Bluesky Is Shrinking and Still Beats Threads on Referral Traffic

Threads passed 400 million monthly active users in late 2025 and its daily actives grew 21.3% year over year through July 2026. Bluesky went the other direction, falling from a 22.1 million monthly-active peak to 10.4 million. Yet publishers still report three times more referral traffic from Bluesky than from Threads. The practical move is to split effort by goal: reach on Threads, clicks on Bluesky.

That opening stat pair confuses a lot of people, and honestly it confused me for a while too. The biggest alternative platform is the worst at sending traffic, and the shrinking one is the best at it. Once you understand why, the allocation decision gets much easier than the "which platform wins" framing suggests.

The scoreboard, and what it hides

Start with the raw numbers, because they set the ceiling on everything else. Meta says Threads is now around 450 to 500 million monthly active users, with daily actives at 147 million as of July 2026, up 21.3% year over year, according to TechCrunch's August 2026 reporting. Bluesky peaked around 22.1 million monthly actives in late 2024 and sat at 10.4 million in June 2026, a 52% drop from that peak, with roughly 3 million daily actives. Mastodon is smaller still, hovering somewhere between 700,000 and a million monthly actives depending on whose count you trust.

Quick metric note, because these get conflated constantly: monthly active users measure how many people opened the app at all in 30 days. Daily actives measure habit. Bluesky's 46 million registered accounts number, which the company itself promotes, measures neither. It counts everyone who ever signed up, including the people who left in 2025. When a platform leads with registered accounts instead of actives, that choice is telling you something.

So Threads has roughly 40 to 50 times Bluesky's active audience. If audience size were the whole game, this article would be one paragraph long. It isn't, because the platforms treat your links completely differently.

Bluesky keeps shrinking and keeps out-converting

Here is the part that should change how you allocate hours. EMarketer's reporting on publisher referral data found The Boston Globe getting three times more traffic from Bluesky than from Threads, with 4.5 times higher conversion to paid subscriptions. EUobserver pulled 3,800 unique visitors in a week from a Bluesky account with 3,300 followers, against 1,320 visitors from an X account with 203,000 followers. Gaming site Aftermath watched its Twitter-to-Bluesky referral ratio go from 9-to-1 to nearly 2-to-1 in three months.

Fair warning on recency: that publisher data is from late 2024, when Bluesky was at its growth peak. The gap has probably narrowed some as actives declined. But the structural reason behind it hasn't changed. Bluesky COO Rose Wang has said it plainly: "Unlike other platforms, we don't depromote your links." X confirms it demotes posts with external links, and Threads inherits Meta's general allergy to sending users anywhere else.

A follower on a platform that suppresses links is worth a fraction of a follower on a platform that doesn't.

The other thing the shrinkage headlines miss: the people still on Bluesky are unusually committed. TechCrunch notes a stickiness rate around 29% in June 2026, which is comparable to Threads despite the platform being one-fortieth the size. Think of Threads as a stadium crowd and Bluesky as a trade-show happy hour. The stadium has more people, but at the happy hour, the people you meet actually take your card.

The action here is measurable. If your content targets journalists, academics, developers, policy people, or the media-adjacent crowd that migrated there, post your links to Bluesky for 30 days with UTM tags. The benchmark from what I've seen in publisher accounts: clicks per follower should come in at 2 to 5 times your X rate. If it doesn't beat X by at least 2x in a month, your audience isn't there, and you can drop the platform with a clear conscience.

Threads is the reach play, and the ads finally exist

Threads in 2026 looks like a platform mid-transition from growth mode to harvest mode. Sprout Social's statistics roundup puts median brand engagement between 3.8% and 4.5%, roughly double what X delivers, and reports that 30% of users interact with brand content at least once a week. Around 57% of brands now post there, though the split is lopsided: 64% of B2C brands versus 19% of B2B. Ads went global in early 2026 through Meta Ads Manager, and eMarketer projects Threads will add $11.3 billion to Meta's revenue by the end of 2026.

That last number matters more than it looks. Platforms with small ad loads give organic content room to breathe. Platforms chasing an $11 billion target do not, at least not for long. We watched this exact movie with Facebook pages a decade ago, and the sequel rarely surprises.

