SPINALE Content Engine v4
Social Media Execution System — February 2026
v3 had the right data. v4 has operator-level advice from someone who's done it.
Built on v3 analysis + Numan (Krea AI) strategy session — 2/19/26 @ 4:15pm EST
Numan: 136k IG / 5M views at Krea. Grew GigaChat IG from 12k. Former #1 artist on Rarible. Viral AI content specialist.
Numan's Audit — What He Said We Should Do
These are direct recommendations from the call. Not paraphrased opinion — operator notes from someone who grew Krea from 89k to 136k followers and 5M views in 6 months.
🔴 STOP IMMEDIATELY
- The three-across grid posts. "Instagram punishes you each time. You're posting three times in a row." Also: it looks messy. Someone scrolling sees half a bracelet and doesn't understand what they're looking at. They don't tap to find out.
- Boosting within Instagram. "It's basically flushing money down the toilet. It doesn't do anything unless a post has already gone viral." Use Facebook Ads Manager on desktop instead — that's how real ads run.
- Spreading engagement across multi-image grid sets. One image = concentrated likes. Three images at once = diluted engagement. Algorithm reads diluted = less reach.
- Pure AI content for conversion posts. For people spending $2k+, "they wanna know what they're buying. Crazy concepts of it turned into a car — it looks great, but that's not gonna convert to sales."
🟡 FIX NOW
- Archive the three-grid posts. Don't delete — archive them. Start posting single images, and quietly archive the old grid in the background simultaneously.
- Reel cover images. Numan noticed two recent reels had the same cover thumbnail. "People don't think it's the same thing if it's about to watch it." Each reel needs a distinct, eye-catching cover. Either set one manually or pick the most visually different frame.
- Establish one person who owns socials. "Have one person own it." Split ownership = inconsistency. Assign the account.
- Post 4:5 ratio for grid posts. "It takes up more screen space on your phone, meaning more eyes on it." Square is fine. 4:5 is better. 9:16 only for Reels.
🟢 DO MORE OF
- Founder story content. "I would take a camera to the jeweler for the day, get a shitload of content, break it up into 10–15 second reels, keep pushing them out." Film the process. People spending $2k+ want to know what they're buying. One shoot = one month of content.
- Posting consistency above all else. "Consistency is the main reason why nothing ever works for people. They just don't commit." Three times a week minimum — Monday, Wednesday, and Sunday/Friday evenings. Lock it in, never miss it.
- Mix carousels + reels. Both. Not just reels. Carousels for the collector/buyer, reels for reach.
- Organic content mixed with AI. Numan's personal page: "intentionally shitty — taken on an iPhone against a white wall. That's purposely done. It's a different mode of content." Not every post needs to be polished. Some of the most viral content is lo-fi by design.
- Use the Krea LoRa trainer. 50 images of ABALLE's visual style → 10-minute custom model → every prompt stays on-brand without lengthy prompting. Numan confirmed this is production-ready.
- Nano Banana Pro + Cling as the AI toolchain. Numan's exact words: "Nano Banana Pro is the best image model out there at the moment, and will be for a long time." Cling for video — "keeps the first frame consistent, lighting and shadows stay stable." These are his only two tools for client work.
- Use the professional dashboard. Check what times your audience is most active. Post then. Every time.
🔵 STRATEGIC MOVES
- Giveaway as marketing budget reallocation. "$2k bracelet giveaway vs. $1k+ photo shoot with uncertain results — look at it as your marketing budget spent differently." High engagement, high visibility. One caution: don't put dollar figures in the caption — may suppress reach.
- Celebrity approach: personal connections, not paid placements. "Get them in. Let's take a few pictures. Personal favors." The story doesn't need to be 'they bought it' — just photos with people who have influence. Numan has a direct line to Jay Balvin (one of the biggest Colombian artists). g.money has proximity to Raúl Alejandro and Bad Bunny. Work those connections.
- Entry-level products as social hooks. Rings at $500, pendants, belts. "The aspirational buyer — the 24-year-old who can't afford the watch bracelet — can wear the ring." Numan used rubber bracelets as an example ($1-2 to produce, sell at $150). g.money's push back is valid — brand preservation matters. The existing $500 ring plan is the right call. Rubber bracelets are not ABALLE.
- Rinse and repeat what hits. "See what hits on Instagram and just rinse that method until it's time to switch up the process." When a format works — same format, different content, again and again.
What Changed From v3 (And Why)
| v3 Assumption | What Numan Added | v4 Fix |
|---|
| Grid posts are a gallery — the three-across format shows a curated world | "It looks messy. If it pops up in my feed as half a bracelet, I don't tap." Also: Instagram punishes rapid consecutive posts with suppressed reach. | Single post strategy only. 4:5 ratio. Archive the three-grid posts while posting new content. |
| Reels are the primary engine — double down hard | Confirmed. But: mix in carousels too. And intentional lo-fi organic content alongside polished AI visuals. | Maintain Reel focus + add lo-fi organic Reels as a fourth format (THE PROCESS). |
| AI visual generation is a pillar, not a tool (THE RENDER) | Confirmed for reach/art. But NOT for conversion. "People spending $2k+ want to see the jeweler, the process, the real thing." | THE RENDER stays for brand awareness. THE PROCESS (jeweler content) added for conversion-oriented audience. |
| Production pipeline: Midjourney/Runway/BLNG | "Nano Banana Pro is the best image model. Cling is the best video model." Explicit toolchain recommendation from someone doing this daily. | Switch toolchain to Nano Banana Pro + Cling. Use Krea LoRa trainer for brand-consistent style. |
| Boosting posts to amplify reach | "Boosting within Instagram is flushing money away. Only Facebook Ads Manager." | No more in-app boosts. All paid amplification through Facebook Ads Manager on desktop. |
| v3 did not address ownership or schedule enforcement | "Consistency is everything. One person owns it. Three times a week minimum." | Formalize ownership and posting schedule. Name the person. Lock the days. No exceptions. |
| Celebrity cosign strategy was implied but not operational | Numan has Jay Balvin. g.money has Bad Bunny/Raúl Alejandro proximity. "Personal connections — just get them in for photos." | Activate both networks with a defined proposal. No payment required to start — photo access only. |
The Four Pillars (v4)
v3 had three. v4 adds a fourth based on Numan's core insight: the highest-converting content for a $2k+ product is not AI — it's real.
PILLAR 1: THE SHOW
From v2, confirmed by Numan. Snuffy as intellectual voice. One idea. Weekly.
No changes from v3. Execute it. Every episode ends with a password drop.
PILLAR 2: THE REACTION
Most proven format in our own data. Zero production required.
No changes from v3. Film it opportunistically. Hand someone the piece. Don't cut until they say something real.
PILLAR 3: THE RENDER
AI visuals as art, not production shortcut. Reach and brand awareness.
Updated toolchain (from v3 Midjourney/Runway → v4 Nano Banana Pro + Cling):
- Image generation: Nano Banana Pro — best output, most consistent style
- Video: Cling 2.6 — best first frame consistency, stable lighting/shadows
- Style consistency: Krea LoRa trainer — train on 50 ABALLE reference images, use for every render
- Workflow builder: Krea text-to-workflow — prompt what you want, it builds the full AI pipeline
New Cling note: Use Cling 3.0 for motion/liquid morphs (it handles dynamics better). Cling 2.6 for everything else.
Render concepts, style guides, and caption bank unchanged from v3. All still correct.
PILLAR 4: THE PROCESS (NEW)
Numan's most important recommendation. For buyers, not for reach.
"Take a camera to the jeweler for the day. Get a shitload of content. Break it up into 10–15 second reels. Milk it for a whole month."
— Numan, 2/19/26 call
What it is: Raw, close-up, real documentation of ABALLE being made. The hands of the jeweler. The tools. The metal. The fire. The finish. No AI. No scripts. Just the thing being born.
Why it works: THE RENDER converts curious people into followers. THE PROCESS converts followers into buyers. Someone spending $2,000+ needs to see that this is real, made by real hands, in a real place. That's the trust builder.
Format:
- 10–15 seconds. Close-up. Vertical. Handheld.
- ASMR sound — the actual ambient sound of the workshop. No music.
- Caption: one process word. "cast" / "set" / "finish" / "ground" / "polish"
- No CTA. Just the thing.
Production rule: One shoot day at the jeweler = 20–30 clips. That's 6–8 weeks of THE PROCESS content.
When it posts: Wednesday Reel slot (replacing Stories-only on Wed from v3). Stories remain, but add one process Reel on Wednesday.
