Topic affinity

Meta’s patents describe possible evidence for topic interest in social-networking systems, rather than documenting how Instagram currently calculates it.23 Meta’s event-invitation patent illustrates this with automobile racing: expressed interests, liked pages, related posts and participation in racing activities can help identify people interested in a racing event.2 Meta’s topic-feed patent treats more frequent posting about a subject as evidence of greater interest and describes posting about one subject as a possible indication of interest in similar subjects.3

Instagram’s description of Threads connects topic interest to discovery: its For You feed uses engagement with authors and content topics, including past interactions with accounts and posts on both Threads and Instagram.4 Instagram also says that likes on Threads help it provide similar content.4 Meta’s topic-feed patent describes a different application of topic interest: ordering topic options by the likelihood that someone will interact with their content, then allowing that person to switch between topic feeds.3 Meta’s recommendation patent describes comparing topics associated with pages or other items someone is already connected to against possible recommendations, while also considering how many other people connect to both items.5

Meta’s messaging patent describes another way to establish topic preferences: asking people to select subjects themselves rather than relying only on interests inferred from activity.6 In the Meta patent, a publisher’s welcome message can offer checkboxes or buttons that add selected topics to a subscription, allowing future messages to contain matching content.6 The same Meta patent connects these explicit choices with later activity by describing further topic suggestions based on articles the subscriber opens.6