Multiple filtering criteria derived from consumer surveys can be used to adjust an advertising schedule to achieve improved performance. Historical performance data related to performance factors, such as gross rating points, can be filtered using the multiple filtering criteria. The filtered data can then be used as the basis for predicting future performance of an advertisement or advertising campaign. The historical data can be updated as new performance data becomes available, and the predicted future performance updated. An advertising schedule can be adjusted to maintain performance factors within designated upper and lower limits.