Attention, Not Volume: A Scorecard for Judging a Calm Token Community
Photo: Ycleymans, CC BY-SA 4.0, via Wikimedia Commons
Every community has a number attached to it, and the number is almost always a count of messages. It appears in dashboards, in update posts, in the quiet implication that a busier room is a healthier one. For a token whose entire proposition rests on composure rather than noise, that metric is close to useless — and occasionally actively misleading.
This is an argument, not a measurement standard. But it rests on a simple observation: message volume responds to incentives that have nothing to do with the health of a project. Incentive campaigns raise it. A price move raises it. A single contentious decision raises it more than a year of competent administration. What volume does not do is distinguish between a community that is being attended to and one that is being entertained.
Why counts flatter everyone
A count is attractive because it is unambiguous and cheap. It requires no judgement, and it can be produced by anyone with an export function. The trouble is that it conflates four different things: the number of people present, the number of things they have to say, the number of things worth saying, and the number of times someone has responded.
In a quiet month, all four fall. That tells you almost nothing. A community where governance questions are answered carefully within three days and one where they are met with emoji and a link to a roadmap may produce identical counts, and the second may produce more.
Five signals that survive a quiet week
The alternative is to measure attention rather than noise. Five signals are worth tracking, and all of them can be observed from outside the moderation team.
1. Time to first reply versus time to a real answer
Record two timestamps: when a substantive question is asked, and when a reply arrives that engages with it. The gap between the first reply and the first useful reply is often the more revealing figure. A community that answers in four minutes with “the team is aware” and in five days with a real explanation is functional but slow. One that answers in four minutes and never returns is worse than it looks.
2. Consistency of moderation
The test is whether the same rule is applied to the same behaviour regardless of who is behaving. Watch how a critical post from a large holder is handled compared with an identical post from a new account. Watch whether deleted messages come with a stated reason. Consistency is dull, and dull is the point: it is the thing that makes a calm room credible rather than merely quiet.
3. Closed loops on issues
Most support threads end without a stated outcome. A community that returns to say “this was a display bug, fixed on Thursday” or “this is intended behaviour, here is the documentation” is doing something rarer than it appears. Count the proportion of raised issues that receive a visible closure within thirty days. That single ratio often separates a project that is being run from one that is being narrated.
4. Substantive answers to governance questions
Governance is where volume is least trustworthy, because it is where the temptation to reply without answering is greatest. The signal is not whether questions get a response but whether the response contains information that could be checked: a date, a threshold, a named decision-maker, a document. A reply that restates the question in warmer language is not an answer, and a community that supplies many such replies is telling you something about how it will handle a contested vote.
5. Who is doing the answering
Track the number of distinct people who give substantive replies in a month. A community served by two individuals is resilient only until one of them takes a holiday. Breadth of knowledgeable participants is a genuine health signal, and it rises and falls more slowly than message counts — which makes it more informative.
A scorecard you can keep
A monthly record of five indicators, scored from zero to four, is enough to detect change without turning observation into a hobby.
| Signal | What to record | Score (0–4) |
|---|---|---|
| Response quality | Typical gap between question and substantive answer | |
| Moderation consistency | Same rule applied to comparable cases, reasons given | |
| Issue closure | Share of raised issues closed visibly within 30 days | |
| Governance substance | Share of governance questions answered with checkable detail | |
| Contributor breadth | Number of distinct substantive responders |
Sample in the same week each month, for the same length of time, in the same channels. Keep the raw notes alongside the scores — a score without an example is an opinion, and you will want to test your own judgement later.
A few practical rules keep the exercise honest. Do not score a month in which you were personally involved in a dispute; wait until the following month. Do not adjust a score because the price moved. And resist the urge to publish individual scores as a verdict on people who are, in most cases, volunteers.
What a falling score does and does not tell you
A decline across several months is a legitimate reason to look more closely. It may indicate a team stretched thin, a moderation process that has drifted, or a genuine change in how a project regards its holders. It may equally reflect a quiet period in which nobody had anything to ask.
What the scorecard cannot do is tell you whether to hold or sell, and it should never be presented as if it could. It measures one thing: whether the attention a community receives is proportionate to the claims it makes about itself. For a project whose public voice is deliberately calm, that is not a soft measure. Composure is easy to project and hard to sustain, and the places where it is sustained are almost never the places that light up a message counter.
Quiet, on its own, is neither good nor bad. Attentive quiet is a different thing entirely, and it is worth the effort to tell them apart.