[nickm12]

My experience working in a large internet services company is that relatively few of our software projects fail, but about 50% of our software systems are failures—in that they deliver insufficient value and are deprecated within 1-3 years of their launch.

For such systems, I don't think there it's fair to assign a single point of failure, but typically the failure is correlated with a couple of factors:

- It was built by a team with a shallow understanding of the problem to be solved (and not enough time/interest/skill to get a deeper one)

- It was built by an org that was not stable enough to provide the long term support it needed to prove its value (1.0 is an important milestone, but 1.0 software will die on the vine without the support to get to 1.1, 1.2, and 2.0)

"Talent" obvious plays a role, but I feel it is hard to answer the question "is this team talented enough?" as compared to "does this team understand the problem well enough?" and "will this teams organization support them long enough?"