The blog post I linked to above specifically addresses this (https://daringfireball.net/2012/09/get_the_fainting_chair). The specific consequences of delaying Apple Maps would have been either:

1. Not having turn-by-turn directions on Apple Maps, arguable the single most important feature for a mapping app.

2. Share more of Apple users data with Google in order to support Google Latitude. (https://en.wikipedia.org/wiki/Google_Latitude)

So we have two choices: Not have the single most important feature in a mapping app, or violating one of the basic tentpole features the iOS brand, user privacy. So obviously they went with the third option, launch their own mapping service despite its flaws, and I've never heard a convincing argument that that wasn't the best choice.

Because Maps most likely needed user feedback and crowd sourced data collection to be improved. The longer the wait the bigger the gap to Google Maps that had been live for many years, happily collecting data about what streets are open, where the stores are, how the traffic works in a certain locations and so on

> Similar to how they secretly kept Intel builds of OS X in their back pocket for years.

I had a G4 PB and then an Intel MBP. The transition was better than expected, but certainly not without bumps. In a lot of ways, it was similar to maps in that to really finish they had to get it out the door.