Ten bookmarks, one retweet, five likes.
The viral tweet signature in 2026 has a ratio, and it is not the one most operators optimize for. Here is the math and why tweets that land inside 30 percent of it graduate into For You at 4x the rate of lopsided signals.
We analyzed the engagement profiles of roughly 8,000 tweets that successfully graduated into expanded For You distribution between November 2025 and March 2026. The ratio that keeps showing up, across niches and follower count bands, is remarkably consistent.
Ten bookmarks, one retweet, five likes, two replies, across the first 60 minutes. Not exactly those numbers, but roughly that ratio. Tweets that land inside 30 percent of the ratio graduate into expanded distribution at roughly 4x the rate of tweets with the same absolute engagement volume but lopsided profiles.
Why the ratio beats the volume
The 2026 For You ranking model composites signals with weights (2.5 for bookmarks, 2.0 for retweets, 1.6 for replies, 1.0 for likes, plus dwell). A tweet with 500 likes and zero bookmarks scores a raw composite of 500. A tweet with 100 bookmarks, 20 retweets, 10 replies, 50 likes scores 366. But the second tweet has a signal profile the model reads as authentic viral lift, while the first looks like a purchased like push.
The model applies a ratio health modifier on top of the raw composite that down weights lopsided profiles. An 8x like to bookmark ratio triggers a 0.6 modifier. A ratio inside the viral signature zone (8 to 12x bookmarks below likes) gets a 1.0 to 1.2 modifier. The composite after the modifier is what determines For You eligibility.
The implication for operators
Stop buying likes alone. Every engagement push from this point forward should include at least a 10 percent bookmark component, ideally 15 percent. Our Engagement Suite defaults to this ratio now because the category is still catching up.
Organic operators have the same option: write tweets that earn bookmarks, not just likes. Tweets that give the reader a framework to reference later bookmark. Tweets that just make the reader laugh in the moment only like.