Resistance to change is part of our primate programming. If a path through the forest was safe and led to food, why wander off of it? Hence the adage, "if it ain't broke don't fix it." Yet, microeconomic theory dictates a company must innovate to remain competitive. After the primary innovation (e.g., a user portal) what is left to do except "improve" it? Sometimes change is beneficial and truly an improvement. Sometimes change is simply lip service intended to satisfy stakeholders (such as managers whose jobs depend on new product). So the answer must be to implement innovative new products only when the marker demands it. But what products and when? Marketers of conventional products solved this problem a long time ago. Surveys, focus groups and limited release prototypes are basic tools that can help prevent flubs and flops. Why megacompanies such as Google and Yahoo don't or won't use such tools is the real question to be asked. - Roger from FriendFeed MT Plugin

Keeping a roll-back option is expensive, I'm sure. Besides, the new designs usually integrate in new central features that are not accessible via the old design and likely fix bugs. Even worse, leaving users with the ability to roll back means that the company will have to support the old design. Let's get a handle on things here, this is getting ridiculous, this is the Internet. Things change all the time and will continue to, they either need to suck it up and get with the program, or they can go off an use one of the many other services that do almost the exact same thing. Whining never solves anything. - Aram Zucker-Scharff from FriendFeed MT Plugin