Suppose you are on a cross-functional team that is discussing how to allocate the costs of a purchasing department. One team member suggested that ‘number of purchase orders issued’ is the best cost driver to use as a cost-allocation base. However, a scatter graph of total costs versus number of purchase orders issued shows the following. Because the data clearly indicate that the single cost driver ‘number of purchase orders issued’ is not a reliable measure of the work done in the department (because there is too much scatter in the data), the team investigated further. It discovered that a significant amount of work of the purchasing department was certifying new vendors in addition to issuing purchase orders. What alternative method of allocation would you recommend?