I found the discontinuation of Hypercard really surprising, since Hypercard was what made Macs useful as "business computers" in the niche places where I saw them used.
For example, the musical instrument store where I worked in during high school used Hypercard to keep all their customer records and maintain an inventory of all their equipment. This was all based on custom "stacks" which were in turn lightly modified from included example stacks to add their specific business logic. (For example: the stacks would calculate rental fees and late rental charges instantly based on the store's arcane formulae. They would produce reports telling us who was late on rent and who was most likely to get really late. The stacks were constantly under development, which usually meant: we need a feature, let's pop the hood and add three lines of code.)
At the end of the day HC was just a single-machine database with an amazing programmable custom interface builder. This would have been obvious to me if I'd been older and understood these ideas, but I didn't realize how useful the combination was until many years later when I started building web apps. It occurred to me that what I was doing was basically making a much crummier version of Hypercard (albeit with a shared Internet-hosted database, which Hypercard did not have back then) at something like 100x the programming effort. I remember seeing that MS Access had some kind of interface builder, but it was a pile of hot garbage compared to Hypercard.
TL;DR if Hypercard had managed to stay the course and improve itself for that use-case I think Macs would have been vastly more useful to (at least small) businesses. Instead Macs went through a dark period where the nearly got killed by Windows, and were basically only saved by Jobs' attention to design plus Microsoft's reluctant benevolence.