With the improvement of people's living standards, the requirements for Number List commodities tend to be personalized. Everyone has their own habits for their preferences for products, and grasping the habits and developing them is also an effective means of promotion and publicity on the e-commerce platform. In this way, the recommendation system came into being.
Recommending, as the name suggests, recommends products that can meet users' demands and needs according to users' preferences and behaviors, so as to achieve the purpose of users' purchase. In a broad sense, all product information that is actively pushed to users can be regarded as the category of recommendation, and the product recommendation with commercial monetization capability is also called advertisement. The management of advertisements is generally handled by a separate advertisement system, and the recommendations described below mainly refer to the recommended content of products other than advertisements.
1. Recommended basic information
The recommendation system is fundamentally to solve the problem of decision-making in marketing product selection, and all need to have some indicators to measure and evaluate the effect to provide the basis for the subsequent adjustment of the parameters of the recommendation strategy and optimization of the method. Common recommendation system metrics and search comparisons include precision, recall, and novelty.
Precision: Indicates the percentage of the recalled products that recommend the correct products to the total recalled products.
Recall rate (Recall): Indicates the percentage of recalled products in the overall product.
Novelty: Indicates the situation of the products in the recommended long-tail range. If the recommended products are all popular products, the novelty is very low; otherwise, the novelty is high.
CTR: Click-through rate, which is also one of the measurement indicators of the advertising system.
CVR: Conversion rate, which refers to the conversion rate from when a user clicks a recommended product to complete a purchase. The formula is CVR=(Conversion/Click)*100%
Let's take a look at the common presentation forms of recommendation systems on the user side of e-commerce platforms, as shown below: