Official-source social-media growth playbooks, October 2026

Load this for platform-native production, hooks, SEO/discovery, timing, video and organic engagement. This is a curated operational synthesis, not a secret formula for proprietary recommendation engines. Every network can change ranking, specs and API permissions. Consult official sources at campaign time and mark non-official tips as hypotheses to test. Do not copy others’ exact skill instructions or automate prohibited participation.

Common rules for every network

• Source before story: show a real release, runnable example, measured result, first-party photo or documented limitation. A strong opening must be repaid immediately with proof.
• One pillar asset, multiple genuinely new treatments: produce an original demonstration or article, extract evidence-supported moments, then reframe for each audience. Avoid verbatim duplicate floods or third-party clip reuploads.
• Audience-native design: first seconds/line express the viewer’s problem and concrete payoff; demonstrate rather than advertise; no false numeric superlatives, fake scarcity, purchased social proof or invented endorsements.
• Search discovery: use words your audience searches for in truthful title, visible caption, spoken explanation and accessible description; useful keywords outperform irrelevant hashtag stuffing. Validate current hashtag, length, music and licensing requirements per network.
• Accessible production: accurate subtitles, meaningful alt text, legible screenshots, clear audio, contrast and understandable context without sound. Show rights/licensing for screenshots, footage and music. Never label a storyboard as a rendered asset.
• Consent + reputation: manual review of claim, exact identity, account, moderation norms and URL before publishing; no auto-liking, network-of-accounts amplification, mass unsolicited replies or subverted rate limits.
• Experiment not superstition: varying hook, first-frame proof, duration, format, thumbnail or time are hypotheses, not guarantees. Predefine metric, observation window and stop rule, and beware selection bias.

LinkedIn: useful professional insight, not empty thought leadership

Documented: LinkedIn’s Page guidance recommends complete company profiles, regular posts, custom visuals, video, thoughtful discussions and analytics. It reports correlational Page performance examples (e.g., images more comments; video more engagement); those are not guarantees for an individual author.

Do:

1. Name the exact professional problem in the opening; show one design decision, diagram, concise demo, validated benchmark or surprising true failure.
2. Explain trade-offs and a replicable lesson. State what evidence supports, and disclose project affiliation. Avoid formulaic inspirational filler.
3. Choose a plain-text lesson, real screenshots, short native video, or an authored PDF/document when a system architecture needs sequence.
4. Add one contextual CTA (reproduce the example, discuss a limitation, try documented open source); respond to legitimate technical questions.
5. Test personal versus Page accounts only with separately authorized permissions; compare meaningful conversations and technical clicks, not just impressions.

Avoid: blanket statements that “external links are punished” or “commenting in the first hour guarantees growth”. Test actual outcomes; do not move links into comments solely because an unaudited skill recommends it.

YouTube long-form + Shorts: discoverable, retained demonstrations

Documented: YouTube describes recommendations as personalized using clicks, watch time, viewer satisfaction, shares and other signals, not one fixed formula: recommendations. Its creator metrics guide highlights CTR, retention and traffic sources.

1. Start an engineering demo with the observable outcome (screen recording), then specify the problem and system; do not spend the opening on logos or generic intros.
2. For searchable long-form, use a title that states the actual task and tool, a thumbnail that makes an accurate promise, descriptive chapters where appropriate, technical prerequisites and links to reproducible source.
3. For Shorts, use one tangible idea / result, visually understandable in the first moments, with correct subtitles and an intelligible standalone explanation.
4. Inspect audience-retention dips, traffic source (Search / Suggested / Browse / External), engaged views and useful downstream visits. YouTube changed its basic views counting from August 24, 2026; preserve definition and period: Analytics help.
5. For eligible public long-form videos, YouTube Studio Test & compare can test up to three thumbnails concurrently and evaluates watch-time share, not CTR alone. Do not present sequential title edits as randomized A/B tests. Not available for Shorts in cited eligibility guidance.

Instagram Reels / carousels: original visual teaching

Documented: Meta’s Instagram creator Best Practices offers personalized creation, engagement, reach and guideline advice in Professional Dashboard; Meta’s Jan 2026 update says 75% of US Instagram recommendations in Q4 2025 came from original posts. This is platform-wide, not an individual account target.

1. Film/produce original screen demos, comparison visuals or visual explainers. Repurpose a topic, not lightly edited copies of a third party’s media.
2. Open with a visible result and concise on-screen problem; make a self-explanatory Reel or carousel (question -> challenge -> real proof -> applicable takeaway).
3. Caption accurately, use legible overlays, meaningful relevant language/keywords and proper credits. Select an honest profile/website CTA rather than generic comment bait.
4. Use the account’s professional dashboard and native insights to inspect reach, shares, saves and retention when actually available.
5. Respect original-content and recommendation-eligibility policies. Meta also describes Facebook original-content prioritization.

TikTok: interest-led discovery, not magic hashtag combinations

Documented: TikTok’s recommender explanation identifies interactions (including watch-through versus skip), content information and user factors; relative weights change. Watch behavior often matters more than fixed posting-time recipes.

