Marketing And Growth

---
name: marketing-and-growth
description: Growth engineering, conversion rate optimization (CRO), technical and local SEO, AI-search visibility, landing page copywriting, analytics, campaign/launch strategy, email, customer research, positioning, and content systems. Use for marketing, growth, launch strategy, SEO, local SEO, GEO/AEO, or funnel optimization.
---

# Marketing and Growth Engineering

Marketing work should be driven by **business truth + customer evidence + current channel data + measurement**, not by a pile of clever prompts.

Use deep references **JIT**, not all at once:

- broad/substantial marketing strategy, research, positioning, copy, email, content systems -> `references/marketing-operating-system.md`;
- SEO/search/growth-loop work -> `references/seo-growth-loop.md`;
- local SEO / Google Business Profile / local AI visibility -> `references/local-seo.md` (normally alongside the general SEO reference when technical/search work is also involved);
- product/feature launches, launch research, video/content production, repurposing -> `references/launch-content-system.md`.

The first reference distills the useful procedures behind the bookmarked “500 prompts / replace an agency” article into reusable capabilities and workflow chains instead of preserving hundreds of prompts verbatim. The other two distill the Grok SEO/Helena and Okara/Motion/faceless-video bookmark clusters respectively.

Load only the reference(s) that actually apply to the current task.

## Core principles

1. **Context before copy.** Inspect the offer, audience, proof, objections, brand voice, goal, and available evidence before generating campaign assets.
2. **Clarity over buzzwords.** Communicate a specific truthful value proposition. Avoid generic “revolutionary/seamless/next-generation” filler.
3. **Research before positioning.** Use customer language, category/competitor evidence, existing analytics, search data, support/review material, and project truth where available.
4. **Procedures over magic prompts.** Reusable knowledge should be expressed as inputs -> decision/artifact -> evidence/QA -> next stage.
5. **Current sources for changing facts.** Search behavior, competitors, platform rules, prices/features, APIs, SEO guidance, regulations, and channel norms can change; verify them live when material.
6. **Proof before claims.** Never fabricate metrics, testimonials, logos, scarcity, guarantees, customer results, or product capabilities.
7. **Instrument important funnels.** Define the success signal and review window before declaring a campaign/change successful.
8. **Close the loop.** Baseline -> hypothesis -> action -> measurement -> learning -> keep/iterate/revert/stop.

## Capability selection

Load/use only the capabilities the task actually needs. Typical capabilities are:

- customer/ICP research;
- competitor/category analysis;
- positioning and message house;
- content strategy;
- sales/landing copy;
- repurposing/content atomization;
- SEO opportunity + content briefs;
- local SEO / GBP / local search and AI-search visibility;
- email lifecycle;
- campaign/launch architecture;
- analytics/CRO/learning loops.

Do not run all ten because a project contains the word “marketing”.

## Technical SEO foundation

For implementation-level SEO, verify against the current stack/search-engine guidance and project requirements. Common concerns include:

- crawlability/indexability;
- canonical URLs;
- sitemap/robots behavior;
- semantic page structure;
- structured data appropriate to the actual entity/content;
- metadata/social cards;
- internal linking;
- performance/Core Web Vitals where relevant;
- duplicate/thin/generated content risk;
- locale/hreflang where the site actually needs it.

Do not add every schema type or metadata field mechanically. Match the page and current official guidance.

For local businesses, use `references/local-seo.md` rather than inventing a second SEO methodology. For generative/answer-engine visibility, treat GEO/AEO as an SEO/search surface unless current platform evidence requires a distinct tactic.

## Copy / value proposition

A useful copy sequence is:

`reader state -> desired action -> promise -> mechanism/difference -> proof -> objections -> information sequence -> copy -> factual/brand review`

Headlines should explore different **strategic angles**, not just synonyms. CTA text should make the next action and expectation clear.

Use real social proof only. If no proof exists, change the composition/message rather than manufacturing it.

## CRO / funnel optimization

Start from observed or testable friction.

```text
funnel evidence
-> friction hypothesis
-> bounded message/design/product change
-> QA
-> measurement window
-> result
-> learning
```

Track events that correspond to the real funnel rather than blindly installing a generic event taxonomy.

## Content / distribution

A content system should connect audience need to a business objective:

```text
questions / pains / search demand / product evidence
-> content pillars or topic clusters
-> strong source asset
-> platform-native derivatives
-> distribution
-> downstream metrics
-> future topic/format updates
```

Repurposing may create X/LinkedIn/short-video/newsletter/carousel/etc. derivatives, but it must preserve factual consistency and adapt format/angle to the platform rather than cloning text everywhere.

## Campaign / launch

For a material launch, prefer a staged pipeline:

`research -> positioning -> message house -> launch brief -> channel architecture -> core assets -> derivatives -> distribution -> questions/data -> follow-up -> review`

Use current launch references when they materially improve the decision; do not copy another launch's wording/brand assets or infer causality from a viral post.

## Email

Design email sequences from the customer state and job of each message, not from a fixed “5-email template”. Respect the actual jurisdiction/platform's consent, suppression, unsubscribe, and messaging requirements.

## Analytics and learning

Choose metrics that match the objective: qualified leads, conversion, revenue/margin, clicks, watch/completion, search performance, retention, objections, etc.

A learning should record evidence, scope, confidence, decision impact, and when to re-check. Do not turn one campaign or two posts into a universal law.

## References and provenance

When external sources materially affect a durable marketing decision:

- distinguish canonical/current evidence from creator/vendor claims and inspiration;
- record the actual principle or factual premise used;
- re-verify dynamic facts when needed;
- update the project's reference ledger when the influence is durable.

Large prompt/source collections are treated as **source material for procedures**, not runtime dependencies. Agency-replacement/pricing framing is not treated as verified outcome evidence.

## Completion questions

Before claiming substantial marketing work complete, be able to answer:

- What audience/business problem did this solve?
- What evidence or current sources informed it?
- Which intermediate decisions/artifacts led to the output?
- Are all material claims supportable?
- What metric/evidence will tell us whether it worked?
- What is the review/next-action condition?