Gravitational lensing magnifies the luminosity of galaxies behind the lens. We use this effect to constrain the total mass in the cluster Abell 1689 by comparing the lensed luminosities of background galaxies with the luminosity function of an undistorted field. Since galaxies are assumed to be a random sampling of luminosity space, this method is not limited by clustering noise. We use photometric redshift information to estimate galaxy distance and intrinsic luminosity. Knowing the redshift distribution of the background population allows us to lift the mass/background degeneracy common to lensing analysis. In this paper we use 9 filters observed over 12 hours with the Calar Alto 3.5m telescope to determine the redshifts of 1000 galaxies in the field of Abell 1689. Using a complete sample of 151 background galaxies we measure the cluster mass profile. We find that the total projected mass interior to 0.25h^(-1)Mpc is (0.48 +/- 0.16) * 10^(15)h^(-1) solar masses, where our error budget includes uncertainties from the photometric redshift determination, the uncertainty in the off-set calibration and finite sampling. This result is in good agreement with that found by number count and shear-based methods and provides a new and independent method to determine cluster masses. ; Comment: 13 pages, 10 figures. Submitted to MNRAS (10/99); Replacement with 1 page extra text inc. new section, accepted by MNRAS