meta-algorithm-guide

name	meta-algorithm-guide
description	Use when building or refreshing a platform-ranking reference from current evidence. Produces platform algorithm guide and pre-publication checklist; use `meta-posting-optimisation` when that neighbouring contract is the closer match.
metadata	{"portable":true,"compatible_with":["claude-code","codex"]}

Platform Algorithm Guide — Organic Reach Reference

Scope: This is an operational reference document, not a strategy document. It covers ranking signals, penalised behaviours, favoured formats, and posting benchmarks for the six primary platforms used in the East African market. Use the pre-publication checklist (Section 8) before every post. For platform-specific strategy, use the relevant platform-* skill.

---

Use When

• Use this skill for building or refreshing a platform-ranking reference from current evidence.
• Confirm that meta-posting-optimisation is not the closer route before proceeding.

Do Not Use When

• Use meta-posting-optimisation when its narrower output is requested.
• Do not publish, spend, change a live account, certify compliance, or invent missing client evidence.

Required Inputs

Artefact	Source/provider	Required?	If absent
Dated first-party platform guidance and account analytics	Client, approved systems, or dated platform exports	Yes	Stop the affected decision; request it or mark the field unknown and narrow the output.
Purpose, audience and approval boundary	Client brief or accountable owner	Yes	Return discovery questions; do not infer approval.

Outputs

Artefact	Consumer	Acceptance condition
Platform algorithm guide and pre-publication checklist	Client lead and next workflow owner	Every recommendation traces to an input, names an owner or next action, and marks assumptions and unassessed checks.

Evidence Produced

Evidence	Format	Acceptance condition
Decision and source register	Table in the deliverable	Each material claim records its source/date or is labelled unverified; missing evidence never becomes a pass.

Capability and permission boundary

Read and search access to the supplied artefacts are required; calculation or file-rendering capability is optional. This is read-only by default: inspect and report without changing source records, accounts, skills or campaigns. Editing the deliverable requires explicit authorisation; publishing, production mutation, destructive action, spend, and certification claims require separate explicit authority and evidence.

Degraded mode

If files, platform access, network, rendering, fonts, or calculation tools are unavailable, return the narrowest useful qualified platform algorithm guide and pre-publication checklist. Mark each blocked check not assessed, state the consequence, and provide the exact evidence needed to resume. Never convert an unavailable check into a pass.

Decision rules

Choice	Action	Failure or risk avoided
Dated first-party platform guidance and account analytics is current and attributable	Produce the full platform algorithm guide and pre-publication checklist and cite the evidence used.	Decisions based on stale or unrelated evidence.
A material input is missing or contradictory	Stop that decision, request clarification, or issue a labelled partial result.	Fabricated precision and false confidence.
The requested outcome belongs to meta-posting-optimisation	Route there and hand over the verified inputs already collected.	Neighbour collision and duplicated work.

Workflow

1. Confirm the requested decision, consumer, market, period and permission boundary; route to meta-posting-optimisation if its contract is closer.
2. Inventory the required inputs and their provenance. Stop any decision whose critical evidence is absent; recover by requesting it or recording a bounded assumption.
3. Apply the domain method in the core sections below, following the decision table whenever evidence conflicts or scope changes.
4. Verify calculations, dates, named platforms and claims against the supplied sources; label inference and uncertainty.
5. Produce the platform algorithm guide and pre-publication checklist, decision/source register and explicit next owner. Do not mutate live systems without separate authority.
6. Run the repository anti-slop ship gate. If a blocking factual, permission or evidence defect remains, fix it or withhold release.

Quality Standards

The output is client-specific, uses British English and the stated market/currency, distinguishes observed fact from inference, exposes gaps, and gives a checkable acceptance condition. Recommendations must be feasible within the confirmed budget, capacity and permissions.

Anti-Patterns

• Using an undated benchmark as the client's result. Fix: use account evidence or label the benchmark as a provisional comparator.
• Producing the platform algorithm guide and pre-publication checklist without dated first-party platform guidance and account analytics. Fix: stop the affected decision or issue a clearly bounded partial output.
• Treating missing access or data as a successful check. Fix: record not assessed, its risk and the recovery input.
• Absorbing meta-posting-optimisation into this workflow. Fix: route the neighbouring output and hand over verified inputs.
• Publishing, spending or editing a live account during planning or review. Fix: obtain separate explicit authority and retain action evidence.