Updated Weekly Cadence (v4)
| Day | Format | Pillar | Caption Style | CTA |
|---|
| Mon | Reel — Talking Head | THE SHOW (STRUCTURE ep) | Hook structure (5 types) | Comment SPINE |
| Tue | Reel or Carousel — AI Visual | THE RENDER | One word or technical phrase | Comment ABALLE |
| Wed | Reel — Process | THE PROCESS | One process word. No CTA. | — |
| Thu | Carousel | Founding Story chapter | Chapter hook sentence | Comment ABALLE |
| Fri | Reel — Reaction | THE REACTION | One emotional word | Comment PRICE |
| Sat | Stories only | Week recap, behind the pieces | — | — |
| Sun | — | Rest or buffer | — | — |
5 feed posts/week. Daily stories. One owner. No exceptions.
Posting time rule: Check the professional dashboard for peak audience hours. Post every single piece at that time. Every time. This is how the algorithm learns your pattern.
Grid Strategy (New in v4)
The three-across grid is over. Effective immediately.
The transition plan:
- Start posting new single-image posts right now.
- Simultaneously, archive old three-grid posts in the background. Don't rush — do 2–3 per week.
- New grid aspect ratio: 4:5 for stills. More screen space. More attention.
- Each new post = one image. One image = concentrated engagement.
Reel covers: Every reel must have a unique, manually selected cover frame. Not the default first frame. Pick the most visually distinct moment. This is what people see when browsing your profile Reels tab.
AI Toolchain (Updated)
| Tool | Use | Why |
|---|
| Nano Banana Pro | All image generation | Best model available. Consistent. Doesn't have the AI sheen that gives it away. Numan's primary tool. |
| Cling 2.6 | Video — standard motion | Best first frame consistency. Lighting/shadows don't drift. Industry standard for quality. |
| Cling 3.0 | Video — liquid/morph effects | Superior motion dynamics. Use when the concept involves transformation or fluid movement. |
| Krea LoRa trainer | Brand style consistency | 50 ABALLE reference images → 10-minute training → every prompt outputs on-brand imagery. No more prompt-by-prompt style fighting. |
| Krea text-to-workflow | Building AI pipelines | Describe what you want in plain language, Krea builds the entire workflow. Great for new concepts. |
| Krea real-time edit | Concept ideation | Sketch + prompt simultaneously. Not ideal for product integration — better for exploring ideas quickly. |
| Brand repo prompts | All image prompts | Brand bible lives in the repo. Every prompt should specify the ABALLE camera style + film grain for cohesion across all AI output. |
Simple prompt philosophy (from Numan): "I don't write paragraphs. My prompts are one or two lines." Complexity doesn't equal quality. Short, precise, stylistically anchored prompts outperform long ones.
Ads Strategy (Updated)
No more in-app Instagram boosts. Period. "It's flushing money down the toilet."
— Numan, 2/19/26
v4 paid strategy:
- All paid amplification runs through Facebook Ads Manager on desktop
- Test: boost a Reel that's already organically performing (1,000+ views). Never boost from zero.
- For a brand at ABALLE's stage: organic first, ads to amplify proof — not to manufacture it
- Product ads (conversion-focused) pair best with THE PROCESS content — real product, real process, real price
Celebrity & Influencer Strategy (Activated)
v3 named the opportunity. v4 names the people and the play.
What Numan Said
"Getting people in, high followers, personal connections — just get them in to say, yo, come and try us out. Let's take a few pictures. Just personal favors. They don't necessarily have to buy it."
The Networks
| Connection | Via | Priority | Ask |
|---|
| Jay Balvin | Numan (direct, talks every few days) | 🔴 High | Photos only — come try it on. No payment to start. |
| Raúl Alejandro | g.money (Puerto Rico proximity) | 🔴 High | Same. Photo access. Personal connection play. |
| Bad Bunny | g.money (proximity) | 🟡 Strategic | Longer-term. Prepare the proposal. One piece made for him. |
| Anima (existing) | Already a buyer | 🟢 Activate now | He already bought two. Ask for a tagged post or feature story. He's a fan — he'll say yes. |
| Lisa (Blackpink) | Via Anima (he bought her one) | 🟡 Timing-dependent | Wait for Anima to gift it. Then approach for a tag when she's wearing it. |
The Proposal Format
Numan asked: "What would be a prime proposal for someone like that?" The answer:
- Come in for 30 minutes. Try it on. We shoot.
- You keep the piece. We use the photos.
- No posting required from your end — just access.
- If they tag: that's the upside. Don't count on it. The photos are the asset.
Giveaway Strategy
Numan greenlit this. Frame it as marketing budget, not a price reduction.
Logic: A $2k bracelet giveaway has a higher expected return than a $1k+ photo shoot that gets 10 likes. Use it as marketing budget deployed differently.
v4 giveaway rules:
- Giveaway of an entry-level piece (ring at $500 or a bracelet)
- Mechanic: comment + follow + tag two people
- Do NOT put dollar figures in the caption. Numan: "Mentioning a dollar sign or figures in captions reduces reach — I've seen it happen." Just name the piece. Let people discover the price.
