...unless features are what you're competing on.

Usually this is the correct advice, but there are markets out there of people who need to take the next step up from a basic product and features are exactly what they're looking for.

Take Basecamp for example. There was recently an article on HN from someone who clearly outgrew the product. The consensus on HN was "this guy isn't in Basecamp's target market", and that's absolutely true.

But that also means he IS in a market of people looking for something like Basecamp but with more features. Something in between Basecamp and, say, MS Project. For these markets, competing on features may work.

Just don't go overboard, or you'll end up remaking MS Project :)