VoIP (Voice over IP) is a transmission technology that allows telephone calls to be made over an IP-based network. A VoIP system may require sampling, quantizing, coding, compression, packetization, and jitter removal to transmit voice signals. The delays introduced by these processes may result in a delay long enough to make the system susceptible to echo problem. Impedance mismatches in analog circuitry and acoustic coupling of the transmit and receive signals at the receiving end is another source of echo. Echo has a direct impact on voice quality, which may determine a user's retention to the VoIP system. Echo cancellation is normally used to remove the echoes, but the removal procedure has itself added another delay value. However, echo not only appeared in long-haul calls but also in short-distance calls of the VoIP system. The objective of this paper is to observe users' retention based on end-to-end delay the users experienced, which is a form of subjective measure on the obtained voice quality. The result can then be used as an input to an objective measure of voice quality of R-Factor. Experimental set up of CRT 3.0, CISCO 3700-series router and analog telephony system were used to investigate the effects. The result confirmed that an end-to-end delay of more than 150 ms would greatly affect the VoIP's voice quality, and thus a user's retention.