The concept of competition has changed from 'competition among companies' to 'competition among supply chains' recently. Therefore, it is necessary to determine safety stock levels through scientific methods by considering customer service level for all companies of a supply chain. In the literature, the mean demand is used to calculate safety stock levels for a specified customer service level. In this study, a rule-based safety stock methodology, considering the order quantity changes immediately before the delivery date of the customer, has been proposed as different from literature. This methodology based on ABC-XYZ analysis provides to classify products according to monetary value and order changes and to propose a safety stock level for each product group in ABC-XYZ analysis. The main motivation is to provide an optimal combination of customer service and stock levels. The proposed methodology has been applied to a company operating in the automotive sector with 1203 products. The safety stock levels were determined according to previous year data and policies on the basis of different behaviors of customers and also the sales.