I don't know that anyone has asked themselves this, but why would a 350 day test be a part of a milestone in a technology development contract? Answer: such long tests are evidence that the technology in question is ready to be inserted into a product for the marketplace.
This implies that when IH licensed Rossi's technology, milestones were written for Rossi to take a demonstrated technology and engineer it to be reliable enough for preparation as a product. The question is, did Rossi's 350 day test demonstrate the contractually required reliability?
I have seen a lot of polarization in the speculations about the test. I hope that the raw data will be released because I would like to analyze it myself as an engineer. The results of the 350 day test are likely to be a mixed result. It is highly likely that Rossi demonstrated that LENR is real, and can produce substantial heat; but the question is, did he demonstrate the contractually required reliability that would put IH in a position to start delivering product and making money?
Many times the first pass at such reliability testing finds core problems in the reliability that requires re-engineering; or worse, re-invention, before the technology can be brought to market. So, it would not be surprising if Rossi's technology worked, but was simply not reliable enough to move to a product development/delivery phase - a phase where IH would realize profit from the delivered technology.
A couple of other clues in this ... Rossi was constantly blogging about babysitting his reactors for long periods and until late hours in the night. A reliably engineered technology would not seem to need such babysitting, because the babysitting implies readiness for manual intervention. Second, as I recall, DGT, whose earlier role was to become the manufacturer of Rossi's technology, claimed that Rossi had not delivered a reliable reactor, and because of this they didn't pay. This reliability issue with Rossi's eCat technology may be a problem that Rossi has not yet resolved.
This issue of reliability of a technology is a contentious point between businessmen and technologists. It is one thing to demonstrate that a technology (LENR) is real, and a completely different matter to demonstrate that is ready for delivery into a product that can be sold to make a return on investment. The businessmen are correct to insist that a technology meet a reliability milestone sufficient for insertion into product before production scale investment.