During a recent consulting project for a major women’s fashion accessory chain, the company tasked this researcher to evaluate the success of campaigns conducted during the previous year. However, this researcher was unable to provide any concrete conclusions due to the small size of the control customers excluded from the campaigns and the small number of customers offered the alternative test versions of the campaign offers.  Not only were the samples too small to determine any significance of difference between the groups, but there were also too few responders in the groups for the modeler to build a response model for the upcoming campaign. This inappropriate sizing, therefore, wasted the funds expended for the campaign versioning, missed the opportunity of income from the withheld control customers, and provided little or no insight into the appropriate campaigns for the present year’s campaigns. This paper examines the examines an appropriate control and test group sizing methodology to ensure desired degree of confidence in final results and includes desired degree of power for of tests to provide results accurate within specified ranges of error. The paper will present an Excel-based model enabling non-statistician business professionals to determine appropriate control and test group sizes to fit their particular business needs.