Reducing shoe returns
ICT & Artificial Intelligence
Client company:Van Gastel Shoes
Project description
The main research question of the project is: "Which product characteristics are most closely associated with high rates of returns, and how can AI be used to address these issues?"
This research question seeks to identify the specific attributes of products that are strongly correlated with high return rates. By analyzing various product characteristics, such as material, size, and style, the project aims to determine which factors contribute most significantly to returns.
Furthermore, the question explores how artificial intelligence (AI) can be utilized to address these issues. The project aims to develop AI models and techniques that can predict return occurrences based on the identified product characteristics. By leveraging AI, the project seeks to provide Van Gastel Shoes with actionable insights, enabling them to take proactive measures to reduce return rates. These measures could include improving the quality of certain products, adjusting pricing strategies, or implementing targeted marketing campaigns.
In summary, the main research question focuses on understanding the relationship between product characteristics and return rates, while also exploring the potential of AI to address these issues and assist in optimizing the company's operations.
Context
The context for this project revolves around Van Gastel Shoes, a company experiencing a high number of product returns. Product returns can be a significant challenge for businesses, impacting profitability, customer satisfaction, and environmental sustainability.
The project aims to leverage AI techniques to analyze and understand the factors contributing to high return rates. By examining various product characteristics, such as material, size, and style, the team aims to identify which attributes are closely associated with returns. This analysis will provide insights into the specific areas that need improvement or optimization to reduce return rates.
The project also takes into consideration the environmental impact of product returns, including the associated carbon dioxide (CO2) emissions, unnecessary kilometers driven for returns, and road safety concerns in residential areas. By addressing return rates, the project not only aims to increase Van Gastel Shoes' profitability but also to minimize their carbon footprint and promote sustainable practices.
Results
In this project, several conclusions were drawn based on the analysis and research conducted. One of the key findings is that almost every shoe in Van Gastel Shoes' inventory has a return rate of 50%, indicating a high proportion of returns. This has significant implications for the environment, as it leads to a substantial number of shoes being returned unnecessarily.
Further analysis revealed that certain types of shoes have higher return rates compared to others. Specifically, children's shoes were found to be more frequently returned than adult shoes. Additionally, sandals had a notably high percentage of returns, particularly during the summer season.
Another interesting insight was the behavior of some customers who purchase multiple pairs of shoes but return the majority of their order, keeping only one pair. This behavior contributes to the high return rates experienced by Van Gastel Shoes.
To address these issues, the project team developed models that predict the return rate percentage of shoes. By utilizing traditional machine learning model, for example: RandomForest & Catboost. Also we used neural networks to predict the return percentages. We also utilized Convolutional Neural Networks (CNN), we were able to predict product characteristics related to returns. These predictions can help Van Gastel Shoes identify specific shoes that are likely to be returned and take proactive measures to reduce return rates, such as improving quality or adjusting pricing.
Overall, the project's conclusions highlight the need to address the high return rates in the shoe industry, particularly for children's shoes and sandals. The developed models and AI techniques provide valuable insights and predictive capabilities to assist Van Gastel Shoes in reducing return rates and minimizing their impact on the environment
About the project group
We work on mondays and wednesdays on the group project, which is in total 16 hours per week. We worked with sprints in every 3 weeks. Also we worked with trelloboard for the planning for our project. Everyone in our group had experience with data and AI, we all followed the semester 4 AI classes. In total we had 18 weeks to work on the project and we are in the semester of advanced AI.
