On my first day of work at Microsoft as a PM six years ago, I sat down with my first manager for our first one-on-one and at the end she asked, Are there any questions? I said yes -- one last one: "When do we decide to remove features?"
She was flustered. It was not a question she or probably any PMs were used to answering.
I clarified: "Well, features aren't always right. Sometimes they're done wrong, or they don't really fit what the user really wants. Do we ever remove features?"
She remained dumbfounded at the question, and feeling like a n00b, I decided not to press further. In time, I realized why she was flustered. As a program manager, you spent so much time trying to
get features in that it seems nuts to want to remove them. We made huge spreadsheets of feature lists, prioritized by P1, P2, P0 and sometimes P-1. Yes,
negative 1. Because it gets sorted higher, right?
Features got removed in other ways though. If nobody really used them, they were obviously chopped out and memorialized as a bullet point in the release notes. But that wasn't really my question. Those are the easy ones to chop.
A product gets bloated not because the obviously bad features stick around. They're bloated because there are features that are
barely OK in there. They're not complete. They aren't done correctly. Maybe the UI is wrong, or the internal states aren't thought out well enough, or don't match what the user expects. And there are egos attached, too. A poor PM's ego, at the least,and maybe a dev and a tester's self worth too. An entire feature team might have emotional stakes in that feature.
So you can't chop it. And you don't have time to fix it, so it festers. You can never remove features. You have to fix them, painfully, over time.
This can be avoided. Go deep on the things that matter. Do less with less. Be minimal in scope and maximal in completeness. If you're a startup, don't hire. Make it happen with who you've got. Don't get a PM to sit in meetings or create meetings. Only hire do-ers / creators. Do more yourself. If you're a big company, give a skunk-works-sized team a whole shit-ton of power (and really mean it).
Be less. Do less. And you'll somehow end up with more.
You should follow me on twitter here.