AOL, for example, set off a user revolt in 2006 when it suddenly transformed the Netscape.com portal into a “social” news service. By the time AOL reversed course the next year, the Netscape portal had lost half its audience.
I remember reading that Calacanis was praised for his role in transforming the site before they switched back. Maybe I'm mistaken (I've always avoided AOL). But half of their audience? That's a very serious hit.
I've been turning to Yahoo more often lately. I've noticed that with more complicated, technical searches, Yahoo now seems to give more useful results than Google.
The home page is still annoying, but at least it's avoidable.
I can relate to the benefit of slowly transforming highly used software. Recently, we did a massive upgrade to our forum software (the most used part of our system), changing the look, some of the feel, and some of the functionality in one fell swoop, and with the instant change, there was quite a revolt from the userbase.
Luckily, were able to quickly revert much back to begin incrementally making the changes to little effect. Had we kept it like that, given the response from the users, we could have lost some long customers.
It was our mistake for making such a huge change in one massive patch, but we've learned from that mistake.
It's like back in the day with eBay and their updates. The users were used to their yellow-ish background, and the company was looking for change. So while revamping their design, week by week they would alter the hex value from yellow to a grey color. Without the users knowing, they dynamically changed and improved the layout as a whole.
I remember reading that Calacanis was praised for his role in transforming the site before they switched back. Maybe I'm mistaken (I've always avoided AOL). But half of their audience? That's a very serious hit.