According to the psychology of waiting lines，a good queuing management policy can largely improve the customer satisfaction with waiting．Inspired by that，this paper compares two kinds of queuing management policies alternatively applied by a capital-constrained service system to improve customer satisfaction with waiting: One is to offer extra service to the customers ranked less than N in queue ( policy H) ，and the other is to offer extra service to the customers ranked over N ( policy T) ．In this paper，we model the service system as an M/M/1 queuing system and give the optimal design for each kind of policy．Our numerical examples show that the performances of these two kinds of policies vary significantly under different budgets and different service loads．When the budget is large，if the service load is low or the extra service makes a large difference to customer feeling of delay，policy H dominates policy T．When the budget is small and the service load is low，policy T dominates policy H．Finally，when the budget is small and the extra service makes a small difference，the effects of both policies are not significant．