Coral geochemical climate reconstructions can extend our knowledge of global climate variability and trends over time scales longer than those of instrumental data. However, such reconstructions can be biased by coral growth and skeletal architecture, such as growth troughs, off‐axis corallite orientation, and changing growth direction. This study quantifies the impact of skeletal architecture and growth on geochemistry using measurements of coral skeletal density, extension rate, and calcification rate, and uses these metrics to improve paleoclimate reconstructions. We present paired geochemistry‐density records at Wolf Island, Galápagos, from three Porites lobata corals: two new paired density and geochemistry records from one fossil coral, and new density data from two previously published modern geochemistry records. We categorize each sampling transect used in this record by the quality of its orientation with respect to skeletal architecture. We observe relationships between geochemistry and density that are not detected using extension or calcification rate alone. These density‐geochemistry relationships likely reflect both the response of coral growth to environmental conditions and the nonclimatic impact of skeletal architecture on geochemistry in suboptimal sampling transects. Correlations of density with Sr/Ca, Ba/Ca, and Mg/Ca are consistent with the Rayleigh fractionation model of trace element incorporation into coral skeletons. Removing transects with suboptimal skeletal architecture increases mean reconstructed SST closer to instrumental mean SST, and lowers errors of reconstruction by up to 20%. These results demonstrate the usefulness of coral density data for assessing skeletal architecture and growth when generating coral paleoclimate records.