Discover · Community health

Measure how your community actually feels

Two one-tap rating scales average into a belonging and value health score between 1.0 and 5.0, one you can trend over time, plus what members want more of and the one thing they would change.

Discover Communities Rating scalesCalculated fieldMulti-selectOpen feedback
What it showcases

Features most survey bots don't have

❤️

A health score you can trend

Two 1-to-5 rating scales, belonging and value, average into a single number per member. Average that across responses and track it release over release.

📊

A backlog, not just a number

A multi-select on what members want more of turns the health signal into a concrete list of things to actually do.

✍️

Catches what you missed

An open 'change one thing' box surfaces the issue you did not think to ask about.

🙋

Finds your helpers

A volunteer question surfaces the members who will help you fix what the survey uncovers.

Worked example This template comes with a a quarterly community pulse example so you can see exactly how it plays. Edit the questions and content to make it your own.

“How is everyone doing?” in chat gets you the loudest few voices and nothing you can track. This template measures how your community actually feels: members rate how connected they feel and how much they get out of being there, which becomes a health score you can trend, then pick what they want more of and name the one thing they would change. You leave with a trendable number and a to-do list.

Run it on a cadence, quarterly or after a big change, to catch disengagement before people quietly leave.

A number and a next step

Two 1-to-5 rating scales average into a health score you can trend over time. The activities multi-select tells you what to do about it. The open box catches the thing you did not think to ask, and the volunteer question surfaces the helpers who will fix it with you. Only the two ratings feed the formula, so tenure and the volunteer ask stay descriptive.

Because a rating answer is a number rather than a picked option, each scale also reports its own average and distribution. That is what lets you tell a belonging problem from a value problem instead of guessing which one moved.

The example is a quarterly community pulse. Swap the community name and it fits any server or org that wants to measure and improve how it feels to belong.

Variants

  • More dimensions: add rating blocks (safety, moderation quality, event satisfaction) and extend the health formula to divide by the new count.
  • Words instead of faces: swap a rating for an opinion scale when the wording of each point matters more than the tap target.
  • Anonymous pulse: switch to anonymous for franker answers on a sensitive community.
  • Segment tracking: slice the health score by tenure to see where engagement drops.
  • Volunteer funnel: route the willing members to a follow-up to onboard the helpers you surface.
How to run it

Live in a few minutes

  1. 1

    Clone and name your community

    Swap the community placeholder for your server or org, and tune the intro to your voice.

  2. 2

    Keep the two rating scales

    Belonging and value are the health-score inputs. Want a richer index? Add another rating block and extend the formula to divide by the new count.

  3. 3

    Run it on a cadence

    Publish quarterly, or after a big change, so members rate how it feels and pick what they want more of.

  4. 4

    Trend and act

    Average the health score, slice it by tenure, read the open feedback, and onboard the volunteers you surfaced.

FAQ

Common questions

How is the health score calculated?

It is the average of the two 1-to-5 rating scales: a calculated field works out (belonging + value) / 2, so every member lands somewhere between 1.0 and 5.0. Respondents do not see it; it is for your analysis. Average it across responses to get a community-health index to trend.

How do members answer the rating scales?

One tap on a face emoji, in Discord or on the web. Each scale runs from 1 to 5 with a word at each end, and because a rating answer is stored as a number, every scale reports its own average and a point-by-point distribution, not five unrelated option counts.

Do the other questions affect the score?

No. Only the two rating scales feed the formula. Tenure and the volunteer question are descriptive, so they stay out of the health score.

How often should I run it?

Quarterly, or after a major change. Trending the score over time is where the value is, so a consistent cadence matters more than any single run.

Can I make it anonymous?

Yes. Switch to anonymous for franker answers on a sensitive community. The score still aggregates.

Use this template

Clone it into your community, swap in your content, and share it. No building from scratch.