In this work, we show the characterization of quantum iterations that would generally construct quantum amplitude amplification algorithms with a quadratic speedup, namely, quantum amplitude amplification operators (QAAOs). Exact quantum search algorithms that find a target with certainty and with a quadratic speedup can be composed of sequential applications of QAAO: existing quantum amplitude amplification algorithms thus turn out to be sequences of QAAOs. We show that an optimal and exact quantum amplitude amplification algorithm corresponds to the Grover algorithm together with a single iteration of QAAO. We then realize 3-qubit QAAOs with the current quantum technologies via cloud-based quantum computing services, IBMQ and IonQ. Finally, our results find that fixed-point quantum search algorithms known so far are not a sequence of QAAOs, e.g. the amplitude of a target state may decrease during quantum iterations.