Marketing Report Generator (Markdown)
Compiles prior marketing audit results into a scored, client-ready report with an action roadmap.
Text-only generation: no network calls, no credential access, no destructive defaults. Data flow is transparent (reads known project files, writes MARKETING-REPORT.md), MIT license, attributable source. Deductions: no explicit user confirmation, no warning before overwriting an existing report, and the 'Revenue Impact' framework explicitly encourages inflated sales-oriented figures ('$32,000/month') framed as a client-opening sales tool — an ethical/misleading-output risk; sensitive business data in input files is not addressed.
Scoring system is internally consistent: category weights sum to 100%, formula explicit, output template complete, and three graceful fallback paths exist when no prior data is present. Deductions: no tests; determinism is weak since scores rest on subjective LLM rubric judgments (e.g., 'Good = 17'); the worked revenue example mixes arithmetic loosely; failure feedback for missing data files is generic; reliance on scripts/analyze_page.py is referenced without this skill owning it.
Trigger conditions are clear (/market report, request for a full report), audience (agencies/solopreneurs) is explicit, and missing-data cases have boundary handling. Deductions: no declared non-fit ranges (non-English sites, non-marketing businesses); depends on web scraping with no Chinese-language support or mainland-China reachability assessment.
Well-structured docs: purpose, triggers, step-by-step instructions, output template, and principles; MIT license present; uninstall path documented. Deductions: no version/changelog, no example outputs or FAQ, referenced benchmarks ('industry standards') are unsourced, and maintenance/update responsibility rests with an unverified individual publisher.
Core task (compiling multi-source audit results into a structured client report) has a complete path and directly usable output template covering scores, findings, roadmap, and appendix