Sir

The probable failure of NASA's Mars Polar Lander, coming so soon after the loss of the Mars Climate Orbiter, will once again call into question the agency's policy of ‘faster, better, cheaper’ missions (Nature 402, 565; 1999). There is a perception that faster and cheaper is only achieved by taking undue risk. The European Space Agency (ESA) is now changing the nature of its science programme to include faster and cheaper missions, so it is important to understand the pitfalls and benefits of the US experience. Using databases of past US space-science missions, I have assessed the impact of the changed philosophy on the NASA programme.

The number of space-science missions fell systematically from 40 in the eight-year period beginning in 1961 to just 11 in that up to 1992. A projection of the programme evolution made in 1992 would have predicted that, early in the twenty-first century, one space-science mission per decade would become the norm. Clearly, there was a need to reverse the trend: to make missions take place within a shorter schedule and cost less.

Since 1992, the trend has indeed been reversed and there have been 25 space-science mission launches — double the rate in preceding years. But what has been the effect on the failure rate?

As a measure of reliability, I have considered the number of NASA missions for which a failure in the spacecraft brought science operations to a halt earlier than otherwise would have been the case. Such an early end to operations does not necessarily mean that the mission was a failure — often the spacecraft failure came after the prime objectives had been achieved. In the 1960s there was an early end to operations in at least 30% of the missions. Each mission usually involved new technology and at that time there were many lessons to be learned. In the 1970s and early 1980s, an early end to operations occurred in only about 10% of space-science missions. In the years immediately before 1993, the rate was 18%. Since 1992 it will be 24% if Mars Polar Lander is unsuccessful.

A second way of making a comparison is to identify the missions for which it is indisputable that prime objectives were not achieved. In the 1960s, problems occurred before prime objectives were achieved in only 5% of missions. From 1968 to the end of 1992, the failure rate was only 2% (one in 55 missions). For the seven years since 1992, the rate will be 24% if Mars Polar Lander is unsuccessful (six out of 25 missions). If there is a clear distinction between missions of the ‘faster, better, cheaper’ era and earlier ones, it is in the increase in the relative number of missions that fail to achieve their prime objectives. Causes of these recent failures include use of a relatively untried launcher, omissions and misinterpretations in the project implementation at several stages, and a failure to convert imperial units to metric.

Before making a final judgement, we should consider the benefits of the new philosophy. The dilemma is that space science carries a high risk, failures are well publicized, and it is less easy to make the same impact with successes. One conclusion is that the US programme has become broader since the change in philosophy. There has been more emphasis on technology demonstration during missions, rather than as a separate activity, and more emphasis on public outreach and education. Broadening the scope of the programme has only been possible because the mission frequency has increased.

For solar terrestrial physics, the emphasis has always been on relatively small spacecraft but in various combinations. Here, the change since 1992 is not dramatic except for a considerable reduction in mission gestation time. For astronomy, the change has been from major missions that one by one extended the wavelength range of observations to smaller missions, each with a more focused set of objectives, building on the observations of the earlier spacecraft. The SWIFT mission is a typical example of a faster, cheaper mission building on discoveries of earlier missions. For planetary science, there has been a major transition from relatively rare but comprehensively instrumented observatories to several much smaller missions with well-focused objectives and with technology demonstration an important ingredient.

To some extent, ‘faster, better, cheaper’ was a natural next step to follow the major observatory missions of the 1970s and 1980s. But the change in philosophy has not been introduced to the exclusion of the ‘slower, expensive’ missions: Chandra, SIM, NGST, TPF and Solar Probe all figure in the programme and are the equivalent of missions before 1992. There is no reason to conclude that the changed philosophy has led to a diminution of the science return. On the contrary, while major missions remain part of the programme (though at lower cost), there are now many additional missions, each with short gestation times. Those faster, cheaper missions provide opportunities and science return that was missing from the programme in the 1980s. Perhaps the individual missions may not be ‘better’ than those of the earlier programmes, but, for the programme overall, ‘better’ is a suitable description.

ESA has learned from its experience with Cluster to be wary of untried launchers. Fortunately, Europe moved away from mixing metric and imperial units several years ago. The other NASA failures give warnings on how far a ‘faster, better, cheaper’ approach can be pushed. An advantage that we have in Europe is that there are faster and cheaper programmes running in parallel with the ESA programme. These national missions have a smaller scope than ESA missions and so are implemented more quickly with smaller budgets. These, too, can provide valuable lessons on how far the quest for low cost can be taken. We need to make sure that there is good communication between the ESA and European national programmes, so that we all can learn.