Run a regression to determine the relationship between sales and promotional efforts, controlling for other variables that may affect sales ("controlling" just means you include these other variables in your regression