Community Building

Community Building

Platform Comparison

| Platform | Best For | Pros | Cons (2026 reality) | |----------|----------|------|------| | Discord | Dev/gaming/crypto/creator communities | Free, real-time, rich (Onboarding, AutoMod, Forum/Media channels, threads, audio Stages) | Noisy, weak native search, public-by-default unless you gate; mod tooling needs setup | | Slack | B2B, professional/internal communities | Familiar, threaded, deep integrations, Canvases & Lists | Free plan limits you to recent message history & file access (~90 days), not the old "10,000-message" cap (as of Jun 2026; verify at slack.com/pricing); paid scales costly per active member | | Circle | Course/membership/paid communities | Clean UX, Spaces, native events, courses, paywall, mobile app | Paid SaaS, less real-time than Discord, you don't own the data layer | | GitHub Discussions | OSS projects | Free, lives next to the code, async, Q&A "answered" marking, categories | Dev-only audience; no real-time, no DMs, limited moderation/analytics | | Reddit (subreddit) | Public discovery & SEO | Indexed in Google + cited by LLMs, massive reach, anonymous low-friction posting | You don't own the audience or DMs; heavy mod load (spam/ban-evasion/brigading); API access is now rate-limited/paid for 3rd-party tools; Reddit's content policy + admin actions override your rules. See the reddit-community-engagement sibling skill for subreddit-specific playbooks. |

Other 2026 options: Skool and Mighty Networks (creator/cohort communities, built-in payments + gamification); Discourse (self-hostable forum, best long-tail SEO + LLM citation); Telegram/WhatsApp (huge in crypto/emerging markets, but near-zero structure, moderation, or analytics — treat as a broadcast/chat layer, not a home base). Pick ONE home base; don't run parallel communities you can't staff. Match the platform to where members already are and to your moderation capacity, not to feature lists.

Discord/Slack Channel Structure

📢 announcements        (read-only, major updates)
👋 introductions         (new members post here first)
💬 general               (main discussion)
❓ help / support        (Q&A, encourage helping each other)
💡 ideas / feedback      (product input, feature requests)
🎯 show-and-tell         (members share what they built)
🔧 off-topic             (human connection, non-work chat)
── Staff/Mod channels (private) ──
🛡️ mod-log               (actions taken)
📊 team-internal          (strategy, planning)

Start with fewer channels (5-6 max at launch). Empty channels signal a dead community — add only when an existing channel's conversation visibly splits into a recurring sub-topic, then create the channel and pin a seed post.

Adapt the structure to the platform — don't force the Discord layout everywhere:

• Discord Forum/Media channels for help, show-and-tell, ideas: each question/post becomes its own searchable thread with tags. Far better than a flat text channel for Q&A and showcases — threads don't get buried and can be marked solved. Use a Media channel for screenshot/build galleries.
• Discord Onboarding (Server Settings → Onboarding): set the server to Community, require members to pick interests/roles at join, and surface 3-5 "default channels" so newcomers land somewhere useful. Pair with Verification Level: Medium/High and rules screening to cut spam-bot raids.
• Slack: there are no read-only channels natively, so use a #announcements channel with posting restricted to admins, a pinned Canvas as the persistent "start here / rules" doc, and a List for tracking events or member projects. Threads are the unit of conversation — train members to reply in-thread, not in-channel.
• GitHub Discussions: model channels as Categories (Announcements [announcement format], Q&A [enables answer marking], Ideas [poll/upvote], Show and tell, General). There is no real-time chat or DM — set async expectations.
• Circle/Skool: model channels as Spaces; gate paid spaces behind the membership and keep one free public space for discovery.

Onboarding Flow

1. Welcome message → prefer an in-server onboarding post/ephemeral system message over a cold DM. Link to the intro channel + ONE quick action.
2. Intro prompt: Template in #introductions (below) — keep it to 3 short fields so it's frictionless.
3. Role assignment: React-roles or native Onboarding to self-select interests (drives relevant channel routing).
4. First value moment: Within 24 hours — answer their question, react to/welcome their intro, or invite to the next event. This single moment is the strongest retention lever you have.
5. Day 3 nudge: A public @mention in a relevant channel or a single opt-in check-in — not a recurring DM sequence.

