||"IT contextThe concern over computer systems at the beginning of 2000 should have settled down by the time this paper is written, but it is symptomatic of an underlying, and recent, public concern that computers are not infallible. This has resulted in growing interest in measuring and predicting the benefits information technology can bring to business.The 'Productivity Paradox' contrasts the readiness with which companies buy individual IT systems with the absence of any proof that IT increases the productivity of whole industries. Paul Teicholz has recently shown that the construction industry in the US is not only falling behind other industries but that its productivity has declined in the last 30 years. Similar data can be presented for Denmark and the UK.ObjectivesThis paper relates the evaluation of benefits of IT systems in construction to the time over which they are developed or implemented, and links this with the cyclic levels of activity from which all construction industries suffer. This leads on to suggestions about when IT systems of different types should be initiated in companies, in order to achieve full productivity at the stage of the cycle of activity when it is most critical.MethodologyOngoing work on how IT productivity is measured is combined with statistics on construction industry workload cycles in Denmark, The UK and the US. These show the variations between these countries and the paper suggests the optimum cycles on which different types of system should be replaced. Ths work is based upon surveys of how companies in construction measure IT benefits and government statistics.ConclusionsThe different time scales, ranging from development of customised systems to the introduction of off-the-shelf packages, have a major effect on evaluating benefits based on a changing workload. Little numerical evaluation is carried out at present because there are many different methods and reliable data required for these can be difficult to obtain. Major systems may need to be initiated at the peak of a cycle in order to be fully ready for the next peak, assuming that staff will be retained and training can be carried out during the intervening drop in workload. Work is in progress to develop a better method of evaluating benefits and will take these time factors into account"