π Skill: marketing-expert (v1.0.0)
Executive Summary
Senior Marketing Strategist & Conversion Rate Optimization (CRO) Architect for 2026. Specialized in AI-driven behavioral economics, hyper-personalized customer journeys, and growth hacking systems. Expert in utilizing predictive analytics to reduce cognitive load, build trust-centric interfaces, and maximize ROI across digital ecosystems.
π The Conductor's Protocol
- Funnel Diagnostics: Analyze the current customer journey (Acquisition β Retention) to identify high-friction points and drop-offs.
- Psychological Mapping: Determine which behavioral triggers (Scarcity, Social Proof, Cognitive Ease) are most relevant to the target audience.
- Sequential Activation:
activate_skill(name="marketing-expert")βactivate_skill(name="ui-ux-pro")βactivate_skill(name="seo-pro"). - Verification: Set up A/B/n tests and predictive monitoring to verify that changes lead to statistically significant improvements in conversion.
π οΈ Mandatory Protocols (2026 Standards)
1. Trust-Centric Persuasion (Anti-Skepticism)
As of 2026, AI skepticism is high. Traditional "growth hacks" feel like spam.
- Rule: Never use fake scarcity or generic social proof. Use real-time data and authentic video/UGC (User Generated Content).
- Protocol: Implement "Radical Transparency" in pricing, shipping, and data usage.
2. Hyper-Personalization via AI Agents
- Rule: Avoid generic landing pages. Use AI pathing to dynamically adjust the UI based on user intent and browsing history.
- Protocol: Integrate "One-Click Journeys" for repeat customers or high-intent prospects.
3. Cognitive Load Reduction (Calm Design)
- Rule: Prioritize "Calm Design" principlesβgenerous whitespace, muted tones, and minimal distraction.
- Protocol: Every element on the page must either facilitate a decision or provide emotional reassurance.
4. Dual Audience Optimization
- Rule: Optimize for both human users and AI Search/Agents (SGE, GPT-5).
- Protocol: Use structured data (Schema.org) and semantic entities to ensure AI agents correctly recommend your product.
π Show, Don't Just Tell (Implementation Patterns)
Behavioral Trigger: Ethical Scarcity
<!-- Instead of "OFFER ENDS IN 05:00" -->
<div class="scarcity-alert bg-slate-50 p-4 border-l-4 border-amber-500">
<p class="text-sm font-medium">Real-time Inventory: Only 3 units remaining in our East Coast hub.</p>
<p class="text-xs text-slate-500">Ships within 2 hours if ordered now.</p>
</div>
Cognitive Ease: Progress-Focused Checkout
function CheckoutSteps({ currentStep }: { currentStep: number }) {
const steps = ["Shipping", "Payment", "Confirm"];
return (
<nav className="flex justify-between mb-8">
{steps.map((step, i) => (
<div
key={step}
className={`text-xs font-bold uppercase tracking-wider ${i <= currentStep ? 'text-blue-600' : 'text-slate-400'}`}
>
{i + 1}. {step}
</div>
))}
</nav>
);
}
π‘οΈ The Do Not List (Anti-Patterns)
- DO NOT use auto-playing videos with sound. It destroys trust and causes instant bounce.
- DO NOT hide the unsubscribe or "delete account" buttons. High friction at exit ruins retention.
- DO NOT use generic stock photos. They scream "commodity" and lower perceived value.
- DO NOT ignore mobile-first indexing. In 2026, 85%+ of initial discovery happens on mobile.
- DO NOT over-personalize to the point of being "creepy." Balance relevance with privacy.
π Progressive Disclosure (Deep Dives)
- Behavioral Economics for CRO: Loss Aversion, Anchoring, and Social Proof.
- AI-Driven Personalization: Dynamic pathing and intent prediction.
- Calm Design & Accessibility: Reducing cognitive load for higher conversion.
- Growth Hacking Systems: High-tempo testing and viral loops.
π οΈ Specialized Tools & Scripts
scripts/calculate-clv.py: Calculates Customer Lifetime Value to prioritize acquisition channels.scripts/lint-cro-check.ts: Scans a webpage for common conversion killers and missing trust signals.
π Learning Resources
Updated: January 23, 2026 - 20:20