To pool or not to pool in call centers

Should service capacities (such as agent groups in call centers) be pooled or not? This paper will show that there is no single answer. For the simple but generic situation of two (strictly pooled or unpooled) server groups, it will provide (1) insights and approximate formulae, (2) numerical support, and (3) general conclusions for the waiting-time effect of pooling. For a single call type, this effect is clearly positive, as represented by a pooling factor. With multiple job types, however, the effect is determined by both a pooling and a mix factor. Due to the mix factor, this effect might even be negative. In this case, it is also numerically illustrated that an improvement of both the unpooled and the strictly pooled scenario can be achieved by simple overflow or threshold scenarios. The results are of both practical and theoretical interest: practical for awareness of this negative effect, the numerical orders, and practical scenarios in call centers, and theoretical for further research in more complex situations.