In this study, we focus on the development of energy efficient and achievable load balancing mechanisms for wireless sensor networks. Due to resource constraint and tremendous amount of sensors, one possible way of achieving maximum lifetime of the network is applying data aggregation on sensor data. However, existing approaches introduce significant computation and control overheads that often not suitable for sensor networks applications. In view of this, we propose an in-network data aggregation scheme by exploiting the inherent spatial correlation of the sensed data. Each sensor multiplies its reading with a random coefficient and sends the product to the next hop to calculate a weighted sum of all the massage. Instead of receiving individual sensor readings, the sink will receive all weighted sum and restore the original data. By doing so, each node only performs one addition and one multiplication in order to compute one weighted sum and consume the same amount of energy. Simulation results demonstrate that the proposed scheme is efficient and outperforms existing schemes in terms of energy gain and network lifetime.