---

name: 38-community-seeding-global description: "Use when the user wants to show up in OTHER communities without astroturfing — Reddit, Facebook Groups, Discord, Slack, LinkedIn groups, Quora, and niche forums: a prioritized community list with the rules of each space, a named participant roster, post types written in full, reply scripts, a participation calendar, and red lines. Trigger on 'community marketing', 'Reddit marketing', 'seeding plan', 'post in Facebook groups', 'answer questions on Quora', 'we got banned from a group'. Not for — building your OWN paid or private community, see 28-community-building-global; monitoring mentions across the web, see 15-social-listening-global; briefing paid creators, see 06-ugc-egc-brief-global." metadata: version: 1.0.1 category: content license: MIT triggers:

• "community marketing"
• "Reddit marketing"
• "seeding plan"
• "post in groups"
• "community strategy"
• "forum marketing"
• "Discord communities"
• "Quora answers" output: "File .md — prioritized community list with rules, named participant roster, 4 post types written in full, response scripts, participation calendar, measurement table, compliance and red-line checklist" related:
• product-marketing-context-global
• 09-customer-insight-global
• 15-social-listening-global
• 35-brand-voice-global
• 37-social-caption-global
• 06-ugc-egc-brief-global
• 39-content-audit-global

---

Community Seeding (Global)

Community marketing done properly is participation, not placement. A brand earns the right to mention itself by being useful first, under a real name, in communities whose rules it has actually read. Manufactured engagement — sockpuppets, brand-written reviews, coordinated voting — is both against every major platform's rules and, in the US, against FTC rules. Getting caught is a bigger loss than never posting.

Full per-platform rules, disclosure mechanics, and cadence limits: read references/platform-rules.md.

Information gathering

Read .agents/product-marketing-context-global.md and the insight file (09-customer-insight-global) first. If information is missing, ask up to 4 questions:

1. What is being marketed, and which communities does the target audience actually spend time in? (subreddits, Discord servers, Slack groups, Facebook Groups, LinkedIn groups, Quora topics, niche forums)
2. What is the objective? Awareness / qualified leads / traffic / genuine social proof / product feedback
3. Who on the team will participate under their real name, and how much time per week can they give?
4. Has the brand posted in these communities before — what happened, and was anything removed or banned?

Principles

1. Real people, real names, one account each. Every participant is an actual member of the team using a single account with their own identity. No personas, no second accounts, no purchased or aged accounts.
2. Value first, at roughly 9:1. For every promotional or self-referential post, contribute nine genuinely useful ones. This ratio is Reddit's long-standing self-promotion norm and travels well to every other community.
3. Read the rules before the first post. Every subreddit, server, and group has its own self-promotion policy. Many require flair, a specific thread, or a minimum tenure. Rules override this plan.
4. Disclose the material connection every time. If the person works for, is paid by, or holds equity in the brand, that must be clear in the post itself — not buried in a bio. This is FTC 16 CFR Part 255.
5. Never write a review as if it came from a customer. Brand-authored or employee-authored reviews presented as independent are prohibited under the FTC fake-review rule (16 CFR Part 465) and carry civil penalties. Customer stories must come from the customer, with permission on file.
6. No vote manipulation or coordinated engagement. No asking colleagues to upvote, no brigading, no cross-account replies to your own posts. Platforms detect this reliably.
7. Warm up before you contribute anything promotional. Minimum two weeks of ordinary participation per community; four is better.
8. Measurable or it does not count. UTM where links are permitted, a "how did you hear about us" field, and a log of removals and bans.

Workflow — 4 phases

Phase 1 — Community research and rules mapping (1 week)

Selection criteria:

Criterion	Threshold
Size	Large enough to matter, small enough to be read — for niche B2B, 3,000 engaged members beats 300,000 dormant ones
Activity	Daily posts with real comment threads, not a wall of unanswered links
Rules	Self-promotion permitted in some form (dedicated thread, flair, tenure gate) — record the exact rule text
Moderation	Active mods. Well-moderated communities are safer to invest in than unmoderated ones
Audience fit	Members match the ICP from 09-customer-insight-global

Output: a table of communities ranked High / Medium / Low priority, with each community's self-promotion rule quoted, its link policy, and the content types that fit it.

Phase 2 — Participant setup and warm-up (2-4 weeks)

Each participant needs:

• A real profile: real name, real photo, a bio that states where they work. Affiliation visible before anyone has to ask.
• Genuine history: joined, commented, answered questions 3-5 times a week, without mentioning the brand.
• A consistent area of expertise that matches what they actually know — a support lead answers support questions, a founder answers business-model questions.
• No brand mention at all during warm-up.

A new account with no history is not usable for this work. If nobody on the team can commit the time, do not run the plan — hand the objective to 06-ugc-egc-brief-global and work with creators who are already members of those communities, under a disclosed paid arrangement.

Phase 3 — Participation content (ongoing)

Four post types, each written in full at 150-300 words, in a real person's voice rather than a brand voice:

Type	Purpose	Brand mention rule
1. Genuine question / research post	Learn from the community, surface demand language	No brand mention at all
2. Expertise share / teardown	Give real, specific value; earn recognition	Affiliation stated up front; product mentioned only if it is genuinely the answer
3. Answering an existing thread	Help someone who already asked	Disclose the connection in the same comment, in plain words
4. Customer story amplification	Real social proof	Written by the customer, posted by the customer or clearly attributed with permission. Never brand-authored

What performs in communities: open questions, specific how-we-did-it write-ups with numbers, short guides, teardowns, and honest post-mortems including what did not work.

