A major online thrift store with multiple locations across Europe and the United States. Launched in 2013, it has grown into a large ecosystem with centralized warehousing and sorting facilities that re-launch second-hand apparel items into online markets. In 2019, its overall turnover has reached $15 million.
The online thrift marketplace works by collecting the apparel items from the sellers, delivering them to the centralized warehousing facility and appraising it to determine its final cost.
One of the key goals of that process is to identify the brand and model of each accepted item. Basic appraisal also includes examining the quality of the item received, which also affects the resell price. Originally, these stages were completed manually, but this approach has led to errors and higher personnel costs. Erratic identification means wrong price being set, and that led to operational and financial setbacks.
Customer needed to:
To tackle these issues, we have developed a system for automatic SKU matching based on photo analytics and machine learning.
Reksoft specialists have created a model that allows employees to accurately appraise and identify apparel items to a database of existing SKUs. The system is built on photo analytics and machine learning technologies based on advanced deep neural networks.
Suggested SKU matches of a processed item are determined by a convolutional neural network. It analyzes the photo and, based on authentic features, gives the probabilities of the depicted apparel to can be, for example, kids jersey. The system selects the most likely option.
The data of one apparel element is processed within a few seconds, whereas previously the process could take a long time, and the estimation accuracy was at a significantly lower level.
The network is trained by manual corrections: if the prediction is wrong, then the parameters are rearranged so as to correct the estimate. This happens until the error percentage is reduced to the required value. In the future, when new data and new errors appear, the network will retrain on its own.
The developed system makes it possible to reduce the influence of the human factor, shorten the assessment time and significantly optimize the logistics process. In addition, the solution can help form a common standard for the application of machine learning methods and can be used for other purposes, apparel options and types of logistics. The system is successfully implemented on warehousing facilities in Europe and the USA.