Methodology write-ups, position pieces, and operator notes — the cornerstone set, built on real production numbers, not hand-waving.
Impressions and follower counts are the metrics every rights holder already has - and the ones that collapse the moment finance asks how many of these are real people. The report built on a verified audience and named fans, instead.
Read the post →A media-value report answers a different question than the one sponsors are now asking. What actually counts as proof that an audience is real, and the report that survives a sponsor's own analytics team.
Read the post →Nobody has published a credible number, and free bot checkers don't answer it either - they sample or infer from engagement rate, both of which a bot network can fake. What 'real' should mean, and the only way to actually find out.
Read the post →Both turn a task list into a reward and a reward into a headcount. They differ in where they live and how they check for humans - and they share the same blind spot once the campaign ends.
Read the post →They get compared because both end in 'pipeline'. Clay finds strangers to contact, Common Room finds members already warming up - and neither activates anyone. How to choose, and the third option.
Read the post →Apollo fills a rep's queue from a database of strangers; Common Room fills it from members already in your community. The right pick depends on which pipeline motion you actually run.
Read the post →Brandwatch is for the people who need to understand the conversation; Khoros is for the people who have to run it. Where they overlap, where the overlap misleads, and the job neither does.
Read the post →Management software runs the room. Intelligence tells you who in the room matters. They get bought interchangeably, which is why teams have three moderation tools and still can't name their ten most valuable members.
Read the post →A Web3 community stack is four jobs, not one tool: moderation, quests, listening, intelligence. The honest strengths and limits of each category, and the order to buy them in.
Read the post →Most dashboards report activity over bot-inflated denominators, which is why leadership stops reading. Real-member base, reciprocity rate, archetype mix, verified actions - with numbers from a real scan.
Read the post →An operator's comparison of the two listening heavyweights - where each wins, where they are identical, and the question to ask before buying either: listening, or activation?
Read the post →The first production scan and what the numbers actually meant. Bot-Kill rates, archetype mix, and why "5,806 real community members" is the metric that matters.
Read the post →The math of rented reach versus owned activation. Why one $50,000 KOL post will always lose to a $5,000 community activation spend on the same audience.
Read the post →Reproducibility, audit trails, cost structure. The case for rules-based archetype classification in a category that has decided language models are the answer to everything.
Read the post →Three is too few to differentiate; five collapses into noise. The full reasoning behind the archetype set, including the candidates we considered and rejected.
Read the post →Activity floor, follower-to-following imbalance, posting-burst patterns, and linguistic uniformity. The four signals that explain how 90.96 percent of a Web3 project's follower list got filtered.
Read the post →What gets verified, what does not, and how the verified-only ledger compounds into reporting your client's CFO will actually trust.
Read the post →The method, step by step: filter the bots, score real accounts on reach and conviction, rank the queue, and verify. How to surface the few hundred who actually matter.
Read the post →The four tiers of follower tooling, the two features most tools still skip (bot filtering and reproducible scoring), and how to run the evaluation on your real list.
Read the post →An operator's playbook: work the ranked queue, match the ask to the archetype, send personal outreach, verify every completion. Ten a day, no automated mass-DMs.
Read the post →Four categories that share a word but do different jobs — listening, quests, sales intel, community intelligence. How to pick by outcome instead of feature list.
Read the post →Why finance dismisses community reporting, and how to build the version that survives the room: verified actions over real members, evidence per line, cost against outcomes.
Read the post →Community work is treated as overhead because its output cannot be checked. How to reframe it as a priced, verified, defensible line item — with cascade economics.
Read the post →New posts land at the top of this list. Subscribe via @communityxos on X for the post drop.
The cornerstone list is opinionated, not exhaustive. Tell us what you would want to read about community intelligence; we will write it.