Every community metric in common use treats a member as if your community were the only one they are in. None of them are. Pull rate is the metric that stops assuming otherwise, and the reason it is worth knowing is not the average — it is the spread.
The definition
Pull rate is the share of a member's total Skool activity that they give to one community.
It is a per-member quantity that can be summarised across a roster. Where creator rate and untapped share ask how many of your members do something, pull rate asks how much of what they do happens here.
What the number is
Among 5,265 member-community pairs in communities where we hold full activity data, 71.5% do nothing in a given community while remaining active elsewhere on Skool. That is the important half of the sentence. They have not left the platform, they have not gone cold on the internet, and they are not unreachable. They are busy somewhere else.
Of the members who do something in the community:
| Share of their Skool activity given to this community | Members |
|---|---|
| More than 25% | 0.78% |
| More than half | 0.27% |
| 90% or more | 0.04% |
Roughly one member in 370 gives a community the majority of their platform activity. The median active member gives it 0.2% of what they do on Skool.
Read this as a description of the internet, not a report card
The reflex on seeing those numbers is to take them personally. They are not personal. Divided attention is the ordinary condition of every online space: people belong to many, give each one a sliver, and treat almost none of them as home. A roster where nearly nobody is a majority-loyal member is what a normal, healthy Skool community looks like from the inside, and the same would be true on any platform that could measure it.
It is also worth putting next to what happens when those members do show up. On the same formula the advertising industry uses — interactions divided by audience size — a post in the communities we track engages a median of 5.1% of members, against 0.30% for a median Instagram post and 0.02–0.23% for organic Facebook (both from the Rival IQ / Quid 2026 benchmark report). Those are brand accounts posting to followers rather than owners posting to members, and organic reach on both platforms is actively suppressed, so the situations differ even though the formula does not.
The two facts sit together. Members give any one community a small share of their attention — and when they spend it with you, they spend more of it than the platforms they came from ever got.
These are floors, and here is why
The denominator counts every recorded action on Skool, including passive ones. The numerator counts only the posts and comments we hold. Both halves of that mismatch push the measured share downward, so the true proportion of majority-loyal members is higher than 0.27% — possibly materially higher.
That is why these figures are used qualitatively. The ordering is trustworthy; the exact percentages are lower bounds and should be read as such.
The scope, stated plainly
This is measured in communities that use Skoolgrades, with full engagement data available by consent. It is not a platform-wide statistic and is not presented as one.
Those owners chose an analytics product, so they skew engagement-conscious, and member enrichment runs in score order, so the sample skews toward active members. Both biases argue for treating this as a well-measured statement about a specific set of communities rather than a claim about Skool.
An independent measure of the same construct, drawn from public data rather than from customers, is being collected. Until it either corroborates this or exposes it as a sample artifact, pull rate stays where it is.
The version you can act on
The aggregate is context. The per-member figure is the one that changes a decision.
Two members with identical activity inside your community can have completely different pull rates. One gives you 40% of their Skool time; the other gives you 0.5% and spends the rest elsewhere. They look the same in your feed and they are not the same person. The first is a candidate for the small group who carry a community — see the engine share. The second is where attention is quietly leaking, and knowing that is worth more than any roster-wide average.
That comparison needs a view of what a member does across the other communities they belong to, which is not something a community's own feed can show you.