The poor benchmarking of central government's IT expenditure is unacceptable. Without this information it will not be possible for the Government to advance effectively its cost reduction agenda. We recommend that the Government should investigate the claims of overcharging put to us and seek to identify reliable and comparable cost benchmarks, and collect accurate information from departments in order to compare with those benchmarks. The Government should use independent and specialist advisers and the NAO to assist with identifying objective benchmarking measurements.I'm pretty sure I was the first person to come up with the "Government IT costs about £13bn a year" quote - sometime in 2001 or maybe early 2002. I did that by asking a lot of people what each department spent, adding all the numbers together, adding some fudge factors to cater for projects with "known unknown" costs (so we knew there was going to be an NHS IT programme but had no idea how much it would cost; by 2003 it was clear of course - 10x more than anyone thought for 1/10th of what was wanted). My £13bn would have struggled under the most elementary challenge, but it is, I fear, still a number in widespread use. More recently I've heard at least one senior government person state that this number may be as high as £20bn if you include all of the staff that government retains to manage their IT (and I know I didn't include those first time round) - one department, many will recognise straight away who, reportedly has some 1,500 staff managing the 3,000 staff that their outsource provider has. Were that to be replicated across government, I could well imagine £20bn.
It all reminded me of William Goldman's quote "Nobody knows anything". And still we don't.
We know, of course, more about some things than we knew then. In 2001/2 we couldn't figure out how much departments were spending on the web - we relied on round robin parliamentary questions (which suffered, as the price of desktops today, from the breakfast problem) or just by asking. Now, at least for a bit, we have the annual COI report - what a shame it would be if that disappeared with the organisation that created it. I still suspect there is something of the breakfast problem about that report.
But ask a similar question - how much is spent on rules engines, how much on ERP, how much on EDRM, what is the average log in time first thing in the morning (ask almost any central department user that question and they will hold their head in their hands so you know you need ask no more), how much is spent on Oracle licences across central and local government, even how many data centres do we have / what is the PUE of each / and how much do they spend on electricity and you will quickly fall short of answers.
And that leads to the better question, "so what's the plan then?"
Interesting times ahead.