Deciding which ideas to pursue is an essential step in innovation management. Organizations increasingly open up their decision-making processes internally and externally by harnessing the collective intelligence of crowds. One mechanism for doing so is enterprise crow dfunding, i.e. inviting employees to propose and support ideas on a crowdfunding platform inside the enterprise. In this paper, we empirically analyze such an enterprise crowdfunding endeavor where hundreds of employees endowed with the discretion to spend company budget proposed dozens of ideas and decided to fund 10 of them. Based on log files and information on the employees' roles in their enterprise, we investigate the time course of individual decision-making - i.e., whether they decide rather quickly or more time-lagged which ideas to support Factors that influence decision-making style include characteristics of proposers, supporters, ideas, supporters' activity level, and the progress of the trial. From this, we derive suggestions for future research into crowdfunding, and we propose relevant design parameters for using crowdfunding as a tool for internal openness and Enterprise 2.0.