Yes. But when you used to have a company culture that looked favourably on niche features, and you later start to remove niche feature after niche feature after niche feature, that you start make a lot of people a little disappointed in you.
In other words:
Lets say you remove 100 "niche features".
You have 100,000 users.
Every niche features was used by 0.05% of your users.
Now you have upset 100000 * 0.0005 * 100 = 5000 or 5% of what is likely among your most vocal users.
This assumes the set of upset users for each feature is disjoint. Not sure it makes a difference to your larger point, but I have a feeling that there would be overlap.
Niche features tend to be disjoint, in that they're mainly defined by use case and audience. If a bunch of little things are used by one audience, that is often treated as just one feature, even when they would be considered separate features if they were for different audiences. If something is used across audiences, the feature generally isn't thought of as niche.
Even if they're not disjoint, you've pissed off what, 1-3% perhaps? now consider the fact that each of the pissed of users has multiple reasons to be pissed off, they're going to be a lot more vocal, which is going to be worse for you.