1. Translate the technical achievement into an immediate visual before/after or one understandable lesson for the intended audience.
2. Make the first visual frame meaningful without knowing the creator or product; use native captions and an accurate searchable topic description.
3. Observe real retention, shares, qualified replies and profile actions; isolate one variable in the next video.
4. Never scrape copyrighted trending audio or assume popularity of an unrelated trend produces relevant technical users.
5. TikTok’s Content Posting API can limit unaudited apps to private visibility; for ReachWeaver prepare manual creator-ready media/script until app approval and permission have actually been verified.

Reddit: community-first help, not a scheduled promotion feed

Documented: Reddit rules require authentic participation, compliance with each community’s rules and avoidance of spam/content manipulation.

1. Find a relevant question that actually needs an answer; read the subreddit’s exact self-promotion and flair policy before considering a submission.
2. Lead with reproducible diagnosis, transparent limitations and a standalone useful answer. Disclose when a referenced tool is yours.
3. Use a link only when permitted and directly relevant; invite technical criticism rather than ask for votes or mass share to communities.
4. Let the human author review and publish; respond as an actual participant. No voting automation, bulk Reddit posting, unsolicited DMs or fabricated testimonials.
5. Evaluate genuinely constructive replies and qualified adopters, not karma loops.

Hacker News: manual original writing and occasional real launch

Documented, strict: HN guidelines say do not automate posting or submit generated / AI-edited text. The Show HN rules require a project people can actually try, with real technical substance and no vote solicitation. Show HN access notice currently warns of temporary restrictions for unfamiliar submitters.

ReachWeaver permitted action: do private research, verify prerequisites, prepare factual evidence inventory/checklist and remind the owner of guidelines. The human must independently compose the HN title and submission (without AI-generated or AI-edited text) and manually post if eligible. Do not export generated copy labeled “HN ready to paste.” Major launch only, not every minor release.

X, Threads, Bluesky, Mastodon: authentic network conversations

• X: concise source-backed stand-alone post or genuinely useful technical thread; publish through approved API or manual owner action. X Rules prohibit artificial amplification, account manipulation and credential misuse. Do not assume an exact best time or limit without live validation.
• Threads: conversational topic, relevant community and constructive dialogue; try original thought plus context, not corporate press copy. Meta reported 500 million MAU in June 2026; this does not promise an individual’s impressions.
• Bluesky: technical observation with source, deliberate accessibility and native discussion; use official post/facets API when authorized. Text-only adapter in ReachWeaver remains live-account unverified.
• Mastodon: respect per-instance policy and community norms, add useful context and alt text; use official status API only when explicitly authorized. No synthetic engagement or mass mentions.

Facebook, Pinterest, Snapchat and opt-in groups

• Facebook: test original Reels or relevant moderated Page/Group lessons where niche users already participate; Meta says unoriginal reuploads may be deprioritized: official March 2026. Not a default first channel for deep enterprise platform engineering.
• Pinterest: for visual how-to diagrams, workflow visuals and keyword-led evergreen discovery. Only relevant if honest vertical graphics solve an audience query; avoid repeated identical pin floods.
• Snapchat: potentially large top-of-funnel audience but often high production cost and less direct conversion for technical B2B topics; test only with a deliberate user persona and original mobile-native video.
• Telegram, Discord, WhatsApp, newsletters: send valuable opt-in release notes, respond to actual questions and build retention. Group member counts are not cold discovery; never mass-invite or auto-DM.

Community and conversion adjuncts (not “mass social” networks)

• DEV Community: original, tested how-to and source code links; tag-specific rules permit some self-promotion but ask for restraint.
• GitHub: make README, releases, examples, issue templates and onboarding actually useful; social discovery cannot rescue a broken first-use experience.
• Product Hunt: launch only when the product fits the audience and current guidelines; human-led engagement, not manufactured comments.
• Search/SEO: the owned documentation site and discoverable tutorial are durable assets; route social visits through transparent UTM links and first-party consented analytics, using MSKazemi/seo for deeper SEO where available.

Reusable “attention test card” for Claude

For each platform variant output:

• Proof: exact original source, claim, timestamp, uncertainty
• Audience/job: named person who benefits, query/problem in their language
• Format: demonstrable asset, truthful 1-line/first-frame payoff, native pacing, rights and accessibility
• Discovery: literal natural search terms, platform/community rules and real approved account
• Action: one useful CTA and correctly tagged destination if possible
• Evidence of success: native metric with denominator/definition, qualified engagement, 24h/7d observation
• Status: factual documented practice vs creator hypothesis, direct publishing vs manual only
• Iteration: single next change if actual results justify it

Anti-patterns: blanket “best 9 AM Tuesday” rules; compulsory 3–5 hashtags; link-in-first-comment myths; fake urgency; generic AI inspirational posts; too many posts; auto-replies; mass subreddit cross-posting; assumed API permission; engagement pods; copying competitor media; marketing claims without reproducible evidence.