Ad-Spend Analytics

Companies allocate resources to marketing and advertising to expand market share, boosting sales, and optimizing profits. This research delves into advertisement spending, a significant corporate expense, to enhance profitability through the judicious allocation of advertising budgets. The objective of this research is to make use of historical data of individual companies to suggest the best forecasting technique. In addition, it also provides an ad-mix strategy for effective budget allocation in a multi-channel scenario. Polynomial Regression, Multiple Regression, and Multivariate LSTM are used to build the data-driven forecasting models. An interactive application is also designed, which includes a dashboard on the front end and models implemented in the backend. The experimentation suggests Multiple Regression as the best-performing strategy that covers 90.9% of the total number of companies, subjected to a MAPE of 30% accuracy. However, the experiments also observed that the performance of multivariate LSTM improves dramatically with the availability of more training data. This research provides valuable insights for companies seeking data-driven approaches to improve their advertising strategies and overall profitability.