Pollination is an important ecological process. However, the needs of plants and pollinators are not always met. Pollen limitation commonly reduces seed set and bees often experience nectar dearth. Using a theoretical cost-benefit optimization model we show that natural selection acting at the level of individual plants and pollinators will result in positive feedback that exacerbates pre-existing imbalances between nectar supply and demand. When pollinators are scarce plants will be selected to produce more nectar to outcompete other plants in attracting pollinators, and when pollinators are abundant plants will be selected to produce less nectar. We encourage the testing of this novel hypothesis and propose several ways of doing this via comparative study and experimental manipulation. We also suggest that evidence for seasonal variation in foraging conditions provides preliminary empirical support. If our hypothesis is correct it means that pollination faces a particular challenge in balancing nectar supply with demand.