Worked example

Given verified dated first-party platform guidance and account analytics, the skill produces a platform algorithm guide and pre-publication checklist with source dates and named assumptions. If that evidence cannot be accessed, it returns only the supported sections plus a recovery list; it does not fill gaps with East African defaults.

Read next

• meta-posting-optimisation for the neighbouring contract.
• anti-ai-slop during production.
• ai-slop-audit at the release checkpoint.

References

• Anti-AI slop production gate
• Follow the directly linked repository skills above and any domain references named in the core sections below. Verify current platform, price, legal and regulatory claims before use.

Required Input

Before generating this guide for a specific client or team, collect the following:

• Client business name — trading name as it appears on social profiles
• Industry — sector and sub-sector (e.g. retail: fashion boutique)
• Country / city — default: Uganda / Kampala
• Primary goal — organic reach growth / engagement rate improvement / content team training / reach recovery after algorithm drop
• Active platforms — list every platform the client currently posts on
• Current posting frequency per platform — how many times per week on each
• Primary content format — video / photo / text / mixed
• Team size — solo operator, small team (2–4), or agency setting

---

1. How Social Media Algorithms Work — Core Principles

Every major platform uses a machine-learning ranking system that predicts whether a given user will engage with a given piece of content. The algorithm is not a fixed rulebook — it is a probability engine trained on billions of interactions. However, consistent signals influence rankings across all platforms.

Universal ranking signals (all platforms):

Signal	What it measures
Completion rate	Did users watch/read to the end?
Saves / bookmarks	High-intent signal — user wants to return
Shares / re-shares	Distribution signal — user vouches for the content
Comments (meaningful)	Conversation signal — not emoji-only
Likes / reactions	Weakest signal but still counted
Profile click-throughs	Interest in the creator beyond the post
Dwell time	How long a user paused on the post

Principle: Platforms serve content that keeps users on the platform. Any format or behaviour that achieves this is rewarded. Any format or behaviour that drives users away is penalised.

---

2. Per-Platform Algorithm Ranking Signals

2.1 Facebook

Facebook uses a feed ranking model that weighs four primary factors: inventory (all eligible posts), signals (engagement data), predictions (likelihood a user engages), and relevance score.

Key ranking signals:

Signal	Weight	Notes
Meaningful interactions	High	Comments and shares outweigh likes
Video completion	High	Native video watched past 60 seconds strongly rewarded
Post type relevance	High	Facebook learns what format each user engages with
Recency	Medium	Fresh content boosted within first 2–3 hours
Profile relationship	High	Pages with consistent interaction history ranked higher
Link posts	Low	Outbound links reduce reach — Facebook penalises posts sending users off-platform

Facebook-specific notes:

• Facebook Groups posts reach further than Page posts in most niches — consider a community group as a supplementary owned channel.
• Stories do not appear in the main feed algorithm but reset daily and keep the page avatar at the front of the Stories bar.
• Reels on Facebook draw reach from both Facebook and Instagram inventory when cross-posted natively (not via third-party tools).

---

2.2 Instagram

Instagram's algorithm operates separately across Feed, Reels, Stories, and Explore. Each surface has its own ranking logic.

Key ranking signals:

Signal	Feed	Reels	Explore
Completion rate	High	Very High	High
Saves	Very High	High	High
Shares (DM shares)	High	Very High	Medium
Comments	High	Medium	High
Recency	Medium	Medium	Low
Account history with user	High	Medium	Low
Hashtag relevance	Low	Medium	High

The engagement window: Instagram's algorithm evaluates a post's early performance in the first 30 minutes after publication. High engagement in this window signals the algorithm to push the post to a wider audience. Low early engagement suppresses reach. Post when the target audience is most active (see Section 6).

Instagram-specific notes:

• Relevance, Recency, and Resonance (the "3 Rs") are Instagram's own published ranking criteria for Feed.
• Reels between 15–30 seconds outperform longer formats for completion rate in bandwidth-constrained markets.
• Carousels generate repeat views (users swipe back) — this dwell time signals quality.
• Do not post a Reel and then post a static image within 2 hours — Instagram favours spacing posts to prevent self-competition.

---

2.3 TikTok

TikTok uses the most transparent ranking model of the major platforms. Its primary signal is completion rate — the percentage of viewers who watch a video from beginning to end. Everything else is secondary.

