Crowdsourcing can be applied to the Internet-of-Things (IoT) systems to provide more scalable and efficient services to support various tasks. As the driving force of crowdsourcing is the interaction among participants, various incentive mechanisms have been proposed to attract and retain a sufficient number of participants to provide a sustainable crowdsourcing service. However, there exist some gaps between the modeled entities or markets in the existing works and those in reality: 1) \textit{dichotomous} task valuation and workers' punctuality, and 2) crowdsourcing service market \textit{monopolized} by a platform. To bridge those gaps of such impractical assumption, we model workers' heterogeneous punctuality behavior and task depreciation over time. Based on those models, we propose an Expected Social Welfare Maximizing (ESWM) mechanism that aims to maximize the expected social welfare by attracting and retaining more participants in the long-term, i.e., multiple rounds of crowdsourcing. In the evaluation, we modeled the continuous competition between the ESWM and one of the existing works in both short-term and long-term scenarios. Simulation results show that the ESWM mechanism achieves higher expected social welfare and platform utility than the benchmark by attracting and retaining more participants. Moreover, we prove that the ESWM mechanism achieves the desirable economic properties: individual rationality, budget-balance, computational efficiency, and truthfulness.