We suggest a theoretical framework for the design and development of a "needy" socially assistive robot (SAR) for the elderly. Our hypothesis is that a needy SAR will be more readily accepted by those that require them and would result in overall more positive care outcomes when compared to a non needy alternative. This is supported by existing works which indicate traits that can be considered needy are viewed favourably when correctly integrated. We provide guidance on the key design considerations that should be made in order to implement these traits successfully into a SAR. Using this newly created framework we will subsequently proceed to evaluate the effectiveness of integrating a SAR with these traits in a real-world scenario.