Key ranking signals (in order of weight):

1. Completion rate — the single most important signal
2. Re-watches — users who replay the video
3. Shares — especially off-platform shares (WhatsApp, SMS) which signal viral pull
4. Comments — open-ended videos that invite response perform best
5. Likes — weakest signal but still counted
6. Follows from the video — indicates the content converted a non-follower

TikTok distribution model: TikTok shows every video to a small test cohort (typically 300–500 users). If completion rate exceeds threshold, it is shown to progressively larger cohorts. This means a new account with zero followers can achieve massive reach on a single video — follower count is not a gating factor.

TikTok-specific notes:

• Keep videos under 45 seconds for EA audiences — data costs mean users rarely watch long-form TikTok on mobile data.
• Use on-screen captions. Many users watch without sound in public spaces (matatus, offices, queues).
• Hook in the first 2 seconds. The algorithm measures whether users skip past the first frame.
• Trending audio accelerates distribution — use audio from the Creative Library or trending local sounds.
• Posting at low-data-cost times (early morning before commute, evening after 8 PM EAT when users are on Wi-Fi) improves completion rates.

---

2.4 YouTube

YouTube's algorithm optimises for watch time and session time — how long a viewer watches a video and how many subsequent videos they watch in the same session.

Key ranking signals:

Signal	Notes
Click-through rate (CTR)	Percentage of impressions that resulted in a click — thumbnail and title are critical
Average view duration	Absolute minutes watched, not percentage
Watch percentage	Percentage of the video watched
Likes and dislikes	Sentiment signal
Comments	Engagement depth signal
Saves to playlist	Strong long-term signal
Subscribers from video	Conversion signal

Series playlists as algorithm signals: Organising videos into series playlists signals topical authority to YouTube's algorithm. When a viewer finishes one video in a playlist, YouTube auto-queues the next, increasing session time and directly rewarding the channel. Create a playlist for every content series, campaign, or topic cluster.

YouTube-specific notes:

• Thumbnails drive CTR — test two thumbnail variants where possible.
• The first 30 seconds determine whether a viewer stays. Front-load the value.
• Low-bandwidth EA audiences often watch YouTube on mobile data — keep videos under 10 minutes for educational content; longer only if the audience is Wi-Fi-dominant (corporate, campus).
• Shorts (under 60 seconds) are ranked separately and drive channel discovery — use them to attract new subscribers who then watch long-form content.
• Consistent upload schedule (same day and time each week) trains YouTube's recommendation engine to push content at predicted release times.

---

2.5 LinkedIn

LinkedIn uses three layered graphs to rank content: the identity graph (who you are), the interest graph (what topics you follow), and the knowledge graph (content quality signals). A post is shown first to your direct connections, then to second-degree connections if it performs well.

Key ranking signals:

Signal	Notes
Early engagement velocity	Engagement in the first 60 minutes is critical
Comment depth	Replies-to-comments (threads) score higher than top-level comments
Dwell time	LinkedIn tracks how long users pause on a post
Content format	Native posts > articles > documents > links > polls
Hashtag relevance	Use 3–5 relevant hashtags — more than 5 suppresses reach
Creator mode	Accounts with Creator Mode active get broader distribution
Native documents (PDFs)	Carousel-style document posts generate high dwell time

LinkedIn-specific notes:

• LinkedIn penalises posts with outbound links in the main text. Move the link to the first comment and reference it in the post ("Link in the first comment").
• Native video autoplay on LinkedIn is silent — include captions.
• LinkedIn's algorithm favours personal profiles over Company Pages for organic reach. Encourage team members to share or comment — employee advocacy multiplies reach without ad spend.
• The professional context of LinkedIn means posts that teach a skill, share a genuine lesson, or offer a professional insight outperform promotional posts.

---

2.6 X / Twitter

X uses a ranked feed (For You) and a chronological feed (Following). The For You algorithm is the primary distribution engine.

Key ranking signals:

Signal	Notes
Engagement velocity	The first 2–3 hours after posting are the critical window
Replies	Conversations started by the post signal quality
Reposts (RTs)	Distribution signal
Likes	Standard engagement signal
Bookmarks	High-intent saves signal
Profile authority	Verified accounts and accounts with high follower-to-engagement ratios ranked higher
Media inclusion	Posts with images or short video outperform text-only

X-specific notes:

• Threads perform well for educational or narrative content — post the thread as a sequence and reply to the first tweet to extend reach.
• Do not include links in the first tweet of a thread — X suppresses reach for posts with outbound links. Add the link in a reply.
• The For You feed now surfaces content from accounts the user does not follow — this is the primary discovery mechanism for new audience growth.
• X is the dominant platform for journalists, politicians, and public discourse in Uganda — useful for PR, reputation management, and media relations rather than direct commerce.

