Every Steam Workshop creator starts at zero. The first mod is the one nobody has any reason to look at: no follower base, no reputation, no search history, no reason for the algorithm to believe you exist.
We wanted to know how much of what happens next is decided in that first release. So we took every active Steam Workshop mod published between 1 January 2022 and 31 August 2026 — 4,582,663 items from 1,382,623 creators across 390 games — ordered each creator's catalog by publish date, and asked a single question of the mods that came second, third and later: did this creator ever land another 1,000-subscriber mod?
The answer turns almost entirely on what the first one did.
The debut ladder
Restricting to the 318,157 creators who published at least three mods in the cohort — so that "later" means something — and bucketing them by what their debut achieved:
| Debut mod's subscribers | Creators | Landed another 1,000+ mod | Landed another 10,000+ mod |
|---|---|---|---|
| Under 10 | 146,018 | 2.82% | 0.50% |
| 10 – 100 | 87,294 | 9.99% | 1.46% |
| 100 – 1,000 | 54,146 | 34.16% | 6.44% |
| 1,000 – 10,000 | 24,153 | 76.47% | 24.70% |
| 10,000+ | 6,546 | 92.00% | 64.86% |
Collapsed to the three bands that matter:
- Debut under 1,000 subscribers: 287,458 creators. 10.90% ever land another 1,000+ mod.
- Debut 1,000–10,000: 24,153 creators. 76.47% do.
- Debut 10,000+: 6,546 creators. 92.00% do.
This is a monotone ladder with no threshold and no plateau. Every step up the debut band multiplies the odds of a second hit — 3.5x from the bottom band to the next, 2.2x again, then 2.4x, then 1.2x. The curve is steepest exactly where most creators live, which is the part that matters.
For scale, the platform baseline: 10.00% of the 4,582,663 mods in this cohort ever reached 1,000 subscribers. A creator whose debut flopped is below the platform average for everything they publish afterwards. A creator whose debut broke out is an order of magnitude above it.
The supply side is brutal
Here is the uncomfortable half of the same table. Of the 318,157 creators who have published three or more mods since 2022:
- 90.35% — 287,458 — never got their first mod past 1,000 subscribers.
- 7.59% landed in the 1,000–10,000 band.
- 2.06% crossed 10,000.
Roughly two in three creators who have ever published to the Workshop in this window (859,849 have exactly one cohort mod) are in the first bucket by definition, and once they are in it, nine out of ten of them never produce a hit at all.
The good news is on the other side of the line and it is real: a debut breakout is close to a guarantee of a second one. 92% of creators who crossed 10,000 on their first release did it again. The Workshop does not have a one-hit-wonder problem at the top of the debut distribution. It has a one-shot problem at the bottom.
It is not catalog size, and it is not the year
The obvious objection is that a big debut is just a proxy for a serious creator who was always going to publish more, and better. So we held catalog size fixed:
| Creator's cohort mods | Debut 1,000+: went on to another 1,000+ | Debut under 1,000: went on to another 1,000+ |
|---|---|---|
| 3 – 5 mods | 69.6% | 6.3% |
| 6 – 10 mods | 84.2% | 12.1% |
| 11 – 30 mods | 91.9% | 19.8% |
| 31+ mods | 97.5% | 32.1% |
Inside every band the debut separates the population by roughly an order of magnitude, and the average catalog size in each row is essentially identical between the debut-hit and debut-flop groups (3.8 vs 3.7 mods in the smallest band; 90.6 vs 77.1 in the largest). This is not the debut standing in for "how much this person publishes."
Tighten it to the creators with the least room for a lucky outcome — exactly three to five mods, where a debut hit and a debut flop had the same number of chances:
| Debut band (creators with 3–5 mods) | Creators | Landed another 1,000+ mod |
|---|---|---|
| Under 100 | 136,333 | 2.98% |
| 100 – 1,000 | 27,618 | 22.80% |
| 1,000 – 10,000 | 11,691 | 64.95% |
| 10,000+ | 3,073 | 87.18% |
And it is stable across publish cohorts — a debut that crossed 1,000 predicted a later 1,000+ mod for 82.2% of 2022 debutants, 78.7% of 2023 debutants and 79.8% of 2024 debutants, against 17.0%, 11.7% and 9.1% for debut flops. If anything the penalty for a slow start is getting worse as the Workshop fills up.
