The empirical analysis in this paper uses individual level advertising, page-log, and transaction data provided by a Chinese health insurance firm under a nondisclosure agreement that restricts public release of the raw data. Consistent with the Marketing Science replication and disclosure policy, the authors have prepared a replication package comprising deidentified random subsets of the user-level data, the corresponding ad-day level aggregated data, and the full set of estimation scripts required to reproduce the tables and figures reported in the manuscript and appendix. Public disclosure of these materials is authorized in writing by the data provider. The package includes a README documenting the sampling procedure, the package contents, and step-by-step reproduction instructions; an renv lockfile and session information ensuring environment reproducibility; and a validation report comparing the replicated estimates against the values reported in the manuscript. The four authors contributed equally. The authors declare no competing interests.