While writing the previous true story about the real world of IT financing, I was inspired to share another horror story that I’ve encountered in my years in IT. It begins with being sent in to sort out a major counterparty risk project in one of the banks that was supposed to be delivering a solution for Basel 2 compliance. What the project was is probably irrelevant to the story but anyone spending money on a project that at that point had been running for 18 months and not delivering anything – spending in excess of 10 million Euros per annum – must ask serious questions. It quickly became clear that the project had only 4 internal staff with more than 50 externals. Cynically, one could say that the externals were milking the programme for whatever they could get. Slashing the project to 15 people delivered more in 6 months than 60 people had in 18. Adding a proper steering committee and asking real questions rather than the previous “nodding dogs” meant real focus on value for money.
All the technical cloud evangelists are going to hate me for dwelling on the practical financial matters of establishing a cloud project but I am afraid these issues cannot be ignored. If organisations cannot break the deadlock on the problem of first man in or last man out with an sensible approach to managing allocation of costs, then the organisation will be in danger of missing out this critical technology revolution.
In case you have not read my earlier posts, or in case it was not clear, then let us take a real worked example of the challenges faced by a typical organisation in transition to the cloud. Smaller, more nimble organisations may not face this issue so starkly, but certainly the medium to larger ones will.
In my last post I hinted at the dilemma of who covers the costs for a cloud project. There would appear to be few issues and some major advantages to scaling out to the Public Cloud, as you can pay as you go or pay for what you use. But is this really the case? The problem with transitions to any shared platform is that nasty subject of who pays. Typically organisations will allocate shared service costs with an allocation key for the costs to various groups using the shared service. This allocation key will determine how much each person pays. It is obvious that as the shared service grows the unit price will trend downwards and will eventually level to a free market unit price but the first or last user may for an interim period pay much higher prices. This “first man in” or “last man out” problem causes all sorts of grief for projects in the real world, not only for cloud but for many wider IT projects.