A notion of scaled consensus is defined, wherein network components' states approach dictated ratios in the asymptote. A linear network dynamics that achieves scaled consensus for a prescribed interaction graph is introduced, and an equivalence with the class of (negative) singular M-matrices is explored. A few further characterizations of the scaled synchronization process are given, and finally a modification that allows tracking on the stable manifold is presented. (C) 2014 Elsevier Ltd. All rights reserved.