Case studies

Given the strong growth Flip Design has experienced over the past 15 years, certain challenges have emerged in inventory management. The company has two major sales periods annually but aims to go into production only once annually. As a result, accurately forecasting sales becomes crucial to estimate the required quantity of each clothing item that needs to be manufactured to meet demand and to limit the need for ad hoc production runs.

The challenge with these forecasts, beyond their required precision, lies in the wide variety of garments the company offers. Indeed,In addition to an impressive range of uniform pieces, each item is available in 12 different sizes and several colors. Manufacturing forecasts must, therefore, be carried out individually for each piece. Doing this by hand is time-consuming, inefficient, and laborious.

Flip Design turned Vooban to integrate artificial intelligence into its sales forecasting and production planning processes. The goal was to obtain precise information to optimize their operations and inventory management.

Vooban developed a web interface that generates sales forecasts for all 700 Flip Design products, each of which is further categorized by color, size, and the specific school logo where the items will be delivered. We also developed a mobile application that allows users to visualize these forecasts through an easy-to-analyze graph. Users can submit queries by selecting their desired constraints (item, color, size, school), and the tool produces an estimate while factoring in a wide range of variables, from past sales data to Canadian consumer habits, and even the unemployment rate. Plus, the AI model that processes all this information re-trains itself automatically every month! But Vooban didn’t stop there. A second system designed specifically for Flip Design was also created: an inventory optimization module.

In forecasting sales, there is always an inherent level of uncertainty, a likely margin of error. Flip Design must decide whether to produce slightly more or slightly fewer garments than it anticipates selling. On one hand, producing more means having surplus inventory and the risk of losses; on the other hand, producing too little could mean not having the right products available to quickly fulfill orders. Using the tool's forecasts and the margin of error, our inventory optimization module suggests the ideal quantity to produce, balancing storage costs and opportunity costs to support better decision-making.

Implementing the sales prediction tool has led to considerable savings for Flip Design. It has helped to reduce the number of occasional production runs and increased the company's ability to supply items more efficiently. As a result, Flip Design has maintained customer satisfaction and built a strong reputation in the school sector.

The algorithms for sales prediction and inventory management operates on a 12-month time frame. This allows for larger production runs, reducing the need for occasional batches. These are costly in terms of resources, time, and money, but often necessary when stock runs out.

This tool also equips Flip Design with the ability to support rapid growth without taking on excessive risk. As a result, the company is now well-positioned to expand into new markets across Canada and into the Northeastern United States.