Mature, well-developed markets are close to zero-sum. Think Wall-Mart or McDonald's mature: niches where there are old, aggressive, amazingly well-optimized players, generally two or three in number, who occupy the landscape so completely that for one to grow the rest are probably going to have to shrink.

It isn't a perfect zero-sum, and even in those markets companies are always looking for new ground to colonize, but it's close enough that to first order it makes sense to analyze things in zero-sum terms.

And it's also the antithesis of start-ups in most cases, which are frequently attempting to create new markets where none existed by serving needs that no one knew they had.

I don't think your example is at all zero-sum, because the overall size of the economy can and generally does grow, and thus even the most well-entrenched markets like discount department stores or fast food can and do grow. Perhaps it's true that a market with a few well-optimized firms, each of which can't increase their market share without taking it from another firm, but even then, the market size can grow.