Data-Driven Decision Making for Online Retail Operations in Sourcing, Distribution and Assortment Personalization

In the current world, online retailing accounts for the major share of retailing due to the unique benefits it provides to customers. However, its success is highly dependent on how it manages its operations. Compared to traditional retailing, online retailing requires new decision-making models to support its operations. In this thesis, the E-SMART framework is proposed which assists e-tailers in operations management in the three areas of the sourcing, distribution, and personalization of an assortment. The E-SMART framework consists of three modules. The first module manages sourcing decisions and develops a framework to assist the e-tailer to decide if it should interact with a powerful main brand supplier or instead, procure items from an alternate supplier. The framework determines the best strategy, taking into consideration selling price, assortment size and the inventory level of the item to be procured. It also considers the importance of considering customers as shopping basket consumers in sourcing decisions in different scenarios. The second module manages distribution decisions and develops a decision-making model to determine the most profitable strategy from drop shipping and batch ordering. The interaction of the supplier and the e-tailer is modeled using the game-theory approach. Another important aspect that needs to be considered carefully in the drop-shipping model is the profit-sharing ratio between the supplier and the e-tailer. By analyzing different scenarios, the model shows which strategy is more profitable for the e-tailer. The third module of the E-SMART framework assists the e-tailer in dynamic assortment personalization to improve its sales. The problem is modeled as a Markov decision process to mitigate customers' churn rate in the case of demand uncertainty and limited inventory. Customer choice and customer lifetime are modeled using the multinomial logit model and survival analysis techniques respectively, and are estimated using historical sales data. Following this, personalized assortment planning to increase the e-tailer’s profit and reduce customer churn is proposed. The results of each module of E-SMART are demonstrated in different scenarios to show how the proposed model is applicable in real-world practice.

Grantor dc:publisher

UNSW, Sydney

Year dc:date

2021

Author dc:creator

• Saberi, Zahra

• Operations management
• Online retailer
• Data driven decision making model
• Assortment personalization
• Sourcing
• Distribution

Statement dc:rights

• open access
• CC BY-NC-ND 3.0
• free_to_read

Licence

• https://purl.org/coar/access_right/c_abf2
• https://creativecommons.org/licenses/by-nc-nd/3.0/au/

Language dc:language

EN

Identifier

https://doi.org/10.26190/unsworks/2315

OAI identifier oai:identifier

oai:unsworks.library.unsw.edu.au:1959.4/70995