Pengukuran dan Optimasi Efektivitas Iklan Menggunakan Bayesian Marketing Mix Modeling dan Bayesian Optimization

Fajar Legianto Asmin; Alz Danny Wowor

doi:10.36080/idealis.v9i2.3753

Authors

Fajar Legianto Asmin Teknik Informatika, Fakultas Teknologi Informasi, Universitas Kristen Satya Wacana, Salatiga, Indonesia
Alz Danny Wowor Teknik Informatika, Fakultas Teknologi Informasi, Universitas Kristen Satya Wacana, Salatiga, Indonesia

DOI:

https://doi.org/10.36080/idealis.v9i2.3753

Keywords:

Advertising Effectiveness, Bayesian Marketing Mix Modeling, Bayesian Optimization, Budget Allocation, Digital Marketing

Abstract

The rapid growth of digital advertising has intensified the need for accurate measurement of advertising effectiveness and efficient budget allocation. However, many businesses still struggle to quantitatively link marketing efforts to sales outcomes due to data uncertainty, channel interactions, and dynamic consumer behavior. This study proposes an integrated framework that combines Bayesian Marketing Mix Modeling (BMMM) with Bayesian Optimization (BO) to simultaneously measure advertising contribution and optimize budget allocation under uncertainty. The proposed framework was applied to a dataset of 456 daily TikTok advertising records collected from January 2024 to March 2025. Bayesian Ridge Regression was employed for model estimation and evaluated using 5-Fold Cross-Validation, yielding strong predictive performance with an R² of 0.8649 and MAPE of 0.55%. Optimization results indicate that Visual Shopping Ads (VSA) and Live Shopping Ads (LSA) deliver the highest marginal returns on sales, while Product Showcase Ads (PSA), Awareness, and GMV Max exhibit relatively lower contributions. Unlike previous studies that typically address marketing mix modeling or budget optimization separately, this research introduces integrated probabilistic framework that simultaneously measures advertising effectiveness and optimizes budget allocation by fully utilizing posterior distributions. This approach offers a more robust alternative to traditional Marketing Mix Modeling and manual allocation methods, particularly in fast-changing digital advertising environments such as TikTok Ads. The findings provide practical guidance for retailers in making data-driven marketing decisions and demonstrate the significant potential of Bayesian approaches in digital marketing analytics, especially for emerging markets like Indonesia.

Downloads

Download data is not yet available.

References

[1] S. Kemp, “Digital 2025: Global Overview Report,” Feb. 2025. [Online]. Available: https://datareportal.com/reports/digital-2025-global-overview-report

[2] Dentsu, “Global Ad Spend Forecasts December 2025,” Dec. 2025. [Online]. Available: https://insight.dentsu.com/ad-spend-dec-2025/

[3] Statista, “Advertising revenue of Google from 2001 to 2024,” Statista. [Online]. Available: https://www.statista.com/statistics/266249/advertising-revenue-of-google/?srsltid=AfmBOoohwTTQTfbuXXnGx2f_f4iBtaKQxpj4_znE-OOLyCg2WN_d3PbB

[4] eMarketer, “Worldwide Ad Spending Forecast 2025,” eMarketer. [Online]. Available: https://www.emarketer.com/content/worldwide-ad-spending-forecast-2025#page-report

[5] B. R. Gordon, F. Zettelmeyer, N. Bhargava, and D. Chapsky, “A Comparison of Approaches to Advertising Measurement: Evidence from Big Field Experiments at Facebook,” Marketing Science, vol. 38, no. 2, pp. 193–225, Mar. 2019, doi: 10.1287/mksc.2018.1135.

[6] J. T. Yun, C. M. Segijn, S. Pearson, E. C. Malthouse, J. A. Konstan, and V. Shankar, “Challenges and Future Directions of Computational Advertising Measurement Systems,” J. Advert., vol. 49, no. 4, pp. 446–458, 2020, doi: 10.1080/00913367.2020.1795757.

[7] Y. Jin, Y. Wang, Y. Sun, D. Chan, and J. Koehler, “Bayesian Methods for Media Mix Modeling with Carryover and Shape Effects,” Google Research, Apr. 2017.

[8] H. Chen, M. Zhang, L. Han, and A. Lim, “Hierarchical marketing mix models with sign constraints,” J. Appl. Stat., vol. 48, no. 13–15, pp. 2944–2960, 2021, doi: 10.1080/02664763.2021.1946020.

[9] X. Ni, Y. Zhang, Y. Pu, M. Wei, and Q. Lou, “A Personalized Causal Inference Framework for Media Effectiveness Using Hierarchical Bayesian Market Mix Models,” Journal of Advanced Computing Systems (JACS), vol. 3, no. 9, pp. 9–23, 2023.