It is not the game, either
Ranking the largest game ecosystems in the cohort by the same split — "did a debut breakout predict a later hit" versus "did a debut flop ever recover":
| Game | Creators | Debut 10,000+: later 1,000+ | Debut under 1,000: later 1,000+ |
|---|---|---|---|
| Trailmakers | 24,001 | 100.0% | 2.1% |
| Scrap Mechanic | 13,068 | 66.7% | 6.6% |
| Stormworks | 11,552 | 94.3% | 8.9% |
| Space Engineers | 11,435 | 92.0% | 5.0% |
| Portal 2 | 10,660 | 100.0% | 4.0% |
| People Playground | 9,998 | 90.8% | 4.2% |
| DayZ | 9,526 | 80.7% | 9.8% |
| WorldBox | 9,111 | 58.1% | 5.2% |
| Garry's Mod | 8,504 | 93.6% | 44.6% |
| Cities: Skylines | 7,215 | 97.7% | 8.1% |
| Barotrauma | 7,206 | 91.3% | 9.8% |
| Arma 3 | 7,157 | 88.1% | 8.4% |
The gap is enormous in every ecosystem on the list, and it is not an artifact of one dominant game. Garry's Mod is the outlier in the flop column — 44.6% of its debut flops recover, far above anywhere else — which is worth its own investigation: it is one of the few ecosystems where the discovery graph appears willing to resurface a creator who started badly.
Then the signal inverts on itself
The debut ladder says your first mod sets your trajectory. It does not say your first mod is your best mod. Those turn out to be two very different claims, and the difference is publishing rhythm.
Take every creator in the cohort with exactly five mods, and split them into two groups: those whose five mods landed across five different months, and those whose five mods all landed inside a single calendar month. Then rank each creator's five mods against each other by subscribers. Under a null of no effect, the first mod is the group's best 20% of the time, and the worst 20% of the time.
| First mod is best | First mod is worst | Fifth mod is best | Fifth mod is worst | |
|---|---|---|---|---|
| Spread across 5 months (3,237 creators) | 25.21% | 20.70% | 16.43% | 24.65% |
| All inside one month (8,077 creators) | 20.47% | 34.10% | 29.64% | 14.36% |
| All inside one day (subset of the above) | 19.21% | 41.66% | 34.79% | 10.93% |
The two groups are mirror images.
When a creator spreads their mods out, the natural pattern shows up: the first mod is over-represented as the best in the catalog (25.21% against a 20% null) and the last is over-represented as the worst (24.65%). That is compounding — an earlier release has had more time to accumulate subscribers, and in most catalogs it is also the founder of whatever audience the rest inherit.
When a creator dumps five mods into a single month, that pattern flips completely. The first mod becomes the most likely one in the entire catalog to be the worst (34.10%, against a 20% null), and the last becomes the most likely to be the best (29.64%). Push further and pack all five into a single day, and the first mod is the group's worst 41.66% of the time while the fifth is the group's best 34.79% of the time.
That inversion is the finding. It also runs against the obvious confounder: mods accumulate subscribers over time, so the earliest mod in a burst has had weeks or months longer to gather an audience than the last one. It should win on time alone. Instead it loses. The only thing that changed between the two columns is whether the creator's own releases were competing with each other.
This is the within-creator version of the crowding penalty and the one-sitting problem. Those posts showed that publishing into a crowded game-hour costs a mod roughly two thirds of its hit rate, and that mods fired in a batch of 31+ land 9x worse than mods published alone. This is the same effect measured against yourself: your own second, third and fourth uploads are the competition that buries your first.
The pattern is not a game-mix artifact. Restricting to bursts where all five mods went to the same game — the airtight version, since there is no cross-game difference left to explain — the numbers are identical to two decimal places (first is worst 34.12% vs 34.10%).
It does not scale down with burst size either, which is what you would expect if it were a real mechanism rather than a rounding effect. Within same-month bursts, the first mod's odds of being the group's worst run 36.2% at four mods, 34.1% at five, 30.4% at six and 27.2% at eight — decaying toward the null exactly as self-crowding should, since a larger batch spreads the damage across more rivals. Spread-across-months catalogs show the opposite gradient, with the first mod's share of "worst" falling far below the null at every size.
