[tomkarlo]

If 5% of users utilize a feature, and it's pulled, you're going to lose a fraction of those users. Being afraid to lose a few users sometimes leads to horrific swiss-army-knife products where the existing users are placated but you lose new users to some simpler product that comes along and covers 50-70% of your use cases but has a much simpler and more efficient UX.

I've seen this firsthand too many times: the loss of users when removing a feature is more concrete than the gain in simplicity (particularly because new users are generally not well represented in feedback vs. old users) so there's paralysis over removing features that just never really got traction.