Many complex assembly lines, such as those in the automobile industry, have dozens or hundreds of stations that are affected by customer-selected options on the jobs being assembled. The various options often require significantly different amounts of processing time, and the role of assembly line sequencing in this context is to smooth out the flow of work to each station. However, most assembly line sequencing algorithms developed for such situations cannot consider so many stations or options effectively. In this paper, we develop an analytical method to compute a criticality index for each station, which can be used to determine which stations are most important to include in an assembly line sequencing algorithm. We report computational results using actual industry data which indicates that substantial improvements can be obtained by selecting stations based upon this criticality index.