Penerapan Strategi Pemasaran Digital B2B dengan Machine Learning untuk Mengurangi Ketergantungan pada Canvassing Tradisional dalam Industri Konstruksi

Hanjoyo, Taryono (2025) Penerapan Strategi Pemasaran Digital B2B dengan Machine Learning untuk Mengurangi Ketergantungan pada Canvassing Tradisional dalam Industri Konstruksi. Masters thesis, Institut Teknologi Sepuluh November.

Text 6032232050-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Ketergantungan terhadap metode canvassing tradisional dalam pemasaran B2B pada industri konstruksi menimbulkan biaya operasional tinggi, keterbatasan segmentasi, serta rendahnya akurasi prediksi prospek. Penelitian ini menerapkan strategi pemasaran digital berbasis Machine Learning untuk mengurangi ketergantungan tersebut dengan studi kasus pada PT Eraputra Rahayu Engineering, kontraktor mekanikal dan elektrikal yang menghadapi tantangan efisiensi akuisisi pelanggan. Pendekatan penelitian menggunakan metode Research and Development (R&D) dengan desain mixed methods yang menggabungkan data kualitatif dan kuantitatif. Data diperoleh dari wawancara semi-terstruktur, survei daring terhadap 20 klien B2B, serta digital log dari sistem Customer Relationship Management (CRM) dan Google Analytics. Algoritma K-Means digunakan untuk segmentasi pelanggan, sedangkan Random Forest diterapkan untuk prediksi konversi prospek, dan hasilnya dievaluasi menggunakan A/B testing serta quasi-experimental design. Hasil penelitian menunjukkan penerapan Machine Learning meningkatkan Efektivitas Pemasaran B2B dibandingkan canvassing manual. Model Random Forest mencapai akurasi 83,6% dan AUC 0,87 setelah penyeimbangan kelas (SMOTE), sedangkan penerapan K-Means menghasilkan tiga klaster prospek utama yang meningkatkan presisi segmentasi dan akurasi penargetan. Secara bisnis, Cost Per Lead (CPL) turun 28,4%, return on investment (ROI proxy) meningkat 31,2%, serta jumlah prospek berkualitas naik 42% selama periode uji. Secara kualitatif, implementasi strategi ini mengubah pola pengambilan keputusan dari berbasis intuisi menjadi berbasis data dan meningkatkan koordinasi antara tim digital marketing dan sales lapangan. Penelitian ini membuktikan bahwa strategi pemasaran digital berbasis Machine Learning efektif untuk meningkatkan efisiensi dan akurasi pemasaran B2B di industri konstruksi. Model integratif KMeans dan Random Forest dapat direplikasi oleh perusahaan sejenis guna memperkuat transformasi pemasaran berbasis data dan mengurangi ketergantungan terhadap canvassing tradisional.
=======================================================================================================================================
Reliance on traditional canvassing methods in B2B marketing within the construction industry results in high operational costs, limited customer segmentation, and low accuracy in prospect conversion prediction. This study applies a Machine Learning–based digital marketing strategy to reduce such dependency through a case study at PT Eraputra Rahayu Engineering, a mechanical and electrical contractor facing challenges in customer acquisition efficiency. The research adopts a Research and Development (R&D) approach with a mixed-methods design, combining qualitative and quantitative data. Data were collected through semi-structured interviews, an online survey of 20 B2B clients, and digital logs from the Customer Relationship Management (CRM) system and Google Analytics. The K-Means algorithm was employed for customer segmentation, while the Random Forest model was implemented to predict prospect conversion. The resulting strategy was evaluated using A/B testing and a quasiexperimental design. The findings indicate that the application of Machine Learning significantly improved the effectiveness of the digital marketing strategy compared to manual canvassing. The Random Forest model achieved an accuracy of 83.6% and an AUC of 0.87 after class balancing using SMOTE, while K-Means produced three major customer clusters that enhanced segmentation precision and targeting accuracy. From a business perspective, the Cost Per Lead (CPL) decreased by 28.4%, the return on investment (ROI proxy) increased by 31.2%, and the number of qualified leads rose by 42% during the testing period. Qualitatively, the implementation of this strategy transformed decision-making patterns from intuition-based to data-driven and improved coordination between the digital marketing and sales teams. This research demonstrates that Machine Learning– based digital marketing strategies are effective in improving efficiency and accuracy in B2B marketing within the construction industry. The integrated KMeans and Random Forest model can be replicated by similar companies to support data-driven marketing transformation and reduce reliance on traditional canvassing methods.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Canvassing, Industri Konstruksi, K-Means, Pemasaran Digital B2B, Random Forest, Construction Industry, B2B Digital Marketing
Subjects:	T Technology > T Technology (General) > T58.6 Management information systems
Divisions:	Interdisciplinary School of Management and Technology (SIMT) > 61101-Master of Technology Management (MMT)
Depositing User:	Taryono Hanjoyo
Date Deposited:	28 Jan 2026 09:52
Last Modified:	28 Jan 2026 09:52
URI:	http://repository.its.ac.id/id/eprint/130831

Actions (login required)

View Item