---

3. Engagement Window Reference

Platform	Critical engagement window	What happens after
Instagram	First 30 minutes	Algorithm decides wider distribution based on early signal
Facebook	First 2–3 hours	Feed ranking locks in based on initial engagement rate
LinkedIn	First 60 minutes	First-degree network determines whether second-degree sees it
X / Twitter	First 2–3 hours	For You algorithm scores and distributes based on velocity
TikTok	First cohort (300–500 views)	Completion rate in first cohort determines next cohort size
YouTube	First 48 hours	CTR and watch time in this period shape long-term search ranking

Implication: Post when the target audience is online. Do not post and immediately go offline — respond to early comments to signal activity to the algorithm.

---

4. Favoured Content Formats by Platform

Platform	Most favoured	Second	Least favoured
Facebook	Native video / Reels	Stories, Carousels	Link posts
Instagram	Reels	Carousels	Static single image
TikTok	Short vertical video	Stitches / Duets	Static images (limited)
YouTube	Long-form video	Shorts	Community posts
LinkedIn	Native text posts, Documents	Native video	External links
X / Twitter	Threads with media	Short posts with image/video	Link-only posts

Universal rule: Native content (uploaded directly to the platform) outperforms content linked from external sources on every platform. Never post a YouTube link on Facebook if the goal is reach — upload the video natively.

---

5. Posting Frequency Benchmarks

Platform	Minimum	Recommended	Maximum before quality drops
Facebook	3× per week	5× per week	1–2× per day
Instagram	3× per week	4–5× per week (inc. Stories daily)	2× per day
TikTok	3× per week	5–7× per week	3× per day
YouTube	1× per week	2× per week	1× per day
LinkedIn	2× per week	3–4× per week	1× per day
X / Twitter	3× per week	5× per week	Multiple per day (threads count as one)

Note: Consistency beats volume. Posting 3 times per week every week outperforms posting 7 times in one week and going silent for two weeks.

---

6. EA-Specific Considerations

6.1 Peak Activity Times (East Africa Time — EAT, UTC+3)

Platform	Weekday peaks	Weekend peaks
Facebook	07:00–09:00, 12:00–13:00, 19:00–21:00	10:00–12:00, 19:00–21:00
Instagram	07:00–08:30, 12:00–13:30, 20:00–22:00	11:00–13:00, 20:00–22:00
TikTok	06:30–08:00, 12:30–14:00, 20:00–23:00	10:00–14:00, 19:00–23:00
YouTube	19:00–23:00 (Wi-Fi hours)	10:00–23:00
LinkedIn	07:00–09:00, 12:00–13:00, 17:00–18:30	Low activity
X / Twitter	07:00–09:00, 12:00–14:00, 20:00–22:00	10:00–12:00

6.2 Data Costs and Video Consumption

Mobile data costs in Uganda remain a significant barrier to video consumption. 1 GB of data costs approximately UGX 3,000–5,000 on most networks (2026 rates).

Implications for content:

• Keep TikTok and Instagram Reels under 45 seconds — longer videos consume more data and are abandoned before completion, hurting the completion rate signal.
• For YouTube, keep educational content under 10 minutes unless the target audience is Wi-Fi-dominant (universities, corporates, Kampala CBD offices).
• Add on-screen captions to all video — users frequently watch without sound to avoid data drain.
• Compress video files before upload — a 720p video performs equivalently to 1080p on mobile screens and loads faster.
• Schedule video posts for evening hours (after 19:00 EAT) when users are more likely to be on Wi-Fi at home.

6.3 WhatsApp as a Near-Zero-Cost Distribution Channel

WhatsApp operates on negligible data relative to video platforms. For many EA businesses, WhatsApp is the primary customer communication and content distribution channel.

