The blog post explores the development of an objective recommendation system using Memgraph and the Cypher query language, focusing on characteristics rather than numerical ratings to suggest products. It critiques traditional collaborative filtering methods that rely heavily on user ratings, which can be subjective and sometimes obscure lower-rated yet valuable items. Instead, the proposed system utilizes graph databases, where nodes represent games and characteristics like genre, developer, and publisher, with edges indicating relationships and assigned relevance to highlight importance. This approach allows for recommendations based on shared characteristics, demonstrated through queries that identify similarities between games like "Stardew Valley" and "Forager," and explore corporate connections in the gaming industry, such as between CD Projekt Red and Electronic Arts. The article suggests that this method offers easy expansion possibilities with more detailed data, different relevance systems, and alternative graph algorithms to enhance recommendation accuracy and system credibility.