amount of IT. Projects get harder the more systems you need to change to make business changes. I had assumed small savings in some of the factors that contribute to the volume of IT, but the model showed how these combine to make big cost savings.

Lots of things add to the volume of IT.

These things are common. A small lack of project focus, a few tactical decisions, a little bit of politics, can cumulatively wreck your IT. Small changes to keep IT focussed, to stop it bloating, to manage back to something smaller (such as system governance), can massively reduce overall costs.

Like the drunken man, we search for IT improvements in the light. We scrutinise our development processes; we argue about splits between test, development and support; we work with strategic sourcing partners. We search here because we can see here. But the model shows that none of these things make a big difference. To make big savings we need to search in the dark corners of IT. The legacy system that no one wants to touch. The old database tool that is not used for new systems. The politically inspired system that no-one really uses, but no-one dares criticise.

Next week I will explain the model in more detail, so that you can see for yourself where to look for savings in your IT organisation.