OBJECTIVES: Since 2005 Austrian regional sickness funds have been facing an increase in drug prescription costs of 7-8 % per year. In order to analyze these costs we developed a benchmarking system comparing drug prescription costs for 2007 of the regional sickness funds including two ratios split into a price and a volume component. Calculating the ratios at the high aggregate of the second ATC level revealed problems: The total savings potential seemed unrealistically high and could not be traced back to a specific field of action. Therefore, the need for a re-design of the benchmarking system was more than necessary. METHODS: Since the cost-function is interpreted as a function of the distribution of prescriptions, price effects can be shown by adjusting the distribution. This is equivalent to shifting prescriptions from specific drugs to others within a cluster. To legitimize this new method, the volume component of each sickness fund is normalized to the average level of prescriptions per beneficiary (ATC level 2), considering the fund’s distribution of prescriptions in lower ATC levels. The process of clustering starts with substitutable drugs according to the national eco-list. After that  these clusters are combined to groups at first ATC level 5 and then  ATC level 4. Within these groups the distributions of the funds are adjusted to the benchmark’s distribution. RESULTS: In each step of the clustering process we used various approaches, like the use of the eco-list or substitution of drugs by similar agents, to affect the price component. Moreover, the new method evaluates the prescription behaviour of contract physicians for each fund. CONCLUSIONS: The original method shows that a savings potential cannot be calculated on an aggregate level. The new method provides a more detailed, step-by-step analysis and allows to identify  – combined with medical quality assurance – possible fields of action.