To address the immediate challenges and to embed a strategic solution that would support future growth, Goodiebox engaged Integrator, an IBM Business Partner, which recommended IBM Decision Optimization CPLEX® solutions.
Before deploying any software, the collective team started with a business framing session to align on the business opportunity, understand the current status, and design the desired outcome. Taking an iterative approach, Goodiebox targeted co-creation of a minimum viable product (MVP). With a working solution in place, the team then applied its experience to improve the data model by adopting new modeling approaches and considering new data sets to yield better outputs.
To facilitate tangible interaction with the data model, Integrator developed a mockup of the user interface and presented the user experience to Goodiebox, which enabled rapid MVP creation.
Juliette Giraud comments, “Integrator has been a key part of the project from the start. One of their data scientists worked with the IBM Client Engineering team during the co-creation phase of the MVP. They produced a data model that automates product recommendations, optimizes existing inventory, and optimizes parameters to maximize customer satisfaction.”
The solution utilizes several IBM Decision Optimization CPLEX components, combined with IBM Watson® Studio Heritage, IBM Watson Machine Learning, and IBM Cloud. As Juliette Giraud adds, “IBM and Integrator worked closely with us to understand our needs and came up with a solution tailored to our business. It was a collaborative effort, and we were very pleased with the outcome.”
After completing the MVP, Integrator and Goodiebox were able to use real-world data to adjust the data model and fine-tune performance. Using stock data and expiration dates, the system suggests possible product mixes for the boxes, enabling Goodiebox to edit the selections, assess the impact on inventory, and reserve items for current or future boxes.