Qualified impressions: what X actually pays for

The number under an X post and the number X pays on are not the same number, and the gap between them is not small. Payout under the Original Content Rewards Program runs on qualified impressions: unique impressions from subscribers to X Premium Basic, Premium, Premium+, or Premium Business, on the Home Timeline, where at least 50% of the post was visible. Paid, promoted, fraudulent, and artificially generated impressions are excluded.

That's four filters stacked on top of each other. Your public impression count applies none of them. So the useful mental model isn't "impressions, minus some shrinkage." It's a funnel where each stage removes a different slice, and the slices are independent of one another. Understanding the shape of that funnel is what stops you from planning a quarter around a number that was never the payable one.

Four filters, and what each one removes

Read the definition clause by clause. Each word is doing work.

Filter	What it removes	Can you influence it?
Unique	Repeat views from the same account	Barely — this is audience behaviour
Premium Basic / Premium / Premium+ / Premium Business	Everyone viewing without one of those subscriptions, including logged-out visitors	No
Home Timeline	Views from search, notifications, profile visits, quote-post surfaces, embeds	Indirectly, through where discovery happens
≥50% of the post visible	Fast scroll-pasts where the post never fully entered the viewport	Somewhat, through post construction

The unique filter is the one people forget, and it hurts most on the content that feels like it's working. A post that stays in circulation for two days accumulates repeat views from the same accounts checking the replies. Every one of those is an impression in your analytics. Only the first is unique.

The Premium-tier filter is the largest and the one you have zero leverage on. It reframes the whole thing: you are not being paid on your audience, you are being paid on the subscribing subset of your audience — including Premium Basic, not only Premium and above. Two accounts with identical reach and different audience composition earn differently, and no amount of format work changes that.

Do not mix this payout definition with the program's entry impression gate. Eligibility asks for 500,000 Home Timeline impressions from verified users in the last 90 days, with reply impressions excluded. That is a different count from qualified impressions. Clearing one does not mean you have measured the other.

The Home Timeline filter is the most counter-intuitive, because it penalises exactly the traffic most creators are proudest of. A post that gets linked from a newsletter, embedded in an article, or found through search is reaching people, and those views sit outside the surface that pays. Off-platform amplification is good for reach, good for follows, good for the top of every other funnel you run, and worth roughly nothing to this specific line item.

The ≥50% visible filter is the only one with a production lever attached, and even there the lever is modest. A post has to actually occupy the screen. Something a viewer flicks past at speed, or that is tall enough that half of it never enters the viewport at once, is doing worse against this filter than something compact that lands whole. That's a reasonable inference from the rule's wording rather than published guidance about specific formats, so treat it as a hypothesis to test on your own account rather than a spec.

Why you can't compute your own ratio

Here is the honest limitation, and it's worth stating plainly because a lot of advice pretends otherwise: X's public analytics do not break impressions down by any of these four filters. You cannot see what share of your viewers subscribe to Premium Basic, Premium, Premium+, or Premium Business. You cannot separate Home Timeline impressions from search or profile impressions. You cannot see viewport depth.

So there is no formula that converts your impression count into an expected payout. Anyone quoting you a conversion percentage either measured it on one account with one audience — which tells you nothing about yours — or made it up.

What you can do is stop letting one metric answer two questions. Impressions answer "did this travel?" Payout answers "did this travel among the people who pay?" Report them on separate lines, and when they diverge, believe both.

Restructuring measurement so it survives this

The practical fix is a reporting change, not a content change. Three moves:

Split reach reporting from earnings reporting. In whatever spreadsheet or dashboard you keep, impressions and payout live in different columns and never get combined into a "value per impression" figure. That figure is unstable by construction, because it moves with audience composition and discovery surface, neither of which you're controlling week to week.
Track payout per post, not per impression. Post-level earnings against post-level format is a comparison that holds up, because the audience is roughly constant across a week's uploads. Per-impression normalising re-introduces the exact noise you're trying to remove.
Log where discovery came from, qualitatively. You won't get the split from analytics, but you know when a post got picked up by a newsletter or trended in search. Flag those posts. When they show strong impressions and weak earnings, that's the Home Timeline filter, not a content failure — and mislabelling it as a content failure is how creators kill formats that were working.