I'll put a number on it: by mid-2027, organic brand reach on Threads drops by at least a third from today's levels as ad load matures. Maybe I'm wrong on timing, Meta has occasionally been patient, but the direction seems close to inevitable. Which means the current window, where a text post from a mid-sized brand can still out-reach its follower count, is the cheapest reach Threads will ever offer you.

What to actually run: five text-first posts a week, native to the platform rather than repurposed link dumps, with replies handled within the hour (reply engagement is where Threads' algorithm pays out, and Sprout's data shows reply-heavy brands growing several times faster than broadcast-only accounts). Benchmark against the 3.8% median. If you're under 2% engagement after six weeks of consistent posting, your format is wrong, not the platform. The dynamics rhyme with what we found digging into LinkedIn's 2026 algorithm: conversation is the ranking signal, and links are the tax.

Mastodon is a community you join, not a channel you run

Mastodon deserves an honest paragraph instead of a polite one. With well under a million monthly actives, it is smaller than many single subreddits. Comparison pieces like Postory's 2026 breakdown keep landing on the same profile: strong norms (alt text expected, content warnings used, hashtags for discovery) and a deep, well-documented hostility to anything that smells like growth hacking. Brands that show up broadcasting get ignored or actively dunked on.

And yet I wouldn't tell everyone to skip it. If your product serves developers, open-source communities, security people, or privacy-focused audiences, those communities are disproportionately there and disproportionately loyal. The way in is a person, not a logo. Have an engineer or founder participate under their own name, answer questions, share things with no CTA. It's a listening post and a credibility play, not a distribution channel, and budgeting it as distribution is how brands end up writing "Mastodon didn't work" postmortems.

Anyway, for most marketing teams the honest Mastodon line item is zero hours, and that's fine.

The split I'd run with five hours a week

Most teams reading this don't get a headcount for alternative platforms. They get a few hours skimmed off the top of someone's week. Here's how I'd spend five:

Three hours on Threads. Five native text posts, reply windows after each, no link in the first post of a thread. Measure engagement rate against that 3.8% median and watch reach per post monthly, because when the organic window starts closing you want to see it in your own data, not in a trade publication six months later.

An hour and a half on Bluesky, but only if your audience skews media, tech, policy, or academic. Links are welcome there, so this is where your actual traffic play lives. UTM everything. Judge it on clicks and conversions per hour spent, not follower growth, because follower counts on a 10 million MAU platform will never impress anyone in a reporting deck. The conversion quality can.

Thirty minutes on Mastodon, as a person, if and only if the audience fit from the section above describes you. Otherwise fold that time into Threads replies.

One habit that applies across all three: platforms this young change fast, and the strategy that works is closer to series-based content thinking than to campaign thinking. Small repeatable formats, weekly review, quick kills.

I don't think there's a single winner here, and I'm suspicious of anyone confident there is. Threads is buying reach with Meta's distribution muscle and will eventually charge rent for it. Bluesky is losing casual users and keeping the ones who click. The teams that do well over the next year will probably be the ones that stopped asking which platform wins and started asking which metric each platform is actually good at.

Questions marketers keep asking

Is Bluesky still worth it for brands in 2026?
For most consumer brands, no. For anyone whose audience includes journalists, developers, academics, or news-heavy readers, yes, because Bluesky doesn't suppress links and its per-follower referral traffic beats X and Threads by multiples in published publisher data. Judge it on clicks, not follower counts.

Does Threads have advertising now?
Yes. Threads ads rolled out globally in early 2026 through Meta Ads Manager, and eMarketer projects the platform adds $11.3 billion to Meta's 2026 revenue. Early CPMs have generally run below comparable Instagram placements, which favors testing budgets now rather than after pricing matures.

Can I just cross-post the same content to all three?
You can, and about half of marketers repurpose across platforms. It works acceptably on Threads, poorly on Bluesky (broadcast-only accounts stall there), and badly on Mastodon, where community norms punish it. If you only have time to nativize for one platform, make it the one carrying your traffic goal.

Notice Me Senpai Editorial