Different product characteristics and consumer expectations must be analyzed when making a product or service recommendation based on use. However, if all types of knowledge are inaccessible, this is known as the clod start problem in a recommendation system. To explain and address the issue of cold start, this paper suggests a structure for user experience-based product recommendation. As a result, this paper introduces a four-phase recommendation scheme. In the first step, site usage data is preprocessed and used to make product recommendations. The frequent pattern-based recommendation is performed in the next step. The decision is made in the third process based on the user’s current search (user click streams) and related user habits seen in network access logs. In addition, the final recommendation is based on filtering the results obtained in the third step, as well as the product or service’s cost, brand, and social review. The results show that the prediction is accurate and precise. Furthermore, compared result of proposed system to traditional system using Apriori Algorithm.