<![CDATA[The growth of digital advertising spending demands more accurate measurement of the effectiveness of multi-platform strategies to optimize the distribution of advertising budgets across various advertising platforms. This study analyzes the impact of the addition of Instagram on the performance of herbal product advertising campaigns previously run solely through Facebook on the Meta Ads platform. The research approach consists of two stages, namely building a TabTransformer-based prediction model using Facebook pre-intervention data and a Bayesian Structural Time Series (BSTS)-based Causal Impact analysis, utilizing TabTransformer prediction results and advertising budget variables as covariates in the construction of counterfactuals. The evaluation results showed the best TabTransformer model was obtained in the second scenario with an MSE value of 1.84, an MAE of 0.95, and an R² of 80%, indicating good predictive performance on the test data. Furthermore, the results of the causal impact analysis showed that the addition of Instagram had a statistically significant positive impact, with an average increase in the number of leads of approximately 58.6%, or the equivalent of 93.21 leads per day in the post-intervention period.]]>
Copyrights © 2026