[10] A. Jha, A. Bhatia, and K. Tiwari, “Bayesian Deep Learning Meets Self-Attention: A Risk-Aware Approach to Advertisement Optimization,” IEEE Access, vol. 13, pp. 86413–86429, 2025, doi: 10.1109/ACCESS.2025.3570537.

[11] K. H. Brodersen, F. Gallusser, J. Koehler, N. Remy, and S. L. Scott, “Inferring causal impact using Bayesian structural time-series models,” Ann. Appl. Stat., vol. 9, no. 1, 2015, doi: 10.1214/14-AOAS788.

[12] C. Gong et al., “CausalMMM: Learning Causal Structure for Marketing Mix Modeling,” in Proceedings of the 17th ACM International Conference on Web Search and Data Mining, New York, NY, USA: ACM, 2024, pp. 238–246. doi: 10.1145/3616855.3635766.

[13] M. Gigli and F. Stella, “Multi-armed bandits for performance marketing,” Int. J. Data Sci. Anal., vol. 20, no. 1, pp. 151–165, Jun. 2025, doi: 10.1007/s41060-023-00493-7.

[14] W. Gao, X. Meng, and Y. Zhang, “Multimodal Reinforcement Learning for Dynamic Cross-Media Advertising Budget Allocation Via DDPG and PPO Combined with Meta-Learning and Adversarial Training,” Informatica, vol. 49, no. 30, 2025, doi: 10.31449/inf.v49i30.11440.

[15] M. Yang, Q. Cao, L. Tong, and J. Shi, “Reinforcement learning-based optimization strategy for online advertising budget allocation,” in 2025 4th International Conference on Artificial Intelligence, Internet and Digital Economy (ICAID), IEEE, Apr. 2025, pp. 115–118. doi: 10.1109/ICAID65275.2025.11034517.

[16] Z. Pan, “A Reinforcement Learning Approach for Adaptive Budget Allocation in Pharmaceutical Digital Marketing: Maximizing ROI Across Patient Journey Touchpoints,” Journal of Sustainability, Policy, and Practice, vol. 1, no. 4, 2025.

[17] H. Wang et al., “ROI-Constrained Bidding via Curriculum-Guided Bayesian Reinforcement Learning,” in Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, New York, NY, USA: ACM, 2022, pp. 4021–4031. doi: 10.1145/3534678.3539211.

[18] Y. Wang, “Application of Bayesian Optimization in Periodic Delivery,” in Proceedings of the 2025 6th International Conference on Computer Information and Big Data Applications, New York, NY, USA, 2025, pp. 556–561. doi: 10.1145/3746709.3746805.

[19] R. A. Astudillo, D. R. Jiang, M. Balandat, E. Bakshy, and P. I. Frazier, “Multi-Step Budgeted Bayesian Optimization with Unknown Evaluation Costs,” in 35th Conference on Neural Information Processing Systems (NeurIPS 2021), Nov. 2021.

[20] S. Daulton, M. Balandat, and E. Bakshy, “Differentiable expected hypervolume improvement for parallel multi-objective Bayesian optimization,” in NIPS’20: Proceedings of the 34th International Conference on Neural Information Processing Systems, Vancouver: Curran Associates Inc., Dec. 2020, pp. 9851–9864.

[21] I. R. Arief and I. Novita, “Penerapan E-Commerce Untuk Mengoptimalkan Penjualan Pada Toko Gita Giza,” IDEALIS: Indonesia Journal Information System, vol. 4, no. 1, pp. 127–136, 2021, doi: 10.36080/idealis.v4i1.2823.

[22] F. R. R. Santoso and E. I. H. Ujianto, “A decision support system for the determination of additional stock items using the Topsis method based on Android,” IDEALIS: Indonesia Journal Information System, vol. 8, no. 1, pp. 42–52, 2025, doi: 10.36080/idealis.v8i2.3339.

[23] R. A. Zilasar, Y. A. Pranoto, and F. X. Ariwibisono, “Sistem Pendukung Keputusan Pemilihan Produk Kaca Dan Cermin Menggunakan Metode Fuzzy AHP,” IDEALIS: Indonesia Journal Information System, vol. 9, no. 1, pp. 21–29, 2026, doi: 10.36080/idealis.v9i1.3617.

[24] S. N. Windrasari, H. Margono, and Y. A. N. S. Putra, “Predictive Sales Analysis in Coffee Shops Using the Random Forest Algorithm,” MALCOM: Indonesian Journal of Machine Learning and Computer Science, vol. 5, no. 3, pp. 1000–1011, 2025, doi: 10.57152/malcom.v5i3.2023.