Outsourcing is usually a way of getting rid of problems a company cannot solve itself. That was what Michael Dell, one of the computer industry's great innovators, told the Harvard Business Review in 1998.
Mr Dell was talking primarily about the information technology business, but how many instances of outsourcing, whether of IT or other activities, does this remind you of? "The classic case," Mr Dell told the journal, "is the company with 2,000 people in the IT department. Nobody knows what they do, and nobody knows why they do it. The solution: outsource IT to a service provider, and hopefully they'll fix it. But if you look at what happens five years later, it's not necessarily a pretty picture."
It is five years since Mr Dell said that, and he is right - it is not a pretty picture. An international survey by PA Consulting last year found that 66 per cent of companies were disappointed with their outsourcing contracts. Only 39 per cent of the companies said they would renew contracts with their existing outsourcing suppliers and 15 per cent said they planned to bring services back in-house
On Friday, Network Rail, owner of the UK's rail infrastructure, decided to take three railway maintenance contracts in-house after Jarvis, a contractor that has been in the news over a fatal accident and two derailments, announced it was giving them up.
It is not just companies that are unhappy with outsourcing. Speak to the swelling crowd of customers dissatisfied with the service they are getting from banks, airlines and utilities and listen to their anger at having to deal with call centres run by outside contractors.
What has gone wrong? Outsourcing started as a set of sound principles that worked when implemented by competent managers. Then, however, outsourcing followed the path of all management trends: less competent managers picked it up because it was fashionable, because management consultants told them it was a good idea, and - above all - because it seemed the perfect way to cut costs.
Outsourcing has become just another fad, Michael Porter, the Harvard Business School professor, told me when I met him in London last week. Many of the companies that think they can cut costs end up disappointed. What looks like a good deal when the service provider first suggests a price turns out to be a lot more expensive when the contract comes up for renewal.
By that stage, bargaining power has shifted. The companies that outsourced their services have often lost the ability to do the job themselves. The provider of the service is in a strong position to demand a higher price. There are other hidden costs of outsourcing, Prof Porter says, including a loss of control over the quality of work done.
To rethink outsourcing, we need to go back to the beginning. Outsourcing is not new; it has been one of the constants of industrial history. As John Micklethwait and Adrian Wooldridge point out in their book The Company, Henry Ford owned the land on which the sheep that produced the wool for his car seats grazed. John D. Rockefeller owned the timber yards that made the barrels for his oil.
As companies developed, they began to concentrate on only part of the process, designing and assembling cars, or exploring for and extracting oil, while buying the wool or the barrels from someone else.
In recent years, faced with pressure to cut costs, companies began to ask other questions: do we need our own security staff, or could we buy the service? And why do we run a staff restaurant when there are catering companies that do nothing else?
As computer technology became more complex, companies turned to IT providers and consultants to do the work for them. These suppliers knew where to buy the latest equipment and how to install, integrate and maintain it. Except that they often did not. Not only could these IT contractors often not get the systems to work; they appeared to have no understanding of the business that was using them. Stand-up rows between employees and IT providers have become staples of modern corporate life.
The point many companies seem to miss about outsourcing is that it does not mark the end of managing an activity but the beginning of managing it in different, and often more difficult, ways. Instead of ensuring your own employees are doing their jobs properly, you have to ensure someone else's are. And you have to do that without the familiar tools of hiring, firing, promoting and rewarding. You can argue with a contractor and threaten to replace them, but once they have intruded themselves into your organisation and systems, that is often not easily done
You may have "service level agreements", specifying the details of what you expect, but these are blunt instruments and deal with poor performance only after it has happened. Some companies handle relations with their service providers well, but it is a job for the highly skilled, particularly when the service provider is on another continent. As Mr Dell said, companies that outsource often turn a problem they cannot manage into one they can manage even less.
The other thing companies need to do is go back to the basic principle that they should never outsource what matters most. This is often not what it seems. Track maintenance might seem a subsidiary task to a railway company, one the consultants would describe as "non-core". But what that railway company is really contracting out is its reputation for safety, and nothing is more core than that.
There is another thing that is core to any business: customers. Only by constant, intimate contact with customers can companies not only provide a decent service, but detect where the business is going next. That so many feel they can safely entrust this task to someone else - with the dismal consequences we see today - will, I suspect, be a source of astonishment to the business historians of tomorrow.
