User-controlled ranking pushes Threads strategy toward relevance

Dear Algo lets Threads users post a public request to see more or less of a topic for three days. Meta's later update introduced Your Algo, a private control that can last one, three or seven days and initially rolled out in the US, Canada, the UK, Australia and New Zealand. These are product controls, not a public ranking formula.

The direction is still meaningful. Threads is making relevance more explicit. Meta also said Q4 optimisations drove a 20% lift in time spent on Threads, while separate product updates added topics, reply controls and discovery reporting. None of these statements prove a brand can generate better performance simply by posting more. They show that the platform is tuning conversation quality and user preference.

Reach tactics become easier to reject

If users can ask to see less of a topic, content designed only to ride outrage or broad attention carries more risk. It may still earn short bursts of visibility, but it also trains people to classify the account as irrelevant. That is costly for organisations that need to be trusted over time.

A more defensible strategy is to define relevance at the paragraph level. Before publishing, the team should be able to name the audience, the live question, the evidence and the next decision. If that cannot be named, the post is probably relying on mood rather than usefulness.

Time spent is not advertiser performance

The 20% time-spent lift is a Meta platform statement. It should not be converted into expected reach, clicks or conversion rates. More time on platform may increase opportunity, but advertisers still need placement-level reporting and content teams still need post-level evidence.

For organic teams, the closest useful measure is retention of topic relevance: repeated replies from the same kind of account, fewer off-topic arguments, more saved reference posts and higher-quality follow-on conversations. For paid teams, Threads should be isolated as a placement when possible so early signals are not hidden inside broader Meta campaign averages.

So what

User-controlled ranking makes relevance a management discipline. The next action is to add a pre-publish relevance check: audience, topic, evidence, response owner and expected follow-up. Review the check against six weeks of posts and remove themes that attract attention without producing useful conversations.