The technical term is futzing. We've all done it: frittered away a day getting a new IT system to work, rather than doing what we're paid for. Futzing is one reason why computers do not seem to be the boon to productivity their makers would like us to believe.

Doubts about IT's contribution to profits are not good for computer makers, especially at a time when customers are taking a hard look at the costs of IT. Recent research shows the "total cost of ownership" to be much higher than the price on the box. One study, for example, showed that an enterprise resource planning (ERP) system costing $4m in software and hardware actually cost $20.5m to install. Other studies show that only a minority of IT projects achieve what they set out to do.

All this creates a climate of scepticism around investments in IT, hitting the suppliers' bottom lines. In the US, spending on IT is set to fall for the third year in a row. This is why IT firms, led by Microsoft, are investing millions on ways of identifying and measuring benefits.

An organisation called the Information Work Productivity Council has launched a programme of research at the Sloan school of management at Massachusetts Institute of Technology (MIT) to identify new ways of measuring benefits. In the words of one sponsor, James Firestone, a senior vice president of Xerox Corporation, one of the eight corporate members: "It will find ways to unlock the latent productivity in technology hardware, software solutions and service offerings."

The "productivity paradox" has been worrying economists since the late 1980s. They noticed that despite massive investments in IT, US productivity in the 1960s and 1970s grew at an average of just 1% a year. As a Nobel prizewinner, Robert Solow, put it: "You can see computers everywhere but in the productivity statistics."

MIT Professor Erik Brynjolfsson, who is leading the $4.5m research project, says the paradox is being resolved. He says that more recent data from the 1990s at last show a correlation between rising IT spending and productivity.

Earlier this year, Brynjolfsson revealed to a seminar sponsored by BT - another firm with a keen interest in proving that technology helps productivity - the results of a long-term study of 800 US firms showing a correlation between IT investment and market value.

He likens the impact of IT to that of the electric motor in factories at the beginning of the 20th century. Manufacturers enthusiastically adopted electric motors to replace steam engines in their plants, but at first the motors merely duplicated their functions. It was only when manufacturers started to distribute electric motors around factories and design buildings for the convenience of production lines rather than around the central steam engine that the great leaps in manufacturingproductivity became possible.

Something similar, Brynjolfsson says, may be happening with IT. The technology itself doesn't make a difference, but the process changes that it enables do. "IT is significantly more productive when combined with organisational change."

Until now, too many IT systems, especially in government, are "paving over the cowpats", he says. The benefits emerge only in a new type of enterprise, the "digital organisation". Such a firm has few tiers of management, rewards initiative and invests heavily in human capital - it hires the best workers and spends a lot of time training them up. His exemplar is Dell, the PC firm.

Brynjolfsson's ideal of the lean, mean digital organisation looks a little battered in the post-dotcom bear market. With stock prices languishing, companies need to measure value in a new way. This particularly applies to information workers, those who need much investment in IT but whose contribution to output is particularly difficult to measure. Hence the Information Work Productivity Council. Its objectives are "to develop standard productivity models and processes to help business improve information work productivity, processes and profitability".

Susan Conway, the council's industry director, says: "Ultimately, the information age will be measured not by the output of technological products, but by the improvement of the actions and decisions that drive business performance."

But the fact that such improvements need such heroic efforts to pin down, will itself fuel scepticism about IT as a magic bullet.