One honest caveat, stated plainly because it matters for how much weight this deserves. The burst effect is concentrated in the low-subscription tail, where most mods sit at zero or single digits and ties are everywhere. Filter the same bursts to require every mod to have at least one subscriber, and the inversion survives but shrinks (first-is-worst 26.49% against a 20% null). Require every mod to have at least 50 subscribers, and it disappears — 1,204 creators, and the first mod is no longer distinguishable from the middle. So: when a five-mod burst goes nowhere, the first mod is disproportionately the one that goes nowhere hardest. When all five mods get real traction, position in the burst stops mattering. That is a real finding about where attention dies, and it is not a claim that slot order is a lever for a mod that was going to succeed anyway.
The exceptions are real, and they are rarer than you would like
The ladder is a strong tendency, not a rule, and the creators who beat it are worth looking at directly.
Of creators whose debut pulled 200,000+ subscribers and whose later mods stayed small:
| Creator | Debut game | Mods | Debut subs | Best later mod |
|---|---|---|---|---|
| Sidies | Mount & Blade II: Bannerlord | 2 | 439,950 | 4,771 |
| Lee Koldun | Garry's Mod | 3 | 359,525 | 4,689 |
| (creator 76561199155182026) | Wallpaper Engine | 2 | 700,528 | 4,508 |
| (creator 76561198958996672) | Don't Starve Together | 5 | 723,266 | 3,062 |
| RobertMod | MECCHA CHAMELEON | 2 | 200,417 | 3,031 |
And the reverse — creators whose debut did essentially nothing, and who broke out anyway:
| Creator | Mods | Debut subs | Best later mod |
|---|---|---|---|
| dug | 5 | 29 | 3,501,778 |
| L9Heb | 129 | 2 | 1,701,679 |
| Nickoz | 4 | 5 | 1,252,767 |
| Foteq | 2 | 0 | 1,164,992 |
Nickoz got five subscribers on a first mod and later published one that reached 1.25 million. EmpireDeadman has 1,103 cohort mods after a debut that pulled 42 subscribers, and a later mod at 918,000. The tail exists. It is just thin: 10.9% of debut flops ever reach even 1,000 subs on a later release, and only 0.50% of the very worst debuts (under 10 subscribers) ever reach 10,000.
What this actually means
The debut signal is a prediction, and the mechanism behind it is genuinely ambiguous. Three explanations fit the data and this dataset cannot separate them:
- Skill and seriousness are persistent. A creator who can get a first mod to 10,000 is the same creator who can do it again; a creator who cannot, mostly cannot. The debut is a measurement of the creator, not a cause of their outcome.
- Audience continuity. A debut with 10,000 subscribers leaves behind 10,000 players with the mod installed and a Workshop page that now has a following — an installed base that the second mod inherits directly. On this reading the debut is a cause.
- Platform treatment. The Workshop's discovery systems may reward accounts with a proven item, giving the second mod a better first-week position than the first one ever got.
All three would produce the curve above, and the honest position is that we cannot yet say which dominates. What we can say is that the curve is real, large, monotone, and stable across games, years and catalog sizes — and that the exceptions, while real, are rare enough that "your first mod was a flop and it's fine" is bad advice at population scale.
The one part of this that is unambiguously actionable is the second finding, because unlike the debut outcome it is a decision you control before you know the result.
Publish your mods apart, not together. A creator who releases five mods one per month ends up with a catalog whose first entry is its best entry 25.2% of the time. A creator who releases five mods inside one month ends up with a catalog whose first entry is its worst entry 34.1% of the time — and if all five land on the same day, 41.7%. Nothing about the mods changed. Only whether they had to compete with each other.
The debut is the single highest-leverage release a creator will ever make on the Steam Workshop. Treating it as one of five things you upload on a Tuesday evening is the most expensive mistake in the data.
Method note: cohort is every item in ModSignal's live Steam Workshop graph with is_active true, visibility public, and a creation timestamp between 1 January 2022 and 31 August 2026 — 4,582,663 mods from 1,382,623 creators and 390 games. "Debut" is the creator's earliest item by creation timestamp; ties broken by item id. A "hit" is 1,000 or more current subscriptions. Rank comparisons use rank() on current subscriptions within creator, so creators with any tied mods contribute those ties to the middle of the rank range rather than to either extreme. Related: the success ladder, the portfolio effect, the one-sitting problem, the crowding penalty.