Goal: New member → first meaningful interaction in 20% for engaged community | | Messages per active user | Total messages / active users | 3-10/week | | Response time | Time to first reply on questions | 40% | | New member activation | % of joiners who post within 7 days | >30% | | Lurker ratio | Read-only members / total | = CURRENTDATE - INTERVAL '30 days' AND event IN ('messagesent','reactionadded','eventattended') ) SELECT d, m, ROUND(100.0*d/NULLIF(m,0),1) AS stickiness_pct FROM dau, mau;

-- 7-day activation funnel for last week's joiners WITH joiners AS ( SELECT memberid, MIN(ts) AS joinedat FROM events WHERE event='memberjoined' AND ts >= CURRENTDATE - INTERVAL '14 days' AND ts >'latencysec')::numeric) FILTER (WHERE event='questionanswered') AS medianlatencysec FROM events WHERE ts >= CURRENT_DATE - INTERVAL '7 days';

**Cohort retention** (group joiners by ISO join-week, measure % active N weeks later):

```sql
WITH cohort AS (
  SELECT member_id, DATE_TRUNC('week', MIN(ts)) AS cohort_week
  FROM events WHERE event='member_joined' GROUP BY 1
),
activity AS (
  SELECT DISTINCT member_id, DATE_TRUNC('week', ts) AS active_week
  FROM events WHERE event IN ('message_sent','reaction_added')
)
SELECT c.cohort_week,
       FLOOR(EXTRACT(EPOCH FROM a.active_week - c.cohort_week)/604800)::int AS week_n,
       COUNT(DISTINCT a.member_id) AS active_members
FROM cohort c JOIN activity a USING (member_id)
GROUP BY 1,2 ORDER BY 1,2;

Read this as a triangle: each row is a cohort, each column is weeks-since-join, the cell is retained members. W4 retention is the number to obsess over — if W4 is below ~40% your activation/onboarding is leaking faster than you can fill it, and growth spend is wasted.

Alert thresholds (page yourself / mod team when a weekly metric crosses these):

| Signal | Warn | Critical | |--------|------|----------| | DAU/MAU stickiness | 4h | >24h | | Weekly net member growth | flat | negative 2 wks running | | % messages from top 1% of members | >50% (over-reliant) | >70% (community = 1 person) |

Engagement Tactics

Events

• Weekly office hours / AMA: Founder or expert answers questions live
• Monthly showcase: Members demo projects (builds connection + UGC)
• Challenges: 7-day or 30-day challenges with public accountability

Async Engagement

• Question of the week: Pinned prompt to spark discussion
• Wins thread: Weekly "share your win" — normalizes participation
• Polls: Quick opinion polls on relevant topics (low-effort engagement)

Recognition

• Shout out helpful members in announcements
• Leaderboard or point system (careful — can feel gamified/hollow)
• Exclusive roles for active contributors

Ambassador / Champion Program

Criteria to join:
- Active for 60+ days
- Consistently helpful (answers questions, welcomes newbies)
- Aligned with community values

Benefits:
- Private channel with team access
- Early access to features/roadmap
- Swag, event invites, reference/resume credit
- Direct influence on product direction

Responsibilities:
- Welcome 3+ new members/week
- Answer questions in support channels
- Flag issues/toxicity to mod team
- Attend monthly ambassador sync

Start with 3-5 champions. Scale to ~1 per 200 members.

Moderation & Trust-and-Safety

Moderation is a safety function, not just tidiness. A commercial community needs a written policy, an audit trail, an appeals path, and a severe-incident escalation plan before the first incident.

Rules / Code of Conduct (inlined template — post in #rules + pin)

{Community} Code of Conduct

You agree to:
1. Be respectful. No harassment, hate speech, slurs, threats, doxxing, or
   personal attacks. This applies to DMs initiated via this community too.
2. Keep it safe & legal. No sexual content, no content involving minors,
   no illegal goods/services, no malware, no sharing others' private info.
3. No spam / undisclosed promotion. Self-promo only in #show-and-tell or with
   mod permission. No unsolicited DMs to members. No referral/affiliate spam.
4. Stay on topic; use the right channel. Search before asking.
5. One account per person. No ban evasion via alts.

Enforcement: mods may remove content and warn, mute, or ban at their discretion.
Decisions can be appealed (see below). By participating you consent to mods
retaining records of enforcement actions for safety and appeals.

Report problems: @mention @mods or use {report channel / form / DM a named mod}.

Tailor #1-#5 to your audience, but keep illegal-content, minor-safety, and anti-doxxing clauses verbatim — they are non-negotiable.

Tiered enforcement (severity-based, not one-size-fits-all)

| Tier | Examples | Action | |------|----------|--------| | Minor | Off-topic, mild rule-bend, first-time low-effort spam | Friendly public nudge or quiet message; no formal record needed | | Standard | Repeated spam, incivility, ignoring mod guidance | Formal warning (logged) → timeout/mute 24h → 7-day ban → permanent ban | | Serious | Targeted harassment, hate speech, doxxing, scams, ban evasion | Immediate timeout + content removal, log evidence, then permanent ban after mod review (skip the warning ladder) | | Severe / emergency | Credible threats of violence, sexual content involving minors (CSAM), self-harm/suicide intent | See "Severe-incident escalation" — do NOT handle as routine moderation |

The Warning → mute → ban ladder applies to the Standard tier only. Serious abuse skips straight to removal; severe incidents leave the moderation track entirely.

Evidence capture & incident log

