How much does Kick pay? Only the 95/5 split is public

Kick publishes exactly one rate, and it is a good one: you keep $4.75 of every $5 subscription, and 95% of every tip, with the other 5% taken as a processing fee. That is Kick’s own help centre, retrieved this morning, and it is the whole of the published rate card.

The thing almost everyone means when they ask this question — the hourly rate Kick pays partners just for being live — is not published anywhere by Kick. Not the rate, not the formula, not the viewer count you need to qualify for it. I checked the revenue-split article, the partner requirements article, and Kick’s streamer site. The only number the requirements page commits to is “stream for at least 5 total hours”. No follower bar, no viewer bar, no rate.

So when you read that Kick pays “$15–20 an hour”, or “$16 to $32”, or that you need “75 average concurrent viewers” to qualify — and you will read all three, because they are on the first page of results — understand what you are reading. Those are third-party estimates. None of the pages carrying them cites a Kick source, because there is not one to cite.

This matters more than it did a month ago, because in September the hidden part of that formula went wrong in public.

What Kick actually said in September

Kick Partner Program creators were paid less than they had earned in September, with rates moving for no reason the creators could see. Kick identified the cause as an internal manual calculation error, corrected the affected figures and issued backpay — which landed in creators’ connected Stripe dashboards rather than showing up in Kick’s own on-site earnings interface, so some people found the money before they found the explanation.

More consequentially, Kick changed the method. Rates had been worked out stream by stream; they are now worked out on a rolling calculation across recent broadcasts. Kick’s stated reasoning, per the reporting, is that the old model let things unrelated to a creator’s work or growth push their rate up and down. Ethan Wright, a director at Kick, put it as: “Something like this tests the trust you put in us and something that no company should ever take for granted.” Kick also said a Trust Centre consolidating KPP eligibility, featuring guidelines and moderation policy was coming.

Label that carefully: the error and the backpay are Kick’s own account, as reported by Streams Charts on 21 September and win.gg. The rolling-window change is confirmed by Kick. The size of anyone’s shortfall is unreported, and I have not seen a single partner’s payout statement.

What nobody has put a number on is the thing the change is actually about: how much a Kick streamer’s audience moves from one broadcast to the next. That part we can measure.

Your audience is not a number, it is a range

We sample every live Grand Theft Auto stream on Kick, Twitch and YouTube once a minute, and each creator’s profile carries their last twelve completed broadcasts with a real duration and an airtime-weighted average audience for each. Take every Kick GTA channel that was live at 04:01:41 UTC on 28 September 2026 and has at least eight completed broadcasts on file — 203 channels, spanning a median of 12 days of streaming — and compare each one’s best recent broadcast against its worst.

The median Kick channel’s best broadcast drew 4.1 times its worst. Not 10% more. Four times.

Ratio of best broadcast to worst	Share of the 203 Kick channels
At least 2×	85.2%
At least 5×	39.9%
At least 10×	20.7%

This is not a Kick pathology, and it would be unfair to write it as one. The same measurement on 796 Twitch GTA channels gives a median of 4.0×, and on 41 YouTube channels 7.0×. Live streaming is simply like this: the same person, doing the same thing, in the same game, draws wildly different rooms on different nights. A rate derived from one broadcast is derived from a coin landing.

Which is a fairly complete vindication of the change Kick made. If your rate is recomputed every stream against a quantity that swings fourfold, your income swings with it for reasons that have nothing to do with you — which is very nearly the sentence Kick used.

The swing is worst for the smallest partners

Here is the part I did not expect, and the part that decides who the change actually helps.

Split those 203 channels at 100 average viewers. The 93 channels below it have a median best-to-worst ratio of 6.1×, and 57.0% of them swing at least fivefold. The 110 channels above it sit at 3.3×, with 25.5% swinging fivefold. Small channels are roughly twice as volatile as large ones.

