Optimal Advertisement Strategies for Small and Big Companies
Many small and big companies in developing countries struggle to make their products or services known to the public. This is especially the case when there are new or have a new product. Most of them use publicity through radio, tv, social networks, billboard, SMS... Moreover, they also need to decide at what time to display their publicity for maximal effects. The companies which have more money typically used a simple strategy which consists in doing the publicity at many sources at different time or at a time such as to maximize the number of viewers. The smaller ones typically target the best popular programs.
However, this strategy is not the best as many users listening to your publicity might not be interested in it. So, you are more likely to miss the interested readers. Moreover, there will be many other competing publicities.
We propose a strategy by using the Multi-Armed bandit problem to optimally solve this problem under realistic assumptions. We further extend the model to deal with many competing companies by proposing the use of a time-division sharing algorithm.
KeywordsBandit algorithm Advertisement Developing countries
- 1.Alon, N., Gamzu, I., Tennenholtz, M.: Optimizing budget allocation among channels and influencers. In: Proceedings of the 21st International Conference on World Wide Web, WWW 2012, pp. 381–388 (2012)Google Scholar
- 7.Maehara, T., Yabe, A., Kawarabayashi, K.: Budget allocation problem with multiple advertisers: a game theoretic view. In: ICML 32, JMLR Proceedings, vol. 37, pp. 428–437. JMLR.org (2015)Google Scholar
- 8.Miyauchi, A., Iwamasa, Y., Fukunaga, T., Kakimura, N.: Threshold influence model for allocating advertising budgets. In: ICML 32, JMLR Proceedings, vol. 37, pp. 1395–1404. JMLR.org (2015)Google Scholar
- 9.Pandey, S., Olston, C.: Handling advertisements of unknown quality in search advertising. In: NIPS 20, pp. 1065–1072. MIT Press (2006)Google Scholar
- 10.Radanovic, G., Faltings, B.: A robust Bayesian truth serum for non-binary signals. In: AAAI 27, pp. 833–839 (2013)Google Scholar