Response handling for every post:

• Prepare honest answers to the 3-5 questions members will realistically ask (pricing, availability, whether it actually works, alternatives — including competitors).
• Reply within the first few hours. Early real engagement is what most feeds reward.
• If someone asks whether you work there, answer plainly. Never deflect.
• Recommend the product only when a member has described a problem it genuinely solves, and disclose in the same breath.

Phase 4 — Measurement

Metric	How to capture
Post reach and engagement	Native post stats per community; screenshot where no export exists
Inbound DMs and mentions	Ask "how did you hear about us" at first contact; log the community
Traffic	UTM per community, only where links are allowed by the rules
Leads and revenue	CRM source field set to the community name
Removals, warnings, bans	Log every one with the rule cited — this is the risk signal that decides whether to keep investing in that community
Sentiment	Track how the brand is spoken about unprompted, via 15-social-listening-global

Output structure

File name: community-seeding-[product]-[YYYYMMDD].md

# Community Seeding Plan — [Product] — [Period]

## I. Target communities
| Community | Platform | Size | Activity | Self-promotion rule (quoted) | Link policy | Priority | Content types that fit |
|-----------|----------|------|----------|------------------------------|-------------|----------|------------------------|
| | | | | | | High/Med/Low | |

## II. Participant roster
| Person | Role | Account | Expertise area | Warm-up status | Communities owned | Weekly time |
|--------|------|---------|----------------|----------------|-------------------|-------------|
| | | | | Not started / In progress / Ready | | |

## III. Four post types — written in full
### Type 1 — Genuine question (no brand mention)
[Full post, 150-300 words, real person's voice]

### Type 2 — Expertise share (affiliation stated up front)
[Full post — real insight, specific, not an advertisement]

### Type 3 — Answering an existing thread (connection disclosed inline)
[Full reply — answers the question first, mentions the product only if it is the honest answer]

### Type 4 — Customer story amplification (customer-authored, permission on file)
[The customer's own words + where permission is recorded]

## IV. Response scripts
| Likely question | Honest prepared answer | Who answers |
|-----------------|------------------------|-------------|
| | | |

## V. Participation calendar
| Date | Community | Post type | Participant | Summary | Target engagement |
|------|-----------|-----------|-------------|---------|-------------------|
| | | | | | |

## VI. Measurement
| Post | Community | Engagement | DMs | Leads | Removed / warned | Notes |
|------|-----------|------------|-----|-------|------------------|-------|
| | | | | | | |

## VII. Compliance log
| Post | Disclosure used | Rule checked | Approver |
|------|-----------------|--------------|----------|
| | | | |

Red lines — never do these

• Never create or use a persona, sockpuppet, or second account. One real person, one account.
• Never write a review, testimonial, or "just a happy customer" post on behalf of a customer.
• Never ask colleagues to upvote, like, or reply to make a post look organic. That is vote manipulation on Reddit and coordinated inauthentic behavior on Meta.
• Never omit the material-connection disclosure when a participant is an employee, contractor, affiliate, or equity holder.
• Never post an identical block of text across multiple communities. Spam filters and members both notice.
• Never drop a product link in the first post to a community, or in any community whose rules forbid links.
• Never cold-DM members who have not asked to be contacted.
• Never argue with moderators. If a post is removed, read the rule, adjust, and move on.

Risks to manage: post removal or account ban (lost warm-up time — prevent by following rules); public exposure as astroturfing (severe brand damage — prevent by never doing it, and by answering affiliation questions honestly); community rule changes (re-read rules monthly); FTC enforcement on undisclosed endorsements and fake reviews (prevent with the compliance log in section VII).

Related skills

• 09-customer-insight-global: identifies which communities the ICP actually uses and supplies their real language.
• 15-social-listening-global: tells you what the community is already discussing, so the angle fits an active conversation.
• 35-brand-voice-global: types 2 and 3 still have to match the brand's core message, even written in a personal register.
• 06-ugc-egc-brief-global: the paid alternative — working with creators and community members through a disclosed paid arrangement.
• 37-social-caption-global: owned-channel posts, which follow different rules from community posts.
• 39-content-audit-global: feed community results into the periodic audit to decide which communities to keep.

Quality checklist

• Every community's self-promotion rule is quoted verbatim, not summarized from memory.
• Every participant is a real, named person with one account and a stated affiliation.
• Warm-up of at least two weeks completed before any promotional participation.
• All 4 post types written in full — no outlines.
• Posts read as a person wrote them, not a brand page.
• No brand-authored customer reviews anywhere in the plan.
• Material connection disclosed in every post where one exists, in the post itself.
• Value-to-promotion ratio across the calendar is at least 9:1.
• Cadence respects each community's rules and does not repeat identical text.
• Measurement defined per post (UTM where permitted, source field, DM attribution).
• Removal and ban log in place.
• No red line is crossed anywhere in the plan.