It’s hard to conceive of it now, but in the late 1960s and 1970s, Whitehall and the public sector knew as much about IT, and in some areas more, than the private.

Then began a process which would later come to be popularised as outsourcing.

Roll forward 40 years, and here is John Manzoni, head of the government’s Major Projects Authority, describing the result. The wave of outsourcing that gathered pace in the 1990s, said the head of a body that monitors £500bn of public projects, has left Whitehall bereft of the ‘critical skills’ needed to procure and manage such projects. Execution and delivery were not ‘well-developed muscles’ in Whitehall, said Manzoni, who lamented that as a result of the outsourcing reflex, the government had lost crucial expertise in IT and technology and now lagged ‘five to eight years’ behind industry in these fields.

When outsourcing began to take off in the 1980s, it was sold as a simple win-win transaction. By focusing on what they did best everyone would benefit, as would the economy as a whole as the use of resources was optimised. ‘Outsource everything except your soul!’ exhorted the excitable Tom Peters.

From the outset, it was clear that it wasn’t quite as simple as that. For a start, outsourcers had to be not just a bit but massively more efficient for the arrangement to provide for both their profit margin and cost savings for the customer. Too often, contract terms gave no incentive for providers to improve. In the longer term, both the value of what was being given away and the hidden cost for the customer of not having an important process under its control turned out to be higher than expected.

Take the electronics industry. The Faustian nature of the outsourcing bargain was graphically revealed when Asian component makers eventually started eating not only the lunch but the entire body of western computer firms that discovered too late that in outsourcing ever larger chunks of manufacturing value they had inadvertently given away their soul too. Or aerospace, where obeying Wall Street’s strictures to minimise its asset base, Boeing outsourced so much of the manufacture of the Dreamliner, the 787, that the complexity of its supply chain outstripped its ability to manage it, causing delays to the launch and cancelled orders.

Or, to bring the story up to date, the UK public sector, whose travails, illustrating all the pitfalls described above, were the subject of Manzoni’s strictures. The e-borders fiasco, where the taxpayer has been landed with a £224m bill in damages and costs awarded to US defence firm Raytheon after a contract to upgrade UK border controls was improperly cancelled, is a timely reminder of the knock-on perils of dismal contract management. Calling the award a ‘catastrophic result’, Keith Vaz, chairman of the Commons home affairs committee, said in so many words that the UKBA didn’t have a clue what it wanted from the project.

The economist Joan Robinson once remarked that the point of learning economics wasn’t to acquire ready-made answers to economic questions, but to avoid being bamboozled by economists who put forward such things. The same is true for technology.

Consider the Department of Work and Pensions’ universal credit scheme, another prominent item on the MPA’s watch list. Curiously, UC doesn’t have a ‘traffic-light’ rating (green, amber-green, amber-red, red) in this year’s assessment. The reason given is that it is a ‘reset’. On examination, a reset turns out to be a variant on a dodge commonly used by service organisations to meet arbitrary targets and service-level obligations: closing a case which can’t be resolved within (say) the target time and then re-opening it as a new one, thus winding the clock back to zero. Opening and closing cases is a common ploy of IT help-desks and other outsourced services where the contractor is paid by volume. No prizes for guessing from which quarter the reset idea is most likely to have come, nor for thinking that those who dreamed it up will not have described it to government as what it is: a fiddle, a scam, a cheat.

Nor is it likely that consultancies whose livelihood depends on selling copious amounts of IT will suggest to government departments contemplating large-scale change that IT is the last, not the first place to start. IT is often glibly called an ‘enabler’ of change; but if done first, it is the reverse, locking in a design of work (and basis of payment) that are impossible to change subsequently except at enormous cost. Although less in the public eye than the central government projects, this is a growing issue all over the public sector. In local government, a number of shared-service and other outsourcing deals have descended into reciprocal recrimination when the promised benefits failed (as they do) to materialise because the starting design was wrong. In some cases councils have found themselves stuck with deals which they have learned the hard way to be disastrous, because they can’t afford to cancel them.

The prudent way of buying IT (and avoiding more cases like NPfIT, the aborted NHS computerisation programme, where large damages costs threaten to swell the £10bn already sunk in the failed scheme), is to put it last, not first. Computers are the servant not the master of change, which begins at the other end, redesigning a service to put humans upfront where, unlike computers, they can deal with the variety that human demand comes in. Such redesigns sometimes result in IT having to be removed as a constraint on doing things better; they are always less IT-intensive (and expensive) than before.

The moral of the story is that in order to outsource a process or service effectively, you need to know how to do it yourself, in every important particular. In which case, of course, you may often conclude that you’ll be doing yourself a financial and strategic favour by doing just that.