This article presents a methodology to measure relative industrial energy efficiency across plants within a manufacturing sector through the use of energy-production signatures (EPSs). Linear programming in combination with regression methods, benchmarking and simulation models are used to study the behavior of a representative manufacturing plant. The methodology is validated using data available from the Department of Energy database. The proposed EPSs can be used for proactive benchmarking and diagnostic purposes leading to improvements in energy consumption for individual companies. The results show that the proposed methodology successfully identifies energy and production inefficiencies within a manufacturing segment. Copyright (C) 2010 John Wiley & Sons, Ltd.