Abstract: A method, implemented on a processor, for reducing user biases and exposure to clutter when ranking user content is disclosed. In this method, users are selected from at least one online grouping of users, such as an interest group or social group. A small subset of users is calculated, which may be from one or more groups. Content is anonymously submitted to the subset of users and they can review and anonymously respond to the content. The weight of the content is increased by positive response and decreased by negative responses. This process is repeated for a predetermined period of time, number of iterations or until a predetermined end point. A numerical value of the content rating is assigned based on the user responses and compared to a calculated numerical threshold. Only content that exceeds the numerical threshold is made available to all users of the online grouping of users.