TikTok Live statistics, measured from real streams

The gap between the median and the average is the whole story of TikTok Live earnings. Gifting is extremely top-heavy: a small number of streams collect most of the diamonds, so any figure quoted as an "average" describes almost nobody. Where a published guide gives you one number for what a live earns, it is usually an average, and usually drawn from streams that were worth writing about.

The concentration repeats inside a single stream. In the median session with any gifting at all, one person accounted for 53% of the diamonds. For a creator, that means earnings hang on a handful of recurring supporters rather than on viewer count, which is why knowing who those people are matters more than the concurrent-viewer number.

These figures come from a sample of creators TikMonitor users chose to track, not a random sample of TikTok, and they cover only sessions we watched end to end. That biases the set toward creators someone considered worth following. We publish it because the alternative figures in circulation have no stated source at all.

The four gifter figures — the median top-gifter share, the Gini, the top-1% share and the count of gifters who make up half the total — are drawn from a shorter window than everything else on this page: 2026-08-05 to 2026-09-27, covering 1,405,408 distinct gifters. We began recording per-gifter detail later than session totals, and the raw events behind the earlier months have since been deleted, so that window cannot be extended backwards. Read them as a picture of that period rather than of the full range above.

Time on air pays, and it pays at an increasing rate. Split every broadcast into ten equal groups by length and the median earnings RATE climbs the whole way up: 0 diamonds an hour in the shortest tenth against 8,468 in the longest. That is worth stating plainly because we have previously described gifting as clustering around moments rather than accruing with minutes — on this data that is wrong. The honest caveat is that causation runs in both directions: a live that is going badly gets ended early, so the shortest streams are partly a record of streams that were already failing.

The clock matters and the calendar does not. The best UTC hour to be live earns roughly 5.0x the median rate of the worst one, while the spread between the best and worst day of the week is only 2.9x. We are reporting the weekday number precisely because it is unremarkable: "post on Fridays" is the kind of advice that gets repeated without anyone checking, and here it is worth almost nothing. Note that a broadcast is credited to the hour it STARTED, so a stream that runs four hours counts once, at its start.

The gift mix is as lopsided as the gifter list. 30% of every gift sent is worth five diamonds or less, and all of those together come to under 1% of the money. Half of all diamonds come from just 12 of 2,175 gift types. The roses and hearts that fill a chat are engagement, not income, and a creator reading the gift feed as a revenue signal is reading the wrong column. These gift figures cover 2026-09-21 to 2026-09-27 only — the raw event stream behind them is kept for days, not months, so this is a snapshot rather than a history.

The payout-minimum figure counts individual broadcasts, not creators. A creator whose single lives each fall short can still cross TikTok's threshold by accumulating across several of them, so read it as "this is what one live typically brings in", not as the share of creators who never get paid. It also measures gross diamonds at the estimated $0.005 rate, which is not the same as a withdrawable balance.