Optimizing procurement processes for Vivo Fashion Group
Team members
Tracy Ly
Grace Austin
Nicole Yang
Ryan Xiong
Betty Wu
Faculty mentor
Itai Ashlagi
Sponsor organization
Vivo Fashion Group Ltd. operates in Kenya, Rwanda, and Uganda, where they design and sell clothing tailored for the body shapes and style preferences of African women.
Project description
Vivo is looking to streamline their internal procurement process for new clothing styles and collections. Vivo currently finds that they run out of styles very quickly and leave a lot of money on the table. Thus, they are seeking a data-driven way to predict future sales demand, and a system to collect actionable feedback from sales teams and customers to inform the manufacturing and distribution of new clothing orders. Such tools would allow Vivo to maximize the percentage of clothing stock they sell at full price and to capitalize on up-and-coming style trends.
Techniques and methods used
Our techniques and methods focused on devising solutions to predict future sales demand using historical data, and finding ways to inform future style trends for which historical data is not available. We believed that addressing these two questions would allow Vivo to paint a better picture of what the demand is for their products, allowing them to better meet that demand.
Initial work involved exploratory data analysis to identify useful trends across different styles, colors, product lines, and price points, as well as potential limitations in their data that would be useful when developing a prediction model. Our team also conducted several interviews with Vivo's data and sales teams to better understand their needs when it came to tools that we were developing. Insights gleaned from these interviews informed the features of the survey and model tools.
Both of the above approaches proved vital in the development of our final deliverables. They also allowed us to identify critical limitations with the data that Vivo currently tracks, allowing us to refine the scope of our model and to identify several recommendations that would enhance the impact of our final deliverables.
Solutions and deliverables
Our team produced two final deliverables. The first involved the creation of a forecasting model developed in Excel to allow Vivo to leverage their sales data to generate future sales predictions. This involved the use of Excel's built-in ETS algorithm to generate future sales predictions at various future points in time based on historical sales. For this deliverable, we wanted to prioritize user-friendly features and intuitive design that would allow the model to be generalizable across different data sets and accessible to all levels of users.
Our second deliverable involved the creation of several multi-layered Typeform surveys to be distributed among sales teams and customers to better identify which styles are popular, which style demands are not being met, what customers are saying they would like to see in the future, and to get feedback on new collections before production. Such an approach is intended to help quantify aspects of demand that data cannot capture, and to minimize risk associated with over- or under-production of certain products.
These deliverables helped inform further recommendations for Vivo regarding data collection and infrastructure that would enhance the power of these tools into the future. We believe that these tools will allow Vivo to gain insight into demand trends, which in turn will allow them to better optimize their design processes, manufacturing, and inventory management. Ideally, Vivo will be able to remain at the forefront of their industry in a world in which the ability to leverage data is becoming increasingly critical.