Mixture got me thinking by mentioning Micro$oft. It was described by a family member thus:
We all know it's easy to complete 90% of a given task; it can take maybe double the time so-far expended (or more) to complete the last 10%. The temptation is to move the developer on after that 'double time' has expired (or sack him/her). Everyone knows that the job is not completely done, but the bean-counters have run out of patience and tell themselves that no-one will notice.
The bigger the company, the more likely for this scenario to hold.
Or so I have been told...
(Reminds me of the British motor industry in times past when it was up to the customer to do the R&D on a new product. The Mark II version of the car was sometimes better.)