Before you delete anything, preserve it — deleted messages can't be un-deleted, and you'll need them for appeals or law-enforcement requests. Screenshot or export the message(s) with timestamp, user ID, and channel, then act. Log every Standard+ action in a private #mod-log (or a sheet) with this format:

[2026-06-07 14:32 UTC] action=warn  target_id=123…  mod=@alice
  rule=#3 spam  evidence=  note="3rd affiliate link today"  appealable=yes

Keep logs in a restricted channel, retain only as long as needed for safety/appeals, and don't expose them publicly.

Appeals

State the path in the CoC: a banned/muted user may appeal once via {a dedicated form, a single email/DM to a named mod, or an appeals server}. A second mod (not the one who issued the action) reviews the evidence and decides; record the outcome in the log. Don't argue enforcement in public channels — it invites pile-ons and rules-lawyering.

Moderator permissions, roles & safety

• Least privilege. Junior mods get timeout + message-delete only; ban/kick and role-management stay with senior mods/admins. On Discord, give mods only the specific permissions they need and require 2FA on the server (Server Settings → Safety Setup → require 2FA for moderation actions). Don't hand out Administrator.
• No solo moderation of serious cases. Two-person review for permanent bans and all appeals.
• Protect your mods. Mods see the worst content and get targeted. Provide a private mod channel for venting/decisions, rotate coverage to prevent burnout, never publish a mod's real identity, and let mods escalate harassment of themselves to admins. Mods are unpaid volunteers in most communities — set humane expectations and thank them.
• Conflicts of interest. A mod should not rule on a thread they're personally arguing in.

Severe-incident escalation (memorize this)

Some situations are emergencies, not moderation calls:

• Self-harm / suicide intent: Respond with care, share region-appropriate crisis resources (e.g. findahelpline.com for international lines; in the US the 988 Suicide & Crisis Lifeline), do not publicly broadcast the person's situation, and if there's imminent danger and you have identifying info, contact local emergency services. You are a community manager, not a clinician — connect, don't counsel.
• Sexual content involving minors (CSAM): Do not download, forward, or "investigate." Remove it, ban the account, preserve the platform-side evidence (report in-app so the platform retains it), and report to the platform and to NCMEC CyberTipline (report.cybertip.org) in the US or your national hotline. This is legally mandatory in most jurisdictions.
• Credible threats of violence: Preserve evidence, remove/ban, and report to the platform and to law enforcement if a specific person/place is threatened.
• Legal / law-enforcement requests & subpoenas: Don't improvise. Route to your company's legal contact; only produce data in response to a valid legal process, and tell members in your privacy notice what you log and when you'd disclose it.
• Data / privacy requests (GDPR/CCPA): If you store member data, honor deletion/export requests within the statutory window and document your retention. Don't dox, and don't publish member PII even in mod discussions.

Tooling (2026)

• Discord AutoMod: built-in keyword/regex filters, the managed Commonly Flagged Words and spam/mention-raid presets, and Raid Protection + join-gate Verification Levels and rules screening. Configure AutoMod to alert mods on borderline terms and auto-block slurs/links — start permissive and tune from false positives.
• Set Server → Safety Setup, enable Onboarding (cuts bot spam), and consider a verification bot for paid/gated communities.
• Slack: restrict who can post in announcement channels, use admin content controls, and on Enterprise plans use DLP/eDiscovery; otherwise moderate manually.
• Add an audit-log/mod-log bot so actions are traceable even if a mod goes rogue.
• Closely related: see the reddit-community-engagement sibling skill for subreddit AutoModerator configs and brigading defenses.

Scaling Stages

| Stage | Focus | Key Actions | |-------|-------|-------------| | 0→100 | Seed & personal touch | Invite individually, be in every conversation, DM everyone | | 100→1K | Habits & rituals | Weekly events, onboarding flow, first champions | | 1K→5K | Systems & delegation | Mod team, ambassador program, documented processes | | 5K→10K+ | Culture & self-sustaining | Members help members, UGC engine, sub-communities |

Critical insight: 0→100 is founder-led. You personally invite, personally welcome, personally engage. There's no shortcut.

Launch Sequence (first 30 & 90 days)

Don't open the doors to an empty room. Seed it first, then invite in waves.

Pre-launch (week -1): seed so it's never empty. Set up the 5-6 channels, AutoMod, Onboarding, rules, and the welcome flow. Hand-invite 10-20 "founding members" (people you know are your ICP), give them a Founding role, and ask them to post intros and a couple of questions/answers before anyone else arrives. A new visitor should see live, on-topic conversation in the first 30 seconds.

Days 1-30 — establish a heartbeat (founder budget: ~1-2 hrs/day):

• Daily: be present. Greet every new join, answer every question within hours, post one prompt/que

…

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: san-npm
• Source: san-npm/skills-ws
• License: MIT
• Homepage: https://skills-ws.vercel.app

Install and usage instructions live in the source repository linked above.