Predicting sales with machine learning solutions and demonstrating their potential business value: A case study for an after-sales department of a B2B industrial company

No Thumbnail Available

URL

Journal Title

Journal ISSN

Volume Title

School of Business | Master's thesis

Date

2022

Major/Subject

Mcode

Degree programme

Information and Service Management (ISM)

Language

en

Pages

88 + 4

Series

Abstract

Organizations of today are eager to develop themselves to become more data-driven in their decision-making, and one method to improve the data-driven capabilities of a business is to implement machine learning solutions. However, developing and implementing machine learning solutions into business processes is a difficult activity, especially when the target is to do it in a way that provides concrete business value for the organization. The target of this thesis was to develop a machine learning solution and to find out if it could predict the outcome of the spare part offers sent by a B2B industrial company’s after-sales department. And more importantly, the target of this thesis was to discover what kind of business value such a machine learning solution could offer to the company and its after-sales department. As the company department was interested in the customer insights machine learning and data analytics could provide, the third target of this thesis was to identify customer features that predispose to certain purchasing behavior. The best performing machine learning solution was capable of predicting the outcome of spare part offers with an Area Under the Curve (AUC) score of 73% and an accuracy of 67%, meaning the developed ML model offered clear prediction power, especially for probabilistic prediction. The built machine learning models were utilized for three different business implications that expressed the business value the machine learning solutions can add to the business processes of the after-sales department. In addition, numerous data-driven business insights about customer purchasing behavior were found and presented in this thesis.

Description

Thesis advisor

Malo, Pekka
Kuosmanen, Timo

Keywords

machine learning, LightGBM, CatBoost, data-driven decision-making, B2B sales, after-sales

Other note

Citation