Connor Hayes on Threads' 'Dear Algo' feature, creator strategy, and why taste is the word of 2026

Summary

Connor Hayes, who runs product for Threads at Meta, lays out how the platform is thinking about feed personalization, creator growth, and monetization — and why it is deliberately not trying to be everything to everyone.

Dear Algo

The feature started as a user behavior: people were already writing "Dear Algo" posts as a way to signal preferences to the feed. The system wasn't built to honor those posts as strong signals, so Hayes formalized it. Now, typing "Dear Algo" into a Threads post triggers a tagged, machine-readable request. Hayes used it himself to block all NFL content for three days and says it worked cleanly. The reason it works better now than it would have five years ago is LLM-powered content understanding — the platform can parse a car reference down to the specific model, brand, and country of manufacture rather than just tagging a post as "cars." Posts recommended as a result of a Dear Algo request are labeled as such in the feed. Requests for content the platform doesn't have enough of get a rejection message; requests for clearly harmful content get blocked.

Feed mechanics

Threads runs a cycles-of-auditions model: new content gets sampled to a small initial audience, and if it performs, it gets broader distribution. The window is tight — content posted more than three days ago is no longer eligible for recommendation to non-followers. Hayes is direct that the smaller the sample, the wider the error bars, so multiple cycles are needed before the system has enough signal.

Engagement through replies drives disproportionate reach. Hayes cites Draymond Green as the clearest example: Green replies aggressively to critics, which signals authenticity and drives feed distribution. The accounts that underperform tend to be the ones that feel performative rather than native to the platform.

Creator strategy

Hayes distinguishes between two types of creator growth. The first is importing talent from Instagram — using Meta's cross-promotion in Facebook and Instagram feeds, which he describes as the best marketing channel Threads could ask for. The second, which he considers equally important, is homegrown talent. He points to YoRush, an NBA fan who started posting on Threads at launch and has become what the community calls the mayor of NBA Threads, as the model. Hayes argues that if a platform only recirculates stars from other apps, it never signals to ordinary users that organic growth is possible.

On creator payouts, Hayes says he has never seen a sustainable direct-payout product work inside an app like Threads. His preferred model is directing traffic to the places where creators already monetize — podcasts, Substack, Patreon, sponsor deals. Threads has built Spotify podcast integration with rich previews and profile pinning as an early version of this. The logic: not all views are equal, and paying per view creates an incentive to chase engagement rather than build an audience with real value. He notes the irony that the creators most attached to platform payouts tend to be those earning around $80 a year, not the ones generating hundreds of millions of views.

Positioning against competitors

Hayes says the most surprising thing from Threads' first few years is how much of the user growth has come from people who never used X or a similar platform. That changes the competitive framing — it's less about pulling users away from X and more about expanding the category of people who engage with text-based social at all. The niche communities driving that growth include dating threads, book threads, and crocheting, alongside the sports and pop culture content that gets more attention.

Hayes is explicit that Threads is not trying to be a video app. Short-form video is Instagram and Facebook's job. The bet is that text format has a large addressable space that isn't fully served, and that Threads can own it without competing directly with Reels.

The team remains small relative to other Meta apps — Hayes doesn't give a headcount but describes it as smaller by orders of magnitude. Growth investment this year is going into two areas: feed relevance and personalization, with Dear Algo as a visible piece of that, and organic discovery so the platform is less dependent on Facebook and Instagram promotion to drive sign-ups.