Competitor Identification Based On Market Commonality And Resource Similarity

With the development of technology, the number of products increases rapidly and the product lifecycle becomes much shorter, which makes it more important to identify competitors broadly and rapidly. Given a focal firm, we want to build a model to calculate the competitiveness of other firms over it to find its main competitors. Furthermore, we want to classify the competitors identified into direct ones, indirect ones and potential ones.In the article, first we construct the theoretical basis of our methods through summarizingthe related work in marketing science. Then we discuss the advantages and drawbacks of proposed data mining models and introduce two typical models “Nearest Neighbor Approach” and “User Coverage Model” based on supply side and demand side respectively. To overcome the drawbacks of current models and comply with the theoretical basis, we propose a new model considering both “Market Commonality” and “Resource Similarity”. When computing “Market Commonality”, we proposealgorithms to learn users’ preference ranks of different firms. Then based on these ranks, we propose a way to aggregate them to get “Market Commonality”.In order to verify the validity of our new model, we crawl information of real merchants and their customer reviews from the internet and use them to identify competitors through “Nearest Neighbor Approach”, “User Coverage Model” and our new model respectively. Through user survey, the new model is proved to have better performance.