This paper proposes a crowdsourcing approach for informative document summarization service. It first captures the task of summarizing a lengthy document as a bi-objective combinatorial optimization problem. One objective function to be minimized is the time to comprehend the summary, and the other one to be maximized is the amount of information content remaining in it. The solution space of the problem is composed of various combinations of candidate condensed elements covering the whole document as a set. Since it is not easy for a computer algorithm to create condensed elements of different lengths which are natural and easy for a human to comprehend, as well as to evaluate the two objective functions for any possible summary, these sub-tasks are crowdsourced to human contributors. The rest of the approach is handled by a computer algorithm. How the approach functions is tested by a laboratory experiment using a pilot system implemented as a web application.