It is not a clean staircase, and I am not going to pretend it is: the very smallest band, under 25 average viewers, comes in at 5.6× — slightly below the 25–99 band’s 6.1×, probably because a channel with four viewers has less room to fall. The honest version of the finding is a step, not a slope. Under a hundred viewers, you swing about twice as far.

That is the group a per-stream rate punished hardest, and the group a rolling average protects most. It is also, worth saying plainly, the group with the least leverage to query a payout they think is wrong.

What Kick’s formula runs on, for one category

Every Kick earnings calculator on the internet asks you for two inputs — concurrent viewers and hours streamed — and then multiplies them by a rate it invented. The rate I cannot give you. The inputs I can, at least for GTA, which is a category where Kick is genuinely strong: 81,276 of the 202,915 people watching GTA live at 04:01:41 UTC were on Kick, across 288 channels, against Twitch’s 97,585 on 1,395.

Across the 281 of those Kick channels with a usable 30-day profile:

30 days to 28 September 2026	25th pct	Median	75th pct
Hours of GTA streamed	26.4	104.8	205.6
Hours watched	1,022	10,114	34,301
Implied average audience	56	127	302

The median Kick GTA channel we can see put in roughly 25 hours a week and accumulated 10,114 hours watched in a month. Clear the published five-hour requirement, and 85.8% of these 281 do. Multiply that median by the widely quoted “$16 an hour” and you get about $1,700 a month, which is the kind of arithmetic every calculator performs and which I would not repeat as a fact, because the $16 has no source behind it.

What this does not show

The hardest limitation first: our Kick population is not Kick. Kick’s discovery surface caps what we can enumerate, and historically we have reached roughly the top of any given category, which is why the median Kick channel here holds 127 average viewers and the median Twitch one holds far fewer. Read every Kick figure above as describing the visible upper part of GTA on Kick, not its long tail. The Twitch and YouTube comparisons are subject to the same instrument differences and are offered as context, not as a league table.

Second, GTA is one category. A category with a large roleplay audience and a Twitch Drops campaign running to 30 September is not a neutral test bed, and some of the swing measured here is that campaign moving people around.

Third, the session records are the last twelve completed broadcasts, which for a median channel spans about twelve days — so this is a fortnight’s volatility, not a year’s. And four of the Kick profiles report more than 720 hours of airtime inside a 30-day window, which is impossible; they are round-the-clock rebroadcast channels and our aggregation sums concurrent broadcasts rather than taking their union. They are excluded from the airtime table above and the defect is ours, not Kick’s.

Finally, and most importantly: this measures audience, not money. We do not see a single payout. Everything above is the input side of a formula whose output nobody outside Kick can check — which is, in the end, the answer to the question in the headline.

The number that is actually published

If you want a figure you can rely on, it is the subscription split, and it is genuinely the best on the market: $4.75 of $5, against Twitch’s standard 50%. Everything else about Kick’s pay is an estimate wearing a decimal point.

Kick moving to a rolling window is a real improvement and the data says so. A Trust Centre that publishes the rate and the thresholds would be a larger one, and it would put every unsourced calculator out of business overnight.

---

Method and citation. Figures are from luciastream.com’s own one-minute census of every live Grand Theft Auto stream on Kick, Twitch and YouTube. Population census at 04:01:41 UTC on 28 September 2026 (202,915 viewers across 1,765 channels), re-read at 04:13:40 UTC (201,591 across 1,728) with both reconciling exactly against the platform totals and no duplicate channels. Session-level figures are from per-creator profiles covering each channel’s last twelve completed broadcasts. Kick’s rates are quoted from help.kick.com, retrieved 28 September 2026. Our sampling and its known limits are described at /methodology/; the live board is at /live/ and the 30-day creator rankings at /leaderboard/. Permalink: https://luciastream.com/blog/how-much-does-kick-pay/. Journalists and researchers are welcome to quote any of it with attribution.