A Credibility Model Based on Influencer Sentiment Analysis and User Reviews on Digital Investment Platforms
DOI:
https://doi.org/10.33005/ic-ebgc.v9i2.201Keywords:
digital credibility, sentiment analysis, influencer, collective reviewer, IndoBERT, retail investment, Source Credibility Theory, dual-source modelAbstract
This study develops a digital credibility model in the context of Indonesian retail investment by integrating two sources that shape public perception, namely social media influencers and collective reviewers on digital platforms. Unlike conventional approaches that position influencers as the dominant source of credibility and reviewers as mere electronic word of mouth, this study reconstructs both as equal and complementary credibility entities. The study employs a quantitative approach based on sentiment analysis of 57,808 Google Play Store review data points and TikTok comments from three major investment platforms in Indonesia — Stockbit, Bibit, and Bareksa — covering the period 2025 to 2026. Data were analyzed using a fine-tuned IndoBERT model, BiLSTM with an attention mechanism, and four classical machine learning models with class imbalance handling techniques based on SMOTE and class weighting. Indonesia's digital investment ecosystem during this period was also characterized by dual validation from the Indonesia Stock Exchange (IDX) through JATS system connectivity and KSEI recording, as well as recognition by TradingView as a global analysis platform that listed all three applications among its Preferred Brokers/Platforms in Southeast Asia. On the marketing side, all three platforms leverage a diverse spectrum of influencers, ranging from macro-influencer entertainment figures such as Raditya Dika, Vidi Aldiano, and Maudy Ayunda, to niche TikTok influencers targeting specific communities. Evaluation was conducted using F1-macro, precision, recall, and ROC-AUC metrics. Results indicate that influencer sentiment tends to build emotional and aspirational perceptions, while collective reviewers form rational evaluations based on real experiences — including responses to specific features such as Bareksa Emas and Robo Advisor. The integration of both sources produces a dynamic Dual-Source Digital Credibility Model (DSDCM). These findings extend the Source Credibility Theory framework to a platform-based digital ecosystem and offer practical implications for investment platform management in designing marketing communication strategies aligned with actual user experiences.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Proceedings of International Conference on Economics Business and Government Challenges

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

