Purpose - The purpose of this paper is to propose a straightforward model for selecting slightly non-homogeneous vendors. Design/methodology/approach - In this paper the use of the interval data envelopment analysis (DEA) is suggested. The bounds of intervals are constant and can be obtained by various estimation techniques. The interval DEA model provides for the decision making units (DMUs) with missing values a lower and an upper bound of their efficiency score corresponding to their most favorable and unfavorable option. Findings - Employing the proposed method for selecting slightly non-homogeneous vendors largely reduced practical difficulties for vendor selection. This method does not exclude any vendor from the selection problem. For all the vendors it provides bounds of the efficiency scores depended on the particular data values that the vendors with missing data assign within the intervals so to maximize their efficiency score. Practical implications - The proposed model considers a slightly non-homogeneous situation for vendor selection. The proposed approach is driven by multiple criteria. The joint consideration of multiple criteria in a slightly non-homogeneous environment helps managers select vendors using a comprehensive approach that goes beyond just purchase costs. Originality/value - This paper is believed to be the first to discuss the problem of slightly non-homogeneous vendor selection with respect to interval mathematics.