Energy is the major bottleneck for wireless sensor networks. Although energy harvesting sensor nodes have emerged as a solution, new challenges in energy management arise due to the intermittently available energy. We also observe that many applications can tolerate certain levels of data inaccuracies. Here, we exploit the application’s tolerance for data inaccuracy to achieve an equilibrium between energy consumption and availability while satisfying application’s accuracy needs, thereby improving network lifetime. Performance studies indicate that our method outperforms other data collection methods. Moreover, our comprehensive energy consumption analysis of quality-aware data collection on different sensor node arrangements reveals the limitations on energy savings when quality-aware data collection is used. We show that a tighter cooperation between MAC layer and Application layer, that is achieved through an Application-driven control mechanism can unlock the full potential of quality-aware data collection.