You're right that you'll (practically) always have more than one feature in a product, so counting features doesn't really make sense. You should approach it from the other end: identify a segment of the market, a narrow set of customers, and aim to build a product that is clearly the best option available for them. Judge every feature through this lens, and don't stray outside the focus until you've won that niche (if ever).

Easier said than done, of course...