Social media producers are currently having a fierce competition worldwide to increase their revenues. To achieve this goal, they are investigating alternative ways to attract more users, generate new user activities, and collect valuable data for personalizing contents and services. One such alternative is crowd computing. Our vision is based on the great potential of a well-coordinated and controlled joint utilization of human intelligence and computer systems that can help solve problems that would be difficult to do with individual capabilities alone. To achieve this vision, which we summarize under our concept of crowd social media computing, we investigate and model the characteristics of the social media ecosystem, we discuss the characteristics of crowd computing, and then we demonstrate how crowd computing can play a pivotal role in emerging social media applications. We also propose a new approach to evaluate the impact of crowd computing on the issue of social media Return of Investment (ROI). (C) 2017 Elsevier B.V. All rights reserved.