Yahoo's blunder with unnecessary changes to their services such as Yahoo! Groups is well explained on slashdot.org by anonymous poster.
More of this person's comment can be found here. If you are interested in exploring the topics brought up by this person, click on the links I added to their quote above. That will take you to Wikipedia articles that will explain each of the terms.
If you have no userbase, the Agile concept of ship (garbage) early and ship (garbage) often even before you really have an MVP actually makes some sense. If you have a 6-month runway of capital before you go belly-up and start over (oh, I'm sorry, "pivot"), there's no point in wasting another month to get it right.
But if you already have a userbase, the developer-centric attitude of leaving what, to users, is core functionality in the backlog while you release half-assed stuff that merely shows off how good you are with AJAX, or how quickly your UX people can change the design from one week to the next, doesn't work. It's bad for your customer base, it alienates them, and it eventually drives them to your competitors.