It is now well established that stellar photospheric compositions can act as a fossil record of the protostellar environments in which planets form.  Stellar spectroscopy of planet-hosting stars is therefore a valuable source of information about chemical conditions for planet formation.  However, the difficulty of resolving planet-scale abundance differences in the photospheres of stars requires a unique approach.  I will present recent results from a five-year-long radial velocity planet survey and complementary stellar spectroscopy focusing on solar twins.  By restricting our sample to a set of “twin” stars, we demonstrate that we can resolve changes in stellar compositions on the scale of only a few Earth masses, potentially offering a new observational window into planet formation signatures.  The techniques developed through our solar twin program will become increasingly valuable as we enter the TESS era of bright, spectroscopy-friendly planet host stars.