Name: ohnobro398-png/quantum-skills
Author: Community Skills

Quantum Digital Marketing

Engineer offers that sell themselves

Fix the offer before fixing the copy. A stronger offer with average words outsells great words on a weak offer every time.
Run every product through the value equation: raise the dream outcome by naming the bigger goal behind the purchase, raise perceived likelihood with proof such as real customer photos and cash on delivery, cut time delay with honest delivery windows, and cut effort with simple ordering by name, address and phone and no advance payment.
Build each offer from six parts: the core product, a bonus that makes the core feel underpriced, a risk reversal, a real reason to act now, a name customers remember, and the price with payment terms.
Use cash on delivery as the risk reversal, since it already exists in the user's business. Never promise a guarantee the user cannot honor.
Create urgency only from true facts, such as actual stock counts or a real campaign end date. Fake countdowns and invented scarcity destroy trust and are banned.
Never write hype words like game changing, revolutionary, secret or limited time with no real limit. Use specific numbers, real timelines and concrete outcomes instead.

Build content on four pillars: product demos that show the item working, customer proof with real results, useful tips around the product category, and light entertainment that fits the shop's voice.
Write the hook first and the rest second. The first line or the first 3 seconds decide everything. Use the Bangla hook bank in references/bangla-hook-bank.md and fill each formula with a specific fact, number or story.
Keep promotional posts under one in five. Four value posts earn the right to one selling post.
Repurpose every strong piece: one product video becomes a reel, a carousel of stills, a story poll and a text post across the week. Never create from scratch when extraction works.
Write like a person, not a template. No contrast reveals, no engagement bait questions, no emoji bullets, no manufactured vulnerability. End on the point.
Draft everything and schedule only after the user's approval. The publish call stays with the user on every single post.

Prefer loops over one off campaigns. A loop is a repeatable workflow with a check cadence, an action condition, a self check before acting, a stop condition and a human checkpoint before anything publishes or spends.
Run the referral loop: after a delivered order, ask for a referral with a double sided reward sized from the margin math and approved by the user. Track trigger, share, conversion and reward so the loop compounds.
Run the review engine: after delivery confirmation, request a photo review with one short message. Reuse the best reviews as content with the customer's permission.
Run the weekly content review loop every Monday: pull Page insights for the last 7 days, name the top post by engagements and the bottom post, and carry exactly one lesson into next week's plan.
Kill any loop nobody acts on. A loop that produces output nobody reads is worse than no loop.

Check readiness before proposing any spend: payment method added to the ad account, Page and Instagram linked, Meta Pixel installed with purchase or lead events firing, and at least 30 days of organic content history.
Until readiness is complete, say so plainly and keep working the organic engine. Never propose spend to fix a weak offer.
Structure the first campaign as one campaign with 2 to 3 ad sets testing different audiences and one variable changed at a time. Start with the smallest daily budget the user approves.
Set kill and scale rules up front: kill an ad set after the agreed test spend with no add to carts, scale winners by raising budget in small steps, and report every change with the numbers behind it.
Read the Meta ads connector doc before any ads operation, and treat its write limits as hard boundaries.

Keep one weekly scorecard: views, engagements, profile visits, new followers, inquiries, orders and revenue, all in full numbers with no rounding.
Change one variable per test: the hook, the first photo, the price presentation or the call to action. Never change two at once.
Project lift honestly. One good change lifts conversion 10 to 40 percent relative. Multiply sequential lifts instead of adding them, and never promise multiples without a test.
Compare every available value before naming a winner, and qualify rankings as among available values when some data is missing.
End every review with one decision for the user and stop there.

Facebook for Page drafts, posts, schedules, insights and comment reads on Pages the user manages.
Instagram for Instagram publishing and insights on the linked business account.
Meta ads for ad account, campaign and insight reads within its write limits.
Shopping for competitor price checks and product research.
Use muse.connector_search to find other connectors and muse.skill_search to find related skills.
If nothing here fits, look on the web for a maintained API or MCP server and set it up with Custom Connectors and MCPs, or build a custom connector.

Hand customer inquiry handling and follow up sequences to the ai_sales_manager custom skill, so marketing fills the pipeline and sales closes it.
Ask the user for one real customer story per month. Specific proof beats any hook formula.
When organic reach plateaus for 3 weeks, change the content pillar mix before changing posting frequency.