A Steam Workshop listing has three pieces of text a creator controls: a title, a description, and an image. We have already measured what happens when two of them are missing — the blank-page problem on the description, the thumbnail divide on the image.

The title is the field nobody has looked at, because it is the only one that is never empty. Steam requires it.

It is also the one every player sees first. So we took ModSignal's live dataset of 6,012,057 active Workshop mods carrying at least one subscription, bucketed every one by the length of its name, and measured the outcome that matters: whether it ever reached 1,000 subscribers.

The result is a clean ladder — and, further down, a signal hiding inside a pair of brackets.

The median mod title is 17 characters

A Steam Workshop mod title, at the median, is 17 characters long. The average is 19.4. That is roughly the length of the word "subscription."

Splitting the active, subscribed catalog into six bands by name length:

Title length Mods Share of catalog Avg subscriptions Reached 1,000 subs Reached 10,000 subs
Under 20 chars 3,592,295 59.8% 1,588 9.66% 3.97%
20–29 1,359,931 22.6% 2,602 18.38% 5.44%
30–39 617,941 10.3% 3,063 23.49% 5.75%
40–49 262,200 4.4% 3,178 26.20% 5.72%
50–69 148,342 2.5% 3,331 26.74% 5.68%
70+ 31,344 0.5% 3,150 26.20% 5.61%

3,592,295 active, subscribed mods — 59.8% of everything on the Workshop — are named in fewer than 20 characters. They average 1,588 subscriptions each and reach 1,000 subscribers 9.66% of the time.

Move up into the 50–69 character band and the same rate is 26.74% — a 2.77x difference in the odds of ever breaking 1,000, on a field that costs nothing to fill.

The 10,000-subscriber rate tells the same story in miniature, then stops. It climbs from 3.97% to 5.75% and then flattens. A longer title does not make a mod a hit. It makes it eligible to be one. Above the 30-character line the success rate locks in around 26%; below it, the floor is much further down.

The obvious confound — and why the ladder survives it

A short title never travels alone. It is the visible marker of a listing that was never finished, and it is not the only field such a listing leaves empty.

Sure enough, name length tracks listing completeness almost step for step:

Title length Mods Share with a preview image and a description
Under 20 chars 3,592,296 55.0%
20–29 1,359,930 73.9%
30–39 617,942 82.4%
40–49 262,200 87.3%
50–69 148,343 89.6%
70+ 31,344 89.1%

More than a third of the shortest-titled mods are missing at least one of the other two fields. So the fair objection is that the ladder above is really just the thumbnail divide and the blank-page problem in a trench coat — short name, no image, no description, all the same unfinished listing.

To test that, we re-ran the ladder only over mods that have both a preview image and a description — the 3,881,726 listings where presentation quality is held constant:

Title length Mods Avg subscriptions Reached 1,000 subs
Under 20 chars 1,974,349 2,887 17.58%
20–29 1,004,549 3,522 24.89%
30–39 509,182 3,717 28.50%
40–49 228,871 3,640 30.02%
50–69 132,842 3,719 29.86%
70+ 27,933 3,535 29.40%

The ladder is steeper, not flatter. Among mods that already show a thumbnail and a description, the shortest titles reach 1,000 subscribers 17.58% of the time and the 50–69 band reaches 29.86% — a 1.70x gap that cannot be explained by anything else on the page, because everything else on the page is identical.

The title is its own lever. It is smaller than the image — the thumbnail divide is a 102x effect, this is under 2x — but it is real, it is additive, and it is the cheapest one to pull.

Where length stops mattering

Notice where the ladder ends: 40, 50, 70 characters all sit at 29–30%. There is no benefit to a longer title past the 40-character mark.

That plateau is informative. If title length were a quality signal — a proxy for effort — it would keep climbing. Instead it saturates just past the point where a name stops being a name and becomes a sentence.

What a short title actually costs is search surface. The Workshop's search indexes the title, and a 12-character title contains fewer words to match than a 60-character one. Past 40 characters you have named the thing, said what it does, and named the game, the mode, or the version it serves — and there is nothing left to add that anyone could search for.

Past that line, extra words do no work. Below it, there is almost certainly a word missing that a player would have typed.

The titles inside square brackets

Length is not the only thing in the title. So we checked the formatting habits creators use — brackets, parentheses, digits, version numbers, ALL CAPS — first across the whole catalog:

Feature in title Mods Reached 1,000 subs
Square brackets [ ] 268,315 35.80%
Parentheses ( ) 503,683 22.65%
Any digit 1,728,544 15.81%
Version-style 2.0, 1.4 141,106 18.04%
ALL CAPS 438,215 8.58%

Square brackets stand out immediately — 35.8% against a catalog average of 20.9%. But that table is contaminated twice over: brackets are more common in long titles, and long titles are more common in games where brackets are normal.

So we held length constant, restricted to a single 40–59 character band, and required a complete listing. In that matched pool of 323,667 mods:

Feature Mods Reached 1,000 subs Mods without it Reached 1,000 subs
Square brackets 55,398 41.8% 268,269 27.6%
Parentheses 106,429 29.3% 30.1%
Any digit 146,903 30.5% 176,764 29.6%
ALL CAPS 4,059 28.0% 319,608 30.1%

Most of the formatting habits evaporate. Parentheses do nothing. Digits do nothing — the apparent catalog-level "number penalty" was entirely a length artifact. ALL CAPS do nothing.

The bracket survives. In a length-matched, presentation-controlled pool, a bracket-titled mod reaches 1,000 subscribers 41.8% of the time against 27.6% for the same-length mods without one.

It is not a couple of outlier games driving it, either: excluding Aimlabs and Counter-Strike 2 — the two ecosystems that dominate bracket usage — the matched effect is unchanged, at 41.8% versus 27.6%. And it holds inside individual games, where taste, tag conventions and competition are all held roughly constant:

Game Mods with brackets Reached 1,000 subs Without Multiple
Trailmakers 2,070 9.4% 1.1% 8.5x
Avorion 3,651 33.2% 10.6% 3.1x
Stormworks 2,911 45.5% 18.1% 2.5x
Scrap Mechanic 4,176 13.4% 5.5% 2.4x
Unturned 2,315 28.7% 12.9% 2.2x
Ravenfield 3,099 73.9% 39.3% 1.9x
Project Zomboid 8,197 42.9% 27.5% 1.6x
Cities: Skylines 6,119 71.2% 45.9% 1.6x
Garry's Mod 39,622 60.7% 49.4% 1.2x
Wallpaper Engine 15,777 95.5% 93.4% 1.0x

In eight of the ten largest bracket-using ecosystems, bracketed titles outperform their unbracketed neighbours — sometimes by a factor of two or three.

What a bracket actually marks

Brackets are not a marketing flourish. Look at what the most-subscribed bracket titles on the platform actually are:

  • training_aim_csgo2 [CSGO Legacy]6,973,836 subscriptions
  • Four Seasons Warmup [Template] — 6,943,050
  • Dust 2 [Creator Studio] Template — 6,940,125
  • Yprac Mirage Guide [CS2 in Description] — 5,728,568
  • Pudge Wars [7.29 Fixed] — 5,136,647
  • [AB+|Rep(+)] External Item Descriptions — 4,377,487
  • Tsar's Common Library 2.0 [Only for B41] — 4,303,216

Every bracket here answers one question: which build, which channel, which dependency does this mod require? [CS2 in Description]. [7.29 Fixed]. [Only for B41]. [Template].

That is the opposite of decoration. A bracket is a compatibility tag — and a compatibility tag only exists because other mods depend on this one. Which is why the bracket population's engagement profile is so strange:

Bracket titles Everything else
Mods 268,315 (4.5% of catalog) 5,743,742
Share of all subscriptions 11.5% 88.5%
Avg subscriptions 5,389 1,935
Favorites per subscription 0.074 0.443
Views per subscription 1.37 1.28

4.5% of the catalog holds 11.5% of every subscription on the platform — 1.45 billion subscriptions concentrated in mods with brackets in their names.

And then the tell: bracket-titled mods earn six times fewer favorites per subscription. People install them and do not love them. They are not browsing for them; they are resolving a dependency. A bracket is what a mod is called when it is required rather than desired.

The composition bears this out. Despite the intuition that brackets mark libraries, only 0.05% of bracket titles contain the word "library" and 0.12% contain "template" — the tag is not self-describing. What it marks is the whole class of mods that exist to make something else work: version gates, compatibility patches, translation variants (1.38% name a language), and the configuration presets that a share code or a server installs without anyone ever reading the name.

The game that inverts it

One ecosystem inverts the bracket effect so completely that it is worth stating plainly, because it defines the boundary of the rule.

On Easy Red 2, bracket-titled mods reach 1,000 subscribers 4.7% of the time. Their unbracketed neighbours reach 36.8% — an 8x penalty in the exact direction opposite to every other game in the table.

Easy Red 2 is a game where a large share of bracket usage appears to be bulk-imported content — map and asset packs named by their source, published in volume. Brackets there are not a compatibility tag; they are a naming convention inherited from an upstream pack, and they mark the opposite population: the mass-uploaded tail we measured in the one-sitting problem.

That is the honest reading of the whole bracket effect. A bracket is not lucky. It is a label, and what it labels — a dependency, a build gate, a bulk import — is what actually determines the outcome. On most of the Workshop the label catches infrastructure. On Easy Red 2 it catches a dump.

What the title is for

Put the two halves together and the Workshop's title field has a clear shape.

The median mod spends 17 characters on a name that will never be searched for. Filling it out to 40–60 characters is worth roughly a doubling of the odds of ever reaching 1,000 subscribers, and it is worth that even after the mod already has a thumbnail and a description. The brackets that outperform everything else in the catalog are not doing it through style — they are the visible edge of the dependency layer, the part of the Workshop that exists to make other mods function.

The image is the gate. The description is the proof. The title is the address — and 60% of the Workshop has given itself a short one.


All figures from ModSignal's live Steam Workshop dataset, snapshot 2026-09-24: 6,012,057 active items carrying at least one subscription, 12.56 billion subscriptions total. Related: the blank-page problem, the thumbnail divide, the one-sitting problem.