CommunityOS classifies every real community member into one of four archetypes: Champion, Amplifier, Builder, Early Adopter. Four is not arbitrary — it is the number at which every class still demands a different action.
A classification system for community members is only useful if each class answers the operator's question: what should I do with this person? That constraint — actionability, not descriptive richness — is what fixed the number at four. This post reconstructs the reasoning, including the archetypes that did not make the cut.
What are the four archetypes?
- Champion — the voice. Speaks about the project with depth and conviction. When a Champion posts, people who trust them listen. The rarest class and the highest value per member. The ask: give them something substantive to talk about, early.
- Amplifier — the reach. Wide audience, fast distribution. Carries a message into rooms the brand cannot enter; the multiplier on everything Champions say. The ask: make sharing effortless at the right moment.
- Builder — the substance. Ships write-ups, tools, threads with original work in them. Builders create the assets a community keeps returning to. The ask: access, materials, and recognition for what they make.
- Early Adopter — the first wave. Shows up before it is obvious, tests, gives the honest first signal. Where every Champion started. The ask: first access and a short feedback loop.
Two independent currencies underlie the set — reach (can this account distribute?) and substance (does this account generate conviction, content, or signal?). Champions and Builders are substance-heavy; Amplifiers are reach-heavy; Early Adopters are a timing signal that predicts future substance. The 60/40 scoring model weighs language over vanity metrics precisely because substance is the scarcer currency.
Why is three too few?
Every three-class scheme we tested merged two groups that need opposite treatment. Collapse Builders into Champions and you send "please post about us" asks to people whose actual value is shipping artifacts — and who find repost-style asks mildly insulting. Collapse Early Adopters into Champions and you treat a predictive signal (arrived early, worth cultivating) as a present-tense one (advocating now, worth activating), which wastes the cultivation window. Each merge produced measurably worse mission-to-completion fit in exactly the way you would expect: wrong ask, no action.
Why does five collapse into noise?
Because the fifth class never survived the actionability test. Every candidate either mapped onto an existing archetype's action, or described accounts that should not receive outreach at all. The graveyard:
- The Lurker. Real, present, silent. Proposed constantly, rejected every time: there is no mission for someone whose defining behavior is inaction. Lurkers stay in the scored population and get monitored — several later crossed into Early Adopter behavior — but "monitor" is a queue state, not an archetype.
- The Critic/Skeptic. Vocal, engaged, negative. Tempting as a class, but operationally it splits in two: high-substance critics are Builders or Champions with a sharp edge (same asks work, often brilliantly), and low-substance ones are noise. The tone axis did not change the action.
- The Whale. Financially significant, communicationally invisible. A real category for a token team, but it belongs to holder analytics, not community activation — nothing about a wallet balance tells you what to ask on X.
- The Mod/Insider. Already inside the tent. Known to the team by name; a CRM row, not a discovery output. A scanner that "discovers" your own moderators is doing make-work.
Five classes also degrade the human interface. A queue where every row is one of four colors reads at a glance; operators internalize the set in a day. Past four, adjacent classes blur, and a classification an operator cannot instantly act on is decoration.
Where does the set sit in the literature?
The lineage is visible and worth being upfront about. "Early Adopter" descends from Everett Rogers' diffusion of innovations; the reach/substance split echoes decades of two-step-flow research on opinion leaders. The difference is purpose. Academic taxonomies describe populations. This one routes work: each archetype exists because it binds to a distinct mission template in the activation surface, and completed missions are verified through Proof Review before they count.
How stable are archetypes over time?
Deliberately not fixed. People move — the canonical path is Early Adopter → Builder or Champion as involvement deepens — and snapshots between scans track those transitions. The set of four is stable; membership is not, and the transitions themselves are signal: a wave of Early Adopters crossing into Builder territory says more about a community's health than any engagement rate. Definitions for every term in this post live in the glossary.