What is your best priority methods for new / old features?

As a leader, prioritizing features are always an art. When I think we have a solid priority system nailed down, in different situations different context and perspective are required. The most difficult is when we need to speed up, but that could end up as a mediocre feature that might not be our best work. What do you do in this case? Do you often prioritize speed? Or risk losing that traction, but rely on the fundamental principle behind the mission of the product. If you have a specific priority method that you use for your product development, please share!