Dropping outliers can be done when outliers cloud the analysis, but doing this in an analysis of startups is inane since startup investors' entire goal is to find outliers.
It's not clear that dropping outliers is a bad idea there. It's also not clear it's a good idea, granted.