Firstly, during data maintenance the focus is on the most critical data (i.e. the ones with the highest payoff per resources spent) before moving on to less critical ones. This implies that the first work of assuring data quality would have the greatest effect, i.e. the costs inflicted by poor quality data decreases exponentially. The second assumption is that the costs of the efforts to ensure high data quality are not causally related to the their importance, i.e. focusing on a set of poor quality data with great impact on costs is not necessarily cheaper than focusing on data with little impact on costs. Thus, the costs of assuring data quality is a linear relationship between data quality and assurance costs.