Enabling Dynamic Pricing of Perishables in Grocery Stores: Identifying Batches for Markdown
Loading...
URL
Journal Title
Journal ISSN
Volume Title
Perustieteiden korkeakoulu |
Master's thesis
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.
Authors
Date
2021-08-25
Department
Major/Subject
Operations and Service Management
Mcode
SCI3049
Degree programme
Master’s Programme in Industrial Engineering and Management
Language
en
Pages
63 + 16
Series
Abstract
The current inventory replenishment practice FIFO is widely adopted from all grocery retailers in the world. By applying FIFO, the older batch is placed in front on shelf. This practice aims to direct consumer purchase to the older batch. However, the appearance of multiple batches of the same product on shelf and the tendency to choose the freshest or newest batch make this current practice ineffective. In order to encourage consumer to purchase the older or near-expired batch, dynamic pricing, especially markdown pricing can be seen as a potential solution to the problem. Although dynamic pricing has been well studied in literature, the main focus is based on SKU level. This shows a research gap for the application of dynamic pricing at batch level. The current set up in information system and product identification technology at grocery retailers does not allow for an efficient and effective dynamic pricing. The main reason is because it is impossible to identify batches of the same product with different remaining shelf life. The proposed solution in this thesis is in a form of a 5-step operational process. This end-to-end process aims to automatically identify available batches of every SKU on shelf and be able to generate the probabilistic forecast for the amount of spoilage. At the end of the process, a new invented policy called (SR, RSL, DP) is introduced and help to categorize batches to different risk profile. Based on this classification, different discount % can be set corresponding to each risk profile. Dynamic pricing and batch-level perishable products identification are still uncommon from current practice in grocery retailers. The demonstration and result of the Bayesian Poisson regression model in proposed operational process has proven good performance in terms of forecasting accuracy and bias. In addition, the design of the operational process is feasible and can be implemented to current real-life situation. Therefore, this thesis not only contributes to current dynamic pricing and product identification literatures, but also help to improve the current solution in practice.Description
Supervisor
Holmström, JanThesis advisor
Loikkanen, LauriKeywords
dynamic pricing, Bayesian, poisson regression, perishables, risky batches, design science