You have a good point, Frodo, but I believe that it is appropriate to at least develop a metric of how much data or science time to expect per dollar invested. I don't care what you use for a metric so long as it is reasonable and reflective of fulfilling the goals of the missions.<br /><br />For example, the Mars Rovers budget was such that we are getting science from Mars at a rate of something like $100/minute (plus whatever the ongoing cost of the mission team at JPL is). Conversely, the ESA's Titan probe produced science data at a rate of millions of dollars per minute.<br /><br />Time may not be an appropriate metric, and you may even conclude that with bytes of data, some data is inherently more valuable than other data. Fair enough, however I think it is still useful for scientists to justify their projects based on some means testing of cost effectiveness, as compared either to other programs in the past, or some standard target of dollars per science minute, or megabyte, or kg of material.<br /><br />How much data a probe recovers can be compared against its designed capacity to recover data, and the capacity per dollar invested. The Mars Rovers are obviously a very effective program because they have been running for two years while they were designed for merely 90 days of operation. That is a credit to the Rover designers and builders. Contractors involved should be rewarded for such great performing equipment. That will incentivize contractors into building excellence into future probes.<br /><br />I read once that back during the Roman Empire, contractors got paid 50% on building completion, and 50% if the building, bridge, or aqueduct was still standing something like 10, 20, 40 years down the road.