The seven triggers. Price moves above a defined threshold for your ticker or a comparable peer. Governance drops: proposals, votes, treasury moves, parameter changes. Audit publications, including security incidents and the post-mortems that follow. Regulatory shifts touching your category (sanction lists, MiCA milestones, SEC moves, jurisdictional bans). Exchange events: listings, delistings, pair changes for your project or for rivals. Cashtag mentions: $TICKER-style references that appear in the broader X timeline. Brand mentions: anything that names your project, your protocol or your team. Each trigger has its own response template family and its own SLA.
The Review queue. Every reply a resident drafts lands in your team's queue with the full thread context, the trigger that fired and the resident's proposed copy. Approve or reject in two clicks, optionally edit before publish. Most teams approve over 90% of items once the brand card is calibrated, usually inside the first 7 to 10 days. The queue is the brand-safety mechanism that makes a thousand monthly touches operationally achievable without a 24/7 review desk on your side. By month two most teams move to a weekly batch review of edge cases only.
Sentiment versus share-of-voice. Sentiment is the polarity of category conversation about your project: positive, neutral, negative. Share-of-voice is the percentage of relevant category mentions that touch your project at all. These two move independently. A project with high SoV and bad sentiment is in a reputation problem, not a visibility problem. A project with neutral sentiment and low SoV is invisible, regardless of how the holders feel. We report both on the monthly review, with the trend line for each rolled forward over the last 90 days.
FUD inside a thread versus outside. Default behaviour on small noise (sub-200 followers, single complaint, no engagement) is silent monitoring. The standing pool does not chase. When a FUD incident crosses a threshold (engagement velocity, account size, replication into adjacent threads), the pool routes a response through Review. The response defaults to context-and-reframe inside the same thread the FUD surfaced in, rather than denial in a new post. New-post denials feed the cycle. In-thread reframes break it without creating a counter-trend.
Brand card. The single document that gates resident output. It defines tone, on-brand talking points, do-not-touch topics (price predictions, KOL beef, named competitors, governance opinions outside scope), how to handle FUD inside a thread, and the legal posture for your sub-vertical. Updated monthly with the team. The brand card is the difference between a pool that reads as a coordinated drop and a pool that reads as a normal slice of category conversation, and most onboarding time goes into calibrating it rather than into ramping touch volume.
The 90-day ramp. Standing pools do not switch on at full output. The first 15 days are calibration: brand card, resident matching against your sub-vertical, trigger map for your category. Days 16 to 45 are ramp: weekly touch volume climbs to plan while the team works through approve and reject cycles in Review. Days 46 to 75 are steady state at 300+ weekly touches across all seven triggers. Days 76 to 90 are a hold review where we hand over the FUD log, the SoV report and the next-quarter trigger map.
Standing pool versus launch pool. A launch pool runs hard for 14 to 21 days around a TGE date and then peels off. A standing pool runs at a lower weekly volume but never stops, watches seven trigger types continuously and is calibrated for narrative-holding instead of narrative-building. Most teams need both: the launch pool to build a category position at TGE, then the standing pool to hold it once the KOLs move on. The same calibrated residents transition between the two pools so the voice does not reset.
Onboarding the Review queue. The first week of a standing pool is dense on the team side, not on the pool side. Most teams spend 30 to 60 minutes a day approving or rejecting items, leaving notes on tone and adding edge cases to the brand card. By week two the approve / reject ratio stabilises in the high 80s percent and the calibration speed drops by half. By week four the team usually moves to a 15-minute morning sweep plus alert-only escalation, which is the cadence that holds through the steady state and beyond. The pool absorbs the load; the team owns the brand card.
Cashtag tuning. Cashtag mentions are the noisiest trigger by default. Generic $TICKER traffic on X includes spam bots, price screenshots, low-effort takes and farm replies, none of which are worth engaging. The standing pool filters cashtag traffic through three layers: account-age threshold, engagement velocity threshold, and a contextual classifier that checks whether the tweet is actually about your project or about a price chart that mentions your ticker. We tune these thresholds with the team weekly so the residents only land on cashtag traffic that is worth a human-readable reply, and the rest goes to silent monitoring.
What gets handed off at the 90-day mark. The end of the ramp is a hold review and a handoff. The team receives the calibrated brand card, the trigger map tuned to your category, the FUD incident log with every responded thread linked, the SoV report rolling forward over the previous 90 days, and a next-quarter map of expected category beats from the news, governance and audit calendars we track. From that point the standing pool runs without re-onboarding, the team takes a lighter Review cadence, and the brand card gets versioned monthly instead of weekly. Most teams keep the pool running indefinitely from there.
