Brandium · Elemental — The Operating Manual

The content strategy

How a global branding agency wins attention in 2026: fewer, stronger, serialized posts; original data; employee voices; and content engineered to be cited by both people and AI engines. This page is the law behind the calendar — read it before running the kit.

← The kit The calendar The standard

01The thesis

Social in 2026 rewards community over virality: platforms rank on watch time, saves and comment velocity — not posting frequency. Trust beats production value, and accounts that feel like people outperform accounts that feel like corporations. Brandium's edge is structural: we're a lab, so our content is evidence. Every post is a finding, an experiment, a specimen or a testimony — formats that get saved, cited and argued with. The rotation isn't decoration; it's the algorithm strategy.

North-star metrics, in order: saves per post → qualified inbound (DMs/emails naming a real project) → profile-to-site clicks → follower growth. Reach is a diagnostic, never a goal. Review weekly, report monthly against the annual report.

02The operating clock

Slot	Time (GMT)	Format	Why this slot
Monday	08:00	Statement / Brand Table element	LinkedIn-first: text framework version posted natively, visual follows as a document post. B2B decision-makers read Monday mornings.
Wednesday	13:00	The Experiment (carousel) or Finding (data)	Midweek save-peak; carousels earn the highest save rates. 13:00 GMT catches London lunch, Lagos lunch, NYC morning.
Friday	17:00	Reel / Showcase	Video and case files ride weekend browse time. Reels: hook in 2s, cut every 2–3s, captions burned in, end on the tile.
Sunday	19:00	Specimen / Question	The quiet slot: grid rhythm and community debate when feeds are calm.
Stories	daily	Teaser → poll → ask-the-lab → countdown	Stories are the community room, not a billboard: one interactive sticker per day minimum.

Volume law: four feed posts a week, no more. If a fifth idea is truly good, it replaces the weakest of the four. Every skipped mediocre post raises the average — and the algorithm ranks the account, not the post.

03Series, not singles

Serialized content compounds: followers return for episodes, algorithms reward the pattern, and each series becomes a bingeable archive pinned in highlights. The kit runs four:

Series	Template	Cadence	Job
The Brand Table	BD-07	Weekly ×8	Edutainment: teaches what each discipline is, sells expertise without selling.
The Experiment	BD-13 + BD-11	Weekly	Original research in public — the save magnet and the GEO engine.
The Reaction Log	BD-05/06/08	Per launch	Proof arc: countdown → drop → testimony → diagnosis. Conversion.
Lab Personnel	BD-10	Monthly	Employee-generated content: human faces outperform corporate voice.

04Platform playbook

Platform	Value proposition	Native behavior
Instagram	The gallery — the grid is the portfolio's front door	Rotation law, 4:5 for reach, stories daily, highlights as archive. Alt text on every image.
LinkedIn	The frameworks — where budgets follow	Text-first posts + document (carousel) uploads. Founders and leads post under their own names; the company page amplifies. Comment on client and prospect posts daily.
TikTok / Reels / Shorts	The process — craft in motion	Same vertical cut everywhere: crits, stress-tests, print-shop texture. Unpolished beats produced.
X	The opinions — fast takes on brand news	Text threads of Experiments; quote-post rebrand news within 24h with a diagnosis, not a dunk.
Behance / Dribbble	The museum — full case files	Every reaction published as a complete case within a week of the feed drop. This kit's presentations upload directly.
YouTube	The library — long-form authority	Quarterly: one 8–12 min case documentary. Feeds GEO harder than any other single asset.
Pinterest	The reference — designers moodboard from here	Pin every specimen, poster and palette board. Slow, compounding search traffic.

05Search everywhere: SEO · AEO · GEO

In 2026 discovery is three-layered: classic search (SEO), answer boxes and voice (AEO), and AI engines that synthesize recommendations (GEO). The winning move is the same acquisition running through this whole kit: be the easiest source to cite. Research shows quotable expert lines lift a source's share of AI answers by ~41%, statistics by ~31%, and citations by ~28% — which is why every Brandium post leads with an answer, a number, or a named method.

Caption law — every post

Line one answers a real question, stated as the question ("What does a brand strategist actually do?"). AI engines and social search both index it.
One original stat or named method per post — from the lab's own log.
3–5 keyword hashtags, never 30. Keywords also live in the caption prose.
Alt text written like a caption for a blind colleague — descriptive, keyword-honest.
Bio and handle identical on every platform: same name, same one-liner, same link. Entity consistency is how machines learn who you are.

The GEO engine — monthly

Publish one piece of original research (The Experiment) on the website as a titled, dated, citable page — social points to it.
Ask ChatGPT, Perplexity, Gemini and Claude: "best branding agencies", "who is Brandium" — log answers monthly, track share-of-model.
Earn third-party citations: design-press writeups, award listings, podcast appearances. AI engines trust what others say about you more than what you say.
Every case file carries named authors — author authority is a ranking input.
Website (next phase) ships FAQ blocks, structured data, and question-shaped H2s. The social captions and site answers must agree word-for-word on positioning.

The one-line GEO test: if an AI engine quoted this post's first line verbatim, would it be a complete, correct answer with our name attached? If not, rewrite the first line.

06Community operations

Always

Reply to every serious comment within 24 hours; the lab's replies are content.
Best comment of the week becomes Friday's Field Note — credit the commenter.
DMs answered by a human with a name, never a bot voice.
Engage 15 min daily on client, prospect and peer posts — comments travel further than posts.
Ask-the-Lab runs weekly; unanswered questions seed next month's content.

Never

Engagement-bait ("tag 3 friends"), giveaways, or follow-loops.
Dunking on named small studios or unsolicited redesigns of real brands.
Posting a claim without its number, or a number without its method.
Trend-jacking that breaks the grid rhythm or the voice.
Publishing client work, quotes or numbers without written permission.

07The measurement ritual

Cadence	Ritual	Output
Weekly (Fri)	Saves, comments, inbound logged per post against the calendar	ONE LINE PER POST IN THE LOG
Monthly	Top/bottom 3 posts examined; GEO share-of-model check; one rule updated	FINDING POST (BD-04) IF NOTABLE
Quarterly	Calendar re-slotted from the data; one experiment graduated or killed	NEXT QUARTER'S CALENDAR
Annually	The audited year-wrap	THE ANNUAL REPORT (BD-23)