The paywall works
Founders, operators, engineers, VCs. The people who decide whether your product gets bought are concentrated on X, and a large share of them pay for Premium. One of the things they are buying is the absence of promoted posts.
Marketing teams tend to treat this as a channel-efficiency question, CPMs are up, targeting is worse, we need better creative. It is not that.
The buyer has purchased a product whose explicit function is to make your advertising invisible to them. No amount of creative fixes a channel the audience has paid to switch off.
What the data says
Buffer analysed 18.8 million posts across 71,000 accounts. The median non-Premium post earned zero engagement, while Premium accounts pulled roughly 10× the reach.
Read that carefully, because the implication is not "buy Premium". The implication is that reach on the platform has consolidated into a smaller set of accounts, and those accounts are the distribution layer. Everything else is shouting into a feed the algorithm has already decided not to serve.
The only remaining route
If paid is switched off and trade press has no reliable readership among your buyers, one channel remains: the accounts those people already chose to follow.
That is a relationship problem rather than a media-buying problem, and it behaves differently. You cannot buy it at auction, you cannot scale it linearly with budget, and it does not work if the content reads like advertising.
What it does do is arrive pre-trusted, because the person carrying it has spent years earning that audience's attention.
Mechanically: quote posts, original takes, and use-case content, posted inside the same window. The algorithm reads engagement velocity, reply chains, and sourcing signals, and pushes wider. The full mechanics are here.
The strongest evidence for this argument comes from people with no reason to make it. The best-capitalised firms in venture have spent the last two years staffing media operations at scale, on the view that reach rather than capital is the part that is now hard to get. The future of distribution reads that shift, including where it leaves companies that cannot fund a newsroom.
How the math changes
The first eight hours seed the climb, then it compounds. An average launch through our network lands around 1M views; strong launches reach 2–5M over 24 to 48 hours, and the shape is a curve over 24–30 hours.
We guarantee nothing, because the honest version is arithmetic: creators engaged multiplied by their median reach sets the floor, and the algorithm sets the ceiling. Anyone promising a number ahead of time is either guessing or buying fake engagement.
What compounds is the base. Higgsfield ran six releases through the network in four months, and each drop landed on an audience the previous one had warmed. The baseline rose across the campaign instead of resetting to zero each time.
That is what accumulation looks like in a chart, and it is why cadence matters: a buyer needs around fifteen encounters and any single launch supplies two or three.
Who this does not apply to
Worth being precise, because the thesis is not universal.
- If your buyer is not on X, some verticals genuinely are not, this argument does not hold and you should ignore it.
- If you sell to procurement rather than to practitioners, you need the credibility artifacts that PR still produces. Distribution complements that, it does not replace it.
- If the product is not yet good, distribution accelerates the discovery that it is not good. That is a real risk and we say so before taking work.
The case against this thesis
The strongest objections we hear, and what would change our minds.
- Views are not buyers. A launch can reach millions of people who will never pay. That is why we report bookmarks next to views: across 1,872 launches, posts saved more often for each like travelled further at every account size.
- X can change the ranker. The thesis rests on followed accounts reaching their audiences. If X stopped serving them, the directory would show it first, as launch multiples falling toward 1x.
- Disclosure could cost trust. Every paid post carries X's paid partnership label. If labelled posts started travelling less than unlabelled ones, we would change how we run launches.
- It does not scale like ads. Reach through people grows with relationships, and those take years to build, which is also why it is hard to copy.
Questions we get asked
Isn't a viral spike just vanity if it doesn't convert?
Reach without conversion is vanity. Every launch we run is engineered backwards from the inbox, we track which creators drove qualified inbound, route comments to your team, and brief follow-up DMs.
Why not traditional channels?
X is where the AI and tech conversation happens. VCs, founders, and operators live on Premium and check it daily.
LinkedIn rewards corporate posts. TikTok reaches a younger consumer audience.
How long does the bump last?
Impressions peak within 48 hours, then taper. The mindshare doesn't. Your buyers saw your product on five accounts they trust in the same day, and that fills the pipeline for weeks after.