Platforms increasingly shape content production not only by aggregating external information streams but also by professionally curating and generating their own content. In doing so, these platforms filter complex information and provide external producers with a structured starting point for their ultimate outputs. However, it remains unclear how reliance on platform content affects the nature of the products offered by external producers. While conventional wisdom might suggest that greater reliance on platform content will make products more similar, we find a more nuanced pattern: we hypothesize that varying levels of reliance on platform content exhibit nonlinear relationships with the similarity of end products. We test these hypotheses in the Danish news industry – where news outlets connect to the news agency and platform, Ritzau, for curated content – and analyze over 5 million articles (2000-2020) using natural language processing. Our results show that moderate levels of platform reliance decrease product similarity, while low and high levels of reliance have positive effects. Thus, relying on platform content is not only a source of standardization but may also encourage a more diverse set of products.
