The point is that the state foots a £2000/day bill to replace a civil servant who'd be paid what, 10% of that? For a total cost of 20% including charges and the like? Which is an odd move if the goal is to reduce costs.
The trouble with this argument is that it is based on the assumption that you could actually hire someone with the skills necessary to fix the much-maligned IT facilities of a major government department for a salary to them of, say, £30k. Given that someone with those skills has a market rate that is an order of magnitude higher than that, this seems implausible.
If this person is actually competent to do the job, then given the amount of money the MoD can expect to save as a direct result of changes she brings in, the rate they're paying is reasonable. It may be politically awkward, because the average voter probably doesn't understand the economics involved, but that's a different issue.
30% (according to the article - 500k pa vs. 150k pa).
In any case, it's a new role that didn't exist before, so there's naturally a lot of set-up effort. It makes sense to get a more experienced person before hiring someone definitively.
That 150k comes with a fantastic pension (final salary!), they have to pay employer taxes, insurances etc. Not to mention that they're almost impossible to fire.
It's closer to a 300k vs 500k pa comparision.