Previous research has indicated that groups can benefit from the combination of face-to-face and asynchronous computer-mediated communication for conducting work. This exploratory experiment compares the effectiveness of four different modes of communication for groups working on the upstream phases of software development: (I) face-to-face (2) synchronous computer conferencing (3) asynchronous computer conferencing (4) combined -- FtF and asynchronous computer conferencing. Teams of graduate students determined the requirements for an automated post office as a course assignment, over a two-week time period. The creativity and quality of solutions produced by groups in the combined condition were higher than those of the remaining three communication modes. Combined groups were generally more satisfied with their solutions, although no differences among conditions were found regarding satisfaction with the process used to accomplish work.