For some purposes it's understandable - we send the probe, get the science it's designed to collect, and we're done. And even in those cases if the mission fails we do not attempt again - we collect the crush data and scrape the mission. For example the schiaparelli lander on mars or the long tether in orbit experiment.
Why don't we launch few Hubble or Kepler telescopes each costing about 500 million to develop and probably 50 - 100 to build now that we have the design. NASA spends 9 billion on James Webb, why not spend few hundred million to launch more of those observatories? Why no more shuttles are built if it's such a great design?
And why don't we ever launch few probes for a backup if one fails?