Checking cross-platform performance from the agent is useful for the first job and useless for the second, which is exactly the point: they're different jobs. The wider question of what to actually track in creative analytics applies here with one X-specific amendment: the platform's own headline metric is a reach metric, not a revenue metric, and treating it as both is the single most common error on monetized accounts.

What this means for format decisions

If earnings are the goal, the funnel argues for a fairly specific posture, and it isn't the posture that maximises reach.

Favour native discovery over exported reach. Content designed to be found in the feed — conversational, timely, self-contained — sits inside the paying surface. Content designed to be linked to from elsewhere doesn't. Both are legitimate strategies; only one of them feeds this line item. The platform-native versus repurposed trade-off is usually framed as a quality question, and on X under this program it's also a payout question.

Favour posts that land whole. A single strong post that occupies the screen beats a sprawling thread against the visibility filter, at least in theory. Test it. The theory is inference, and your account's data is the only evidence that counts.

Stop chasing repeat views. Bumping your own post, replying to keep it alive, and re-quoting yourself all generate impressions from accounts that already saw it. Those impressions are real and they are not unique. The tactic still has a purpose — keeping a conversation going has value beyond payout — but it should never be mistaken for an earnings play.

Don't let this dictate everything. Off-timeline reach builds the audience that eventually becomes on-timeline reach. Optimising a young account purely for qualified impressions starves the growth that makes qualified impressions worth counting. This filter is a lens for reading results, not a mandate for every decision.

Production notes for an X-native cadence

The formats that do well natively on X tend to be short, self-contained, and specific, which is a friendlier brief for generated video than the long-form requirements other platforms attach to their payout programs. A clean generated clip with a strong first frame is a legitimate native post here; it doesn't need to be padded to a length floor.

For the assembly side: keep a reusable timeline rather than posting one-off raw generations, so you can re-cut the same material for a second attempt without regenerating it. Vary the look across a run — a feed that reads as one template is a problem under the program's originality requirement, separately from anything in this post. And if you're building the X format from scratch, the X and Twitter workflows page and the video-on-X brand strategy cover the creative side.

One measurement habit worth building early: keep the finished posts organised. Saving generations to a project means that when a format outperforms, you can find and rebuild the exact assets rather than reconstructing what you did from memory three weeks later.

FAQ

Does a view from a non-Premium follower earn nothing?

Under the qualified impressions definition, a view from someone who is not subscribed to Premium Basic, Premium, Premium+, or Premium Business is not counted toward the payable metric. That doesn't make the view worthless — it's still a person seeing your work, still a potential follow, still audience — but it doesn't register for this specific program's calculation.

Do impressions on a quote post of my post count for me?

The definition counts unique impressions from a Premium-tier subscriber (Basic, Premium, Premium+, or Premium Business) where at least half of the post was visible, on the Home Timeline. A quote post is a different post, and its impressions belong to the account that made it. Read the program terms on X's own help pages before you build a strategy that depends on the answer.

If I can't measure the conversion, what number do I plan against?

Historical payout per post on your own account, which is the only number that already has all four filters applied to it. It's backward-looking and audience-specific, and that's a feature: it's measured on the exact population you're going to be paid on next month.

Is it worth buying Premium for my own audience to see more of my posts?

Your subscription affects your eligibility to earn, not who sees you. The Premium-tier filter applies to your viewers, and their subscription decisions aren't something you influence with your own.

The definitions above come from X's published description of the Original Content Rewards Program. Terms change, and the help page is the authority — check it directly before making a plan that hinges on any specific clause here.