By manipulating a new ranking model, Pinterest had published an overview to update algorithms in the home feed. This has enabled it to display a broader range of content rather than relying on engaging data to drive Pin recommendations. Traditionally, Pinterest ranks content using a click-through prediction model that maximizes user engagement. But this is not an apt model for surfacing content types.
It is a unique approach introducing a real-time ranking system called controllable distribution. This facilitates to focus on video content or other content from a more culturally diverse set of creators without just compromising user experience.
This explains the feature of Pinterest that highlights other contents rather than past activity and also to pin video content to surface actionable ideas from the creator. Video content generates more user engagement and is an extra boost within the algorithm beyond the past user activity. Pinterest will still focus on click-through the prediction model to find relevant content and then apply controllable distribution to diversify the types of content.
Controllable distribution is a flexible real-time system to specify a target at what percent a certain content type should receive its readability.ie, it tracks the percentage of feed in the past and then demotes content accordingly to make it close to the specified target. Hence Pinterest seeks to amplify certain content types to meet variable goals in addition to basic ranking on each user’s engagement.
Pinterest is investing in more resources to diversify its platform ensuring that users are exposed to more content from different sources. If you want to maximize your reach on Pinterest, it’s worth to conserve your content flowing with new, elegant Pins for the system to rank, and also to assess different content types, particularly video, new post options and tools that Pinterest may be looking to give more exposure via algorithm ranking