The value efficiency approach is one possible way of incorporating information preferences into performance analysis of Decision-Making Units (DMUs). In this paper, we propose a novel geometric interpretation of value efficiency while plugging it into radial and non-radial DEA (Data Envelopment Analysis) models under the assumption of variable returns to scale. In addition, this novel geometric interpretation of value efficiency is extended to Additive Slacks-Based Measure (ASBM) modeling. This is achieved by linearization of non-radial DEA models using multi-objective programming. Performance of such proposed approaches in terms of reliability and discriminatory power are compared through a case study involving Finish bank branches. Research implications are then derived and conclusions are drawn.