In order to improve their skin product recommendations, Skincity teamed up with Modulai to build a system that combines the unique knowledge of the skin therapists with product- and customer behavior data to increase product recommendation accuracy with the aim of driving higher conversion rates and customer satisfaction.
Background
Skincity is an online skincare clinic that offers a finely-tuned selection of professional skincare products and make-up. The key to their success is their deep understanding of their products and customers. Central to their operation are the customized product recommendations customers receive from their experienced skincare therapists. Many customers taking the skin test are first-time customers and the Skincity’s therapists have to rely on the information filled in, such as age, skin type, and condition, as well as general preferences for organic or vegan products as a basis for their decisions.
Solution
To assist the therapists in this process, and to leverage their unique skill set even further, we developed a machine-learning-based system generating relevant and accurate recommendations of brands to the skin therapists for each skin test submission. The automated recommendations together with the expertise of the therapists result in a customized set of products and treatment plans for the customer.
We analyzed years of test submission data and purchase patterns and interpreted the results into a coherent picture In collaboration with Skincity’s skin therapists and tech team. The team developed a hybrid collaborative filtering model based on user and product attributes, as well as past interactions and test submissions.
Tools/Tech
Models were trained using algorithms for tabular data and collaborative filtering. The backend was primarily developed in Python and deployed as a set of dockerized services. Various AWS services were used for data pipelining and fast and scalable deployment.