On Tuesday, Twitter announced it will use new “behavioural signals” to push down more tweets that “distort and detract” from conversations and searches.
Currently, Twitter (TWTR) uses policies, human reviewers and machine learning to decide how tweets are organized and presented in conversations and search.
Its new approach will now also take into account thousands of behavioral factors — such as whether an account has confirmed its email address or how often an account mentions users that don’t follow them back — into how tweets are ranked. For example, tweets that are off topic, spammy or troll-like will appear lower in conversations and search results.
Other new behavioral signals that Twitter will use include whether the same person signs up for several accounts simultaneously or behavior that suggests a coordinated attack, such as multiple accounts disrupting a conversation with the same hashtag.
A Twitter spokesperson told CNN these types of tweets will not be hidden, muted or removed from Twitter because they don’t violate its policies. The tweets will still be available if users select “Show more replies” or choose to see everything in search.
“By using new tools to address this conduct from a behavioral perspective, we’re able to improve the health of the conversation, and everyone’s experience on Twitter, without waiting for people who use Twitter to report potential issues to us,” the company said in a blog post.
In its “early testing,” the company said the new approach resulted in a 4% decline in abuse reports from search results and an 8% drop in abuse reports from conversations.
Users will not be notified if their tweet was pushed down.
Twitter acknowledged it expects to make mistakes, and that the system will change, learn and improve over time. It does not expect the majority of users to be impacted by the change.