• Share links to published content (blog posts, YouTube videos, Facebook posts) via WhatsApp Business broadcast lists.
• WhatsApp Status (24-hour Stories equivalent) reaches opted-in contacts at near-zero data cost.
• WhatsApp does not have an algorithm — delivery is chronological and universal to the contact list. Use this to guarantee reach to the highest-value contacts (existing customers, warm leads).
• WhatsApp engagement (replies, reactions) does not count towards social platform algorithm signals. Use WhatsApp to drive traffic to the platform post, not as a replacement for it.

---

7. Algorithm-Penalised Behaviours

Avoid the following. Each behaviour either directly suppresses reach or trains the algorithm to associate the account with low-quality content.

Behaviour	Platforms affected	Why it is penalised
Engagement bait ("Like if you agree", "Tag 3 friends to win")	Facebook, Instagram	Facebook explicitly demotes engagement bait posts
Outbound links in post body	Facebook, LinkedIn, X	Sends users off-platform — algorithms suppress
Cross-posting identical content	All	Algorithms detect duplicate content; native upload always preferred
Clickbait headlines	YouTube, X, LinkedIn	CTR drops when viewer does not stay — algorithm penalises mismatch
Buying followers or likes	All	Inflates metrics without engagement; signals low-quality content to algorithm
Posting in bursts then going silent	All	Inconsistency reduces algorithmic trust and scheduling priority
Hashtag stuffing (30 tags on every post)	Instagram, LinkedIn	Instagram reduced hashtag weight significantly in 2024; LinkedIn caps effective reach above 5 tags
Reposting competitors' viral content without adding value	TikTok, X	Flagged as low-effort duplication
Ignoring comments	All	Non-response to comments signals low engagement depth; algorithm deprioritises

---

8. Pre-Publication Checklist

Use this checklist before every post goes live. Adapt to platform as indicated.

---

BEFORE YOU POST — QUICK REFERENCE CHECKLIST

Content fundamentals

• Does the post serve the audience first (inform, entertain, inspire) before it serves the business?
• Is the hook clear in the first line / first 2 seconds of video?
• Is the call to action specific and single (one CTA per post)?
• Is the copy free from spelling errors and in British English?

Format and upload

• Is the content uploaded natively to the platform (not linked from another platform)?
• Is the video under the recommended length for this platform and audience context?
• Are captions / subtitles included if the post contains video?
• Is the image or video resolution correct for the platform and placement?

Algorithm signals

• Is there a reason for the audience to comment, save, or share (not engagement bait)?
• Are outbound links in the first comment, not the post body? (LinkedIn, Facebook, X)
• Are hashtags relevant, specific, and within the platform limit (3–5 for LinkedIn, up to 10 for Instagram)?
• Is the posting time within the platform's peak activity window for EAT? (see Section 6.1)

Penalised behaviours — confirm none are present

• No engagement bait phrases ("Tag a friend", "Like if you agree", "Comment YES")
• No direct outbound link in the post body (Facebook, LinkedIn, X)
• Content is original or properly adapted — not copy-pasted from another platform
• Headline / caption accurately represents the content (no clickbait)

Post-publication (first 30–60 minutes)

• Monitor and respond to all comments within 30 minutes of posting
• Share to WhatsApp Business broadcast list if content is relevant to existing customers
• Share to relevant Facebook Groups if applicable (with group permission)
• Check early analytics (reach, views in first 30 min) to gauge early signal

---

Quality Criteria

Output produced by this skill meets the standard when:

1. Platform specificity — every platform section identifies its unique primary signal (completion rate for TikTok, saves for Instagram, watch time for YouTube) rather than giving generic advice applicable to all platforms.
2. EA market grounding — data cost constraints, WhatsApp distribution logic, and EAT peak times are integrated into the operational guidance, not appended as an afterthought.
3. Checklist usability — the pre-publication checklist in Section 8 is formatted so a content producer can print it or pin it to their screen and use it without reading the full document.
4. Penalised behaviours are specific — each penalised behaviour names the platforms affected and explains the mechanism of penalisation, not just that it "hurts reach".
5. Engagement windows are actionable — the timing guidance in Sections 3 and 6.1 gives specific EAT times, not vague recommendations like "post when your audience is active".
6. Formats table is decision-ready — a content producer can look at Section 4 and immediately know which format to prioritise for their platform without further research.
7. No strategy bleed — this document does not recommend content pillars, audience personas, or brand voice. Those belong in the relevant strategy skills. This skill stays operational.
8. British English throughout — organisation, behaviour, favoured, recognise, analyse — no American spellings anywhere in the document.