- Run giveaways maximum twice per year — preserves exclusivity
- Pair with a strong THE REACTION Reel in the same week
Entry-Level Products (Social Hook)
Numan recommended it. g.money's pushback (brand preservation) is valid. The v4 position:
Do: Rings at $500, pendants, belts — same design language, lower access point. Lets the 24-year-old who can't afford the bracelet enter the brand. Social content for rings drives awareness of the bracelet. The aspiration ladder works.
Don't: Rubber bracelets or steel versions. "If I make these in steel and sell them for $300, I'm not gonna wear it. Anima is not gonna wear it." The Ed Hardy trap is real — brand dilution via mass accessibility has no recovery. Chrome Hearts avoided it. ABALLE should too.
Content angle for rings/pendants: Same design language as the bracelet. Frame them as "the entry point into a fine art practice" — the most affordable piece of Snuffy's work, not a cheaper version of the bracelet.
Parallel Opportunity: Collectibles Market
Numan flagged this independently. Noting here for g.money's awareness.
Numan's thesis: The Boo Boo Pop / Bear Brick market is massive, underserved, and ripe. Low production cost (China), limited drops, colorway variations, influencer seeding (a girl puts it on her bag, it goes viral). He was about to do this with Joe Jonas before legal complications killed it. Snuffy's sculpture background is perfectly suited.
This is a separate business from ABALLE, not a ABALLE sub-product. Numan's explicit framing: "I wouldn't even link this to ABALLE. This is just the collectibles market being insane."
Snuffy already has 200 sculptural pieces unreleased. Keith wants them. This is an activated asset sitting dormant.
Potential: Low cost, high margin, viral-friendly, scalable via colorways. File this for a future sprint.
What To Build First (Priority Order, v4)
- Archive the three-grid posts. Do 2–3 today. Keep going quietly in the background.
- Book the jeweler shoot day. One day. One camera. 20–30 clips. This is 6–8 weeks of THE PROCESS content. Do this week.
- Assign one social owner. Decide today who runs the account. They set the schedule, they post, they monitor insights. One person.
- Fix the reel covers. Go through existing reels — update any that have the same or weak cover frame.
- Set up Krea LoRa trainer. Pull 50 ABALLE reference images. Train the model. All future AI output flows through this for consistency.
- Reach out to Numan about Jay Balvin. He offered. He talks to Balvin every few days. Draft the proposal. Send it this week.
- Lock posting schedule. Three days minimum — pick Mon/Wed/Fri or Mon/Wed/Sun. Post at the same time every day. Never miss a week.
- Shoot 3 Reaction Reels — invite @mr.k_tattoo or a collector. Film the reaction. Post Friday.
- Record STRUCTURE Episode 1 — Snuffy, studio, the spine-as-architecture idea. Post Monday.
- Write Founding Story Chapter 1 — Thursday carousel. It's ready. Just shoot it.
What v4 Is NOT
- Not a rebrand — the ABALLE voice, aesthetic, and world-building direction from v2/v3 are correct
- Not an abandonment of THE RENDER — AI visuals are still a pillar, but they're now clearly positioned as reach/brand content, not conversion content
- Not a volume play — ABALLE is not trying to post every day. 5x/week is the ceiling, not the floor
- Not a rubber bracelet brand — g.money's brand preservation instinct is correct
The Numan Conversation Summary (For Reference)
Numan (Numan Nuke): AI content specialist, currently at Krea. Grew Krea IG from 89k → 136k + 5M views in 6 months. One million views on a single reel the day before this call. Previously grew GigaChat IG from 12k. #1 on Rarible multiple times. Dubai-based. Open to collaboration after ABALLE tests independently for a few weeks.
His honest assessment of current ABALLE:
- "Very unique. Very niche. Hard to sell."
- "Some of the stuff is really good — the hooks where you start with a computer screen turning into a car on a roof."
- "The product shots with the tattoos and gold Rolex are beautiful. Really nice images."
- "The three-grid — I'd completely not do that at all."
- "I would focus on the story behind it, not the AI."
- "Consistency. That's the main thing. That's why nothing ever works for people."
SPINALE Content Engine v4.0 | February 2026
ABALLE AI CMO — Jackson
Source: v3 analysis (23 Reels · 30 grid posts · 136 competitor posts) + Numan strategy session (2/19/26)