Systematic purchase behaviour in big transactional data: measurement innovations and applications
With the rapid advancement in technology and the emergence of new sources of data, consumer buying behaviours have become increasingly dynamic. This has created a growing need for new measures and models to understand and predict complex buying patterns, aiming to enhance customer experience, satisf...
| Main Author: | |
|---|---|
| Format: | Thesis (University of Nottingham only) |
| Language: | English |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://eprints.nottingham.ac.uk/80038/ |
| _version_ | 1848801156064608256 |
|---|---|
| author | Mansilla, Roberto |
| author_facet | Mansilla, Roberto |
| author_sort | Mansilla, Roberto |
| building | Nottingham Research Data Repository |
| collection | Online Access |
| description | With the rapid advancement in technology and the emergence of new sources of data, consumer buying behaviours have become increasingly dynamic. This has created a growing need for new measures and models to understand and predict complex buying patterns, aiming to enhance customer experience, satisfaction, and loyalty.
This thesis seeks to address this need through three interconnected studies that utilise a combination of traditional statistics and modern machine learning models and techniques. The objective is to explore systematic purchase behaviour (SPB) measurements and their real-world applications. Each study builds upon the previous one to provide a comprehensive understanding of SPB and its implications.
The primary objective is to develop a new measure of SPB by directly assessing the predictability aspect of basket composition. The thesis demonstrates the effectiveness of the proposed measure using both synthetic and real-world data sets. It also highlights the limitations of existing measures and introduces a new measure called Bundle Entropy (BE), which provides a precise indication of predictability, with zero denoting SPB and one indicating total unpredictability. The research also explores real-world applications of BE using two different large transactional datasets from leading UK retailers. The research delves into SPB at various levels of aggregation, providing novel insights into consumers' choices across different retail settings. The final aim of the research is to provide a comprehensive understanding of the main drivers of SPB by analysing variables from historical transactional, demographic, and psychographic data. Machine learning models and variable importance methods are used to understand the influence of each group of variables on SPB.
This research endeavours to advance our understanding of consumer behaviour dynamics and to explore the broader implications of its findings. It emphasises the significant potential of big transactional data, particularly individual loyalty card data, in various aspects of consumer research, including forecasting buying behaviour and explanatory modelling. Additionally, the thesis acknowledges the different limitations encountered within the studies and the common challenges presented by big data. It concludes by offering actionable recommendations and suggesting potential areas for future scholarly inquiry in this field, thereby contributing to the ongoing discourse on consumer behaviour research. |
| first_indexed | 2025-11-14T21:02:58Z |
| format | Thesis (University of Nottingham only) |
| id | nottingham-80038 |
| institution | University of Nottingham Malaysia Campus |
| institution_category | Local University |
| language | English |
| last_indexed | 2025-11-14T21:02:58Z |
| publishDate | 2024 |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | nottingham-800382024-12-12T04:40:14Z https://eprints.nottingham.ac.uk/80038/ Systematic purchase behaviour in big transactional data: measurement innovations and applications Mansilla, Roberto With the rapid advancement in technology and the emergence of new sources of data, consumer buying behaviours have become increasingly dynamic. This has created a growing need for new measures and models to understand and predict complex buying patterns, aiming to enhance customer experience, satisfaction, and loyalty. This thesis seeks to address this need through three interconnected studies that utilise a combination of traditional statistics and modern machine learning models and techniques. The objective is to explore systematic purchase behaviour (SPB) measurements and their real-world applications. Each study builds upon the previous one to provide a comprehensive understanding of SPB and its implications. The primary objective is to develop a new measure of SPB by directly assessing the predictability aspect of basket composition. The thesis demonstrates the effectiveness of the proposed measure using both synthetic and real-world data sets. It also highlights the limitations of existing measures and introduces a new measure called Bundle Entropy (BE), which provides a precise indication of predictability, with zero denoting SPB and one indicating total unpredictability. The research also explores real-world applications of BE using two different large transactional datasets from leading UK retailers. The research delves into SPB at various levels of aggregation, providing novel insights into consumers' choices across different retail settings. The final aim of the research is to provide a comprehensive understanding of the main drivers of SPB by analysing variables from historical transactional, demographic, and psychographic data. Machine learning models and variable importance methods are used to understand the influence of each group of variables on SPB. This research endeavours to advance our understanding of consumer behaviour dynamics and to explore the broader implications of its findings. It emphasises the significant potential of big transactional data, particularly individual loyalty card data, in various aspects of consumer research, including forecasting buying behaviour and explanatory modelling. Additionally, the thesis acknowledges the different limitations encountered within the studies and the common challenges presented by big data. It concludes by offering actionable recommendations and suggesting potential areas for future scholarly inquiry in this field, thereby contributing to the ongoing discourse on consumer behaviour research. 2024-12-12 Thesis (University of Nottingham only) NonPeerReviewed application/pdf en cc_by https://eprints.nottingham.ac.uk/80038/1/PhD_Thesis_Roberto_Mansilla_NUBS_Nov2024_with_viva_corrections.pdf Mansilla, Roberto (2024) Systematic purchase behaviour in big transactional data: measurement innovations and applications. PhD thesis, University of Nottingham. Consumer Behaviour; Systematic Purchase Behaviour; Transactional Data; Bundle Entropy |
| spellingShingle | Consumer Behaviour; Systematic Purchase Behaviour; Transactional Data; Bundle Entropy Mansilla, Roberto Systematic purchase behaviour in big transactional data: measurement innovations and applications |
| title | Systematic purchase behaviour in big transactional data: measurement innovations and applications |
| title_full | Systematic purchase behaviour in big transactional data: measurement innovations and applications |
| title_fullStr | Systematic purchase behaviour in big transactional data: measurement innovations and applications |
| title_full_unstemmed | Systematic purchase behaviour in big transactional data: measurement innovations and applications |
| title_short | Systematic purchase behaviour in big transactional data: measurement innovations and applications |
| title_sort | systematic purchase behaviour in big transactional data: measurement innovations and applications |
| topic | Consumer Behaviour; Systematic Purchase Behaviour; Transactional Data; Bundle Entropy |
| url | https://eprints.nottingham.ac.uk/80038/ |