In Odnoklassniki, a machine learning-based algorithm for analyzing group content begins to work. The algorithm evaluates the public content of groups by dozens of parameters and, depending on the assessment, ranks them in the recommendations section and in the news feed.

It’s easier to say, now Odnoklassniki has its own smart feed.

Read also: How to upload a video to Odnoklassniki

How recommendations and smart feed work in Odnoklassniki

The main task of the new algorithm is to improve the quality of content in users’ news feeds. In particular, he knows how to determine when and on which page the content first appeared in OK, and to find duplicates in other groups, and also takes into account many other factors.

Thanks to the new algorithm, authors of unique and interesting content will have more opportunities to reach the mass audience of the social network, the media will be able to interact more often with their readers and viewers, and entrepreneurs and brands will organically attract more customers.

Now higher ranked those groups that publish a high percentage of unique content not previously published by other groups and profiles in OK. In addition, the algorithm takes into account the content engagement in relation to its coverage in the news feed, the proportion of bots among the group’s subscribers, and other factors. Most of the factors are taken into account in comparison with other groups in OK, which allows you to objectively evaluate authors in the general ecosystem of content on the social network.

The first tests showed that thanks to the new algorithm, the OK feed has more diverse content and a significant decrease in the number of groups that regularly publish pictures and posts that are repeated from time to time, for example, of a humorous nature.

Groups with a low score from the algorithm receive less coverage in the news feed and may not be included in the group recommendations section. At the same time, coverage has grown among small and new groups that are interested in creating unique content for users.

Groups that, as a result, began to rank lower in the feed and are less likely to get into recommendations, will be able to improve the quality and uniqueness of their content at any time and, over time, again claim a higher coverage. The algorithm analyzes changes in the content of the groups on a daily basis.

Earlier Odnoklassniki launched interactive advertising in the mobile news feed, and also added a menu of links for groups.