We propose a new starting point for comparing planar dose distributions using a Jacobian-based measure. The measure is normalization independent, i.e., is sensitive to dose patterns while not being biased by global dose scaling. The measure is free of tunable parameters and is bounded between zero and one. We also propose an integrated version of the measure, normalized over the integrated area: the Jacobian index. We demonstrate analytically and verified numerically that the Jacobian index tends to a unique value of 2/pi when comparing unrelated dose distributions. We present a preliminary demonstration of the sensitivity and general characteristics of this measure. Examples of intensity modulated photon dose distributions from a linac as predicted with the convolution-superposition algorithm for various scenarios of irradiation mismatch and a film measured distribution are compared.