A Machine learning and Empirical Bayesian Approach for Predictive Buying in B2B E-commerce

Aug 4, 2026

Upload by Maeve

Predictive modeling of buyer order placement is critical for sustainable growth in B2B e-commerce, especially in emerging markets where purchasing depends on trust, credit arrangements, and relationship building. The work targets telecaller-driven order placement by combining an ensemble of XGBoost with a modified Poisson-Gamma empirical Bayesian model and selecting relevant features to capture customer ordering dynamics. The proposed approach significantly increases customer order rates, demonstrating practical value for improved targeting and efficient resource utilization.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages8

Reading Time20 min

AccessFree to download

Abstract

Introduction

B2B e-commerce context and telecalling

Motivation for predictive buyer behavior

Prior work and purchasing patterns

A Machine learning and Empirical Bayesian Approach for Predictive Buying in B2B E-commerce Tuhin Subhra De∗ [tuhinsubhrade@iitkgp.ac.in](tuhinsubhrade@iitkgp.ac.in)[ ](tuhinsubhrade@iitkgp.ac.in)Indian Institute of Technology Kharagpur, West Bengal, India Pranjal Singh[pranjal.singh@udaan.com](pranjal.singh@udaan.com) Udaan Bangalore, Karnataka, India Alok Patel [alokpatel.a@udaan.com](alokpatel.a@udaan.com) Udaan Bangalore, Karnataka, India arXiv :2403 .07843v1 [ cs .LG] 12 Mar 2024 ABSTRACT In the context of developing nations like India, traditional businessto-business (B2B) commerce heavily relies on the establishment of robust relationships, trust, and credit arrangements between buyers and sellers. Consequently, e-commerce enterprises frequently employ telecallers to cultivate buyer relationships, streamline order placement procedures, and promote special promotions. The accurate anticipation of buyer order placement behavior emerges as a pivotal factor for attaining sustainable growth, heightening competitiveness, and optimizing the efficiency of these telecallers. To address this challenge, we have employed an ensemble approach comprising XGBoost and a modified version of Poisson Gamma model to predict customer order patterns with precision. This paper provides an in-depth exploration of the strategic fusion of machine learning and an empirical Bayesian approach, bolstered by the judicious selection of pertinent features. This innovative approach has yielded a remarkable 3 times increase in customer order rates, showcasing its potential for transformative impact in the e-commerce industry. CCS CONCEPTS • Theory of computation → Bayesian analysis; • Applied computing → E-commerce infrastructure; • Computing methodologies → Classification and regression trees. KEYWORDS Personalization, E-commerce, Poisson-Gamma Model, XGBoost ACM Reference Format: Tuhin Subhra De, Pranjal Singh, and Alok Patel. 2024. A Machine learning and Empirical Bayesian Approach for Predictive Buying in B2B E-commerce. In 2024 The 8th International Conference on Machine Learning and Soft Computing (ICMLSC 2024), January 26–28, 2024, Singapore, Singapore. ACM, New York, NY, USA, 8 pages. [https://doi.org/10.1145/3647750.3647754](https://doi.org/10.1145/3647750.3647754) ∗Work performed while [interning at Udaan.com](interning at Udaan.com), a B2B e-commerce company Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than the author(s) must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission [and/or a fee. Request permissions from permissions@acm.org](and/or a fee. Request permissions from permissions@acm.org). ICMLSC 2024, January 26–28, 2024, Singapore, Singapore © 2024 Copyright held by the owner/author(s) . Publication rights licensed to ACM. ACM ISBN 979-8-4007-1654-6/24/01. . . $15.00 [https://doi.org/10.1145/3647750.3647754](https://doi.org/10.1145/3647750.3647754) 1 INTRODUCTION Established in 2016 with a vision to revolutionize trade in India through technology, Udaan is the country’s largest business-tobusiness (B2B) e-commerce platform. Udaan operates across diverse product categories, including lifestyle, electronics, home & kitchen, staples, fruits and vegetables, FMCG, pharma, and general merchandise. With a network of over 3 million registered users and 25,000-30,000 vendors and sellers spanning across 900+ cities in India and encompassing more than 12,000 pin codes, Udaan facilitates over 4.5 million transactions per month. In the traditional B2B environment in India, purchasers are situated in both major

FAQs

Why is predicting buyer order placement important in B2B e-commerce?

It supports sustainable growth and competitiveness by improving the efficiency of telecaller efforts and targeting customers with a higher likelihood of future purchases.

What modeling approach is used to predict customer order patterns?

An ensemble combining XGBoost and a modified Poisson-Gamma empirical Bayesian model is used, along with judicious feature selection.

How does the proposed method improve outcomes for the business?

It yields a reported threefold increase in customer order rates, indicating stronger predictive performance for order placement behavior.

Aug 4, 2026

Upload by Maeve