Dating platforms play the role of the traditional village matchmaker when they suggest potential partners that would be a good fit (‘match’). This paper reports from an in-depth study of the matching machinery of four dating platforms using a recommendation system based on a matchmaker model to suggest matches. While content-based recommendation systems form suggestions based on the users’ behaviour and interaction patterns, a matchmaker model uses information about the user to form recommendations. In the matchmaker model, what the IT system characterises as the ideal formation and a ‘good match’ is revealed. By using the reverse-engineering method, we find that of the four platforms investigated, three construct and form matches based on the couple’s degree of similarities along psychological and personal aspects, while one platform is based on a ‘the more similar along all kinds of axes, the better’-model. None of the platforms employs the anthropological hypergamy principle, which refers to the tendency of women to choose partners of similar or higher social status, while men do the opposite, into its matching account. Match value, which we conceptualise as the match score assigned by the platforms to couples, is a key component in the platforms’ matching machinery. Match value is a numeric value presented as an objective and scientific score, representing the degree of how well two persons ‘fit’ together. The platforms reduce individuals and relationships to a numeric value based on a psychological personality model, which ignores the person’s wider social network, class and context. The ranked order of matches does not consequently correspond with the match value, which suggests that the platforms provide benefits for paying members.

No abstract