We analyze the metal accumulation in dwarf and spiral galaxies by following the history of metal enrichment and outflows in a suite of 20 high-resolution simulated galaxies. These simulations agree with the observed stellar and gas-phase mass–metallicity relation, an agreement that relies on large fractions of the produced metals escaping into the circumgalactic media. For instance, in galaxies with Mvir ∼ 109.5–1010 M, we find that about ∼85% of the available metals are outside of the galactic disk at z = 0, although the fraction decreases to a little less than half in Milky-Way-mass galaxies. In many cases, these metals are spread far beyond the virial radius. We analyze the metal deficit within the ISM and stars in the context of previous work tracking the inflow and outflow of baryons. Outflows are prevalent across the entire mass range, as is reaccretion. We find that between 40% and 80% of all metals removed from the galactic disk are later reaccreted. The outflows themselves are metal-enriched relative to the ISM by a factor of 0.2 dex because of the correspondence between sites of metal enrichment and outflows. As a result, the metal mass loading factor scales as v metals circ 0.91 h µ - , a somewhat shallower scaling than the total mass loading factor. We analyze the simulated galaxies within the context of analytic chemical evolution models by determining their net metal expulsion efficiencies, which encapsulate the rates of metal loss and reaccretion. We discuss these results in light of the inflow and outflow properties necessary for reproducing the mass–metallicity relation.