There is a piece of advice that circulates in every modding community, usually stated with total confidence and no evidence: publish when the platform is quiet.
The logic is intuitive. The Workshop's front page has a fixed amount of room. Post at 20:00 UTC on a Tuesday and you compete with everyone else who decided the same thing. Post at 06:00 and you are the only new arrival in the window, so the small slice of traffic that does arrive lands entirely on you.
We have the data to test that. ModSignal tracks the creation timestamp of every Workshop item, in UTC, alongside its current subscription count. So we took every active mod carrying at least one subscription published between 1 January 2022 and 31 December 2024 — 2,047,042 of them, a cohort old enough that its final subscriber totals have had years to settle — and bucketed all of them by the hour of the day they were published.
The naive answer is that the advice is right. The real answer is stranger, and it says more about the people who give the advice than about the hours themselves.
The curve: a 1.28x gap between the quiet hours and the evening
Here is the outcome that matters — the share of mods in each publish hour that ever reached 1,000 subscribers:
| Publish hour (UTC) | Mods published | Reached 1,000 subs |
|---|---|---|
| 00 | 79,568 | 12.13% |
| 01 | 78,191 | 12.31% |
| 02 | 73,465 | 13.43% |
| 03 | 70,123 | 14.34% |
| 04 | 63,783 | 14.99% |
| 05 | 59,020 | 15.99% |
| 06 | 56,288 | 16.12% |
| 07 | 57,434 | 15.96% |
| 08 | 61,536 | 15.78% |
| 09 | 65,640 | 15.62% |
| 10 | 70,879 | 15.51% |
| 11 | 78,197 | 15.61% |
| 12 | 86,835 | 15.80% |
| 13 | 96,525 | 15.09% |
| 14 | 104,825 | 14.96% |
| 15 | 110,454 | 14.70% |
| 16 | 111,833 | 13.53% |
| 17 | 111,290 | 13.11% |
| 18 | 113,050 | 12.85% |
| 19 | 111,606 | 12.28% |
| 20 | 108,705 | 12.21% |
| 21 | 101,119 | 12.17% |
| 22 | 91,392 | 12.01% |
| 23 | 85,284 | 11.56% |
The shape is unmistakable. Success falls through the evening and bottoms out in the last hour of the UTC day, then climbs through the small hours and peaks in the early morning.
Grouping the two extremes:
- Quiet hours (04:00–11:00 UTC): 15.68% reach 1,000 subscribers. 512,775 mods.
- Evening peak (18:00–23:00 UTC): 12.21% reach 1,000 subscribers. 611,156 mods.
That is a 1.28x difference — a 28% relative lift in the odds of ever breaking 1,000 subscribers, purely from the clock.
It is not a rounding artifact, and it is not a long tail of average subscriptions either. The 09:00 UTC hour averages 2,428 subscriptions per mod, the highest of any hour; 04:00 averages 2,142. The median mod sits between 22 and 35 subscriptions depending on the hour, so nothing here is being driven by a handful of monsters.
The supply side is also working against the crowded window. The evening hours absorb 29.86% of all published mods while producing 26.27% of all mods that reach 1,000 subscribers. The quiet hours absorb a smaller 25.05% of supply and produce 28.30% of the winners. Fewer arrivals, more survivors.
Every standard control holds
This is the point where a careful reader should push back. Publishing at 06:00 UTC is not a random assignment. It correlates with where you live, what you build, and how serious you are. So we ran the controls, one at a time.
Hold the publish year fixed. The gap survives all three years, and gets wider, not narrower:
| Publish year | Quiet hours (04–11) | Evening peak (18–23) |
|---|---|---|
| 2022 | 17.39% | 14.42% |
| 2023 | 15.78% | 12.65% |
| 2024 | 14.31% | 10.26% |
Hold the creator's catalog size fixed. A creator with one mod and a creator with three hundred are different animals. Splitting by their total catalog, the gap holds in every band, and again widens as creators get more serious:
| Creator catalog size | Quiet hours | Evening peak |
|---|---|---|
| Sole-mod creator (1 mod) | 8.02% | 6.19% |
| 2–5 mods | 9.66% | 7.60% |
| 6–20 mods | 13.52% | 10.43% |
| 21–100 mods | 19.63% | 15.36% |
| 100+ mods | 29.92% | 23.92% |
Hold the listing quality fixed. The thumbnail divide and the blank-page problem both showed that an unfinished listing is a much bigger deal than anything else on the page. So we split on whether the mod has a preview image and a description. Restricting to the 1,430,978 mods with a complete listing, the gap narrows but does not close: 23.95% versus 19.39%, still 1.24x.
And the 738,004 mods in this cohort with neither a preview image nor a description reach 1,000 subscribers exactly zero times. Not rarely — zero. (In this cohort the two fields move together completely: every mod missing a description is also missing a preview, and vice versa.) Holding listing quality constant can only shrink the gap by removing that entirely dead weight from the pooled average.
Hold the game fixed. Across the 12 largest games in the cohort by volume, the quiet hours win in 8 of 12. Across all 294 game-and-year pairs with at least 200 mods in each window, the quiet hours win 154, the evening wins 129, and 11 tie.
Remove the structural outliers. Aimlabs, Counter-Strike 2 and Wallpaper Engine are the three ecosystems where the Workshop page is demonstrably not the distribution channel — they distort almost every aggregate on this platform. Take all three out and the gap is 13.30% versus 11.19%, still 1.19x.
So the effect is not a proxy for year, game, creator seriousness, or listing quality. It clears every one of those hurdles.
Which is exactly when you should stop trusting it.
The hour is not a lever — it is a fingerprint
There is a better test available than any of those, and it is the one the pooled numbers cannot pass.
Instead of comparing a quiet-hour mod to an evening mod, compare a creator to themselves. Find every creator who published at least one mod in the quiet hours and at least one in the evening window, then ask which of their windows produced the better rate.
62,215 creators qualify. In 48,904 of those cases the two windows tied — no measurable difference at all. Of the 13,311 creators with a winner and a loser:
- 6,529 did better in the quiet hours.
- 6,782 did better in the evening.
The evening wins. Not by much — a 3.9% edge, inside any reasonable margin — but the 28% advantage is gone completely. Tighten the test further, to the same creator publishing to the same game in both windows — 59,683 comparable creator-game pairs — and the evening again edges it, 6,254 to 5,995.
And the reason the pooled number looked so convincing is visible the moment you stop pooling. Even restricting to the 866,788 mods from creators who publish in both windows, the pooled rates still read 19.51% versus 17.76%. The gap survives a restriction that, at the creator level, has already been shown to be a coin flip. That is the signature of composition: pooled averages are dominated by which creators and which game mixes happen to fall into each window, not by the window itself.
The mechanism shows up in the geography.
The effect is timezone-shaped, and it is absent where the audience is largest
Split the cohort by the creator's known country of registration (a field we hold for 457,324 of these mods — 22.3%) and the picture changes shape entirely.
Who publishes in the quiet hours is not who publishes in the evening. Among creators from Asia-Pacific countries, 44.4% of their output lands in the 04:00–11:00 window, and only 10.1% in the evening. For US and Canadian creators the proportions are the reverse: 21.5% quiet, 35.3% evening.
And the two populations do not have the same base rate. APAC creators reach 1,000 subscribers 30.21% of the time in quiet hours against 25.91% in the evening — while US and Canadian creators sit at 16.82% and 16.13%.
That last pair is the finding. In the largest national group in the dataset, the 28% advantage is 4%.
Run the full hour curve for US-registered creators alone and it does not look like the headline curve at all. It is nearly flat — 16.4%, 16.6%, 16.4% through 18:00, 19:00 and 20:00 UTC, against 16.6% at 06:00. The only genuine dip is 23:00 UTC, at 13.95%.
European creators are the opposite. Pooling British, German, French, Polish, Dutch and Swedish creators, the extremes separate sharply: 25.64% at 04:00 UTC and 25.30% at 05:00 UTC, against 17.60% at 23:00 and 16.88% at 21:00.
Across the 14 countries with enough volume to read in both windows, the quiet hours lead in 11 — but the size of the lead is nothing like the headline. The United States, with by far the most mods, is a +2.3% difference. Russia, France and Sweden invert outright.
So the most defensible version of the claim is narrower than the one that circulates: within the same country, the quiet-hour advantage shrinks from 28% to somewhere between zero and a few percent, and in the single largest market it is essentially nil.
What this actually means
The honest summary is uncomfortable for anyone hoping for a scheduling trick.
The publishing hour correlates strongly with success across the Workshop. It clears year, game, catalog size, listing quality and outlier controls. It is not noise, and it is not a rounding error.
But it fails the one test that would make it actionable. If the hour were a lever, moving your own release from the evening into the morning should move your own odds. It does not. The same creator, given mods in both windows, does no better in one than the other — and if anything does marginally worse in the quiet hours.
What the hour is actually measuring is who you are. It is a fingerprint, and it is legible because the Workshop's creators are distributed across the planet. A creator in Tokyo publishing at 14:00 local time and a creator in Texas publishing at 23:00 local time are not running the same experiment in two conditions. They are different creators, in different ecosystems, with different base rates, and the pooled average reads that difference as a property of the clock.
We can show that the effect is composition. We cannot show from this data why the underlying populations differ — audience geography, the maturity of each regional modding scene, the game mix each region over-indexes on, and the way a mod's earliest hours interact with who is awake to see it are all live candidates, and this dataset does not separate them.
There is one genuinely useful, non-obvious thing in here, though, and it is not about the hour.
The incomplete listings are the real story. 738,004 mods in this cohort — 36% of it — have neither a preview image nor a description, and not one of them reached 1,000 subscribers. Zero. Not one.
If you are choosing where to spend an hour of effort before publishing, the hour of the day is not it. The thumbnail is. The description is. On this evidence the field order is not close — and unlike the publish clock, both of those are things you can actually change.