This paper considers economic models for the selection of target mean and variance of a process that has both lower and upper specification limits. When the product quality characteristic falls within these limits, an item is classified as good, or non-defective. The problem of finding the target mean is solved by first making the assumption that variance is known for different conditions. The Taguchi loss function is then incorporated, and variance is optimized for minimum cost, based on the expected loss function which assumes the mean is on target. The paper suggests that the process is first optimized for the mean with a given variance, subsequently optimized for the variance with a given mean, and this process is repeated until their optimal values converge.