VeryBuy uses "style tags" to create data, turning new customers into regular customers In addition to helping "refresh" physical activities, machine learning can also "reshape" e-commerce players who are already in the digital world. E-commerce platforms usually understand customers' "past consumption behavior" through digital footprints, and then get a glimpse of "future consumption choices"; however, what if they are new visitors who have just arrived and haven't left any information? VeryBuy, a machine learning application with AWS, actively creates data through "style tags" to turn new customers into regular customers.
In the past, e-commerce websites often divided visitor behavior into four stages: “browse, compare, check, and purchase” (browse, compare prices, view products, and add to shopping cart). Therefore, “purchase” is often the last mile of all e-commerce efforts. , everyone resorted to all kinds of techniques, just hope that customers can quickly enter the final stage. However, the whatsapp list cross-border e-commerce platform VeryBuy does the opposite, using AWS machine learning to focus on data application in the first three stages of "browse, compare, and check". They structured the data of women's clothing on the station through "style tags", and divided them into different style elements such as preppy style, OL style, etc.
For example, a dress may be marked with 30% preppy style and 70% OL style. When a new customer visits the website for the first time, even if they have not registered or entered the "purchase" stage, the platform can recommend other products of similar styles to the buyer based on the weight of the style elements of the products that the customer has just browsed, compared prices, and viewed. , so that new customers are as familiar as old customers; and when encountering e-commerce festival events such as Lady Night, they can also pick out suitable festival products based on style labels.