Most of the argument about community models is conducted entirely in anecdote. Someone's paid community is thriving, so paid wins. Someone's free community has 8,000 members, so free wins. Here is what it looks like when you measure the whole platform instead.
What model do most Skool communities actually use?
Across 6,521 communities whose membership model can be identified from outside, the split is not what the loudest advice implies.
About 44% run a free door with paid rooms behind it. That is the plurality — more common than a pure free community and more common than a pure paid one.
By size band:
It is the most common model at every size on the platform. Meanwhile, only about 18% of communities over 5,000 members are pure paid — at scale, the fully gated door is the exception.
The limit worth stating: 49% of the mapped corpus has no membership model recorded at all, so this describes the half that can be read from outside. It is also a front-door observation — it records what a community advertises on the way in, not what it sells once you are inside.
Why does the hybrid model dominate?
Because it solves the problem the other two models create.
A pure paid door is a filter applied before anyone has seen anything. It collects a smaller, more committed roster, and it caps your top of funnel at people willing to pay on the strength of a sales page.
A pure free door collects everyone, including the large share who join on impulse and never return. It builds an impressive member count on top of a roster where most people have no particular reason to come back.
The hybrid takes the free door's reach and puts the filter inside, where members can see what they would be paying for. That is a structurally better position, and the measurement bears it out in the place that matters most.
Freemium communities lead show-up rates in all four size bands. They beat pure-free and pure-paid at every scale measured. That is a clean result, and it has an uncomfortable implication for benchmarking: any figure you have read about how "free communities" perform is blending two completely different businesses. A free-door-with-paid-rooms community and a genuinely free community share a price of zero at the entrance and almost nothing else.
Does a paid door actually make people participate?
A little, in a specific range, and for reasons that are probably not the ones you would assume.
In communities between 200 and 5,000 members, a paid door runs about 1.4 times the contribution rate of a free one: a median 9.1% of members have ever posted behind a paid door against 6.6% behind a free one. The result is statistically solid — a 95% confidence interval of 1.13 to 1.81, p = 0.0007.
Now the three qualifications, all of which matter:
It does not hold everywhere. Below 200 members the difference cannot be distinguished from no difference at all. Above 5,000 there are no readable paid communities, so nothing is claimed. The finding is specifically about the middle of the platform.
No individual size band publishes a multiplier. Between 200 and 999 members the gap fails to separate from zero once tested rather than eyeballed. Above 1,000 the effect is real but lands anywhere between 1.1x and 3x — too wide to state as a number. Only the pooled figure survives, and that is deliberate.
It shrank as the sample grew. The 200-999 band read 2.0x at an early cut, 1.5x at a larger one, and 1.2x at the finished crawl with an interval that includes no difference. That is the signature of an effect that was partly an artefact of a small sample. The pooled figure held through all three cuts; the band-level one did not.
So is the paid door causing the participation?
Probably not, or not mostly — and this is the single most important sentence on this page.
A paid roster is filtered by willingness to pay before anyone arrives. A free roster collects people who joined on impulse and never came back. Some of that 1.4x is a cleaner denominator rather than a door that makes people write.
Read it as a difference between two kinds of roster, not as a lever you can pull. Putting a price on an existing free community does not convert your quiet members into contributors — it removes some of them from the count, which improves the ratio without improving the community. If your goal is participation rather than a better-looking metric, the work is elsewhere: why your Skool community is quiet covers what actually moves it.
Which model should you pick?
The data supports three fairly concrete recommendations.
If you are starting, the hybrid is the measured default. It is the plurality model at every size, it leads show-up rates at every size, and it gives you a top of funnel that a pure paid door does not. The burden of proof is on the other two options.
If you run a genuinely free community, know which benchmark applies to you. You are in the arm of the data with the messiest roster and the lowest contribution rates, and comparing yourself to blended "free community" figures will flatter you or panic you more or less at random. Compare against pure-free communities in your size band.
If you run a pure paid community, your advantage is the roster, not the door. You have fewer, better-qualified members. The 1.4x is describing who you let in. Everything downstream — welcome, first two weeks, whether people ever write anything — is still the same job it is for everyone else.
And whichever you pick, the price is a separate decision with its own evidence behind it: the median priced Skool community charges $19 a month, and what Skool communities charge has the full ladder.
These figures come from The State of Skool 2026, an independent measurement of 18,412 Skool communities. See the full report for the full model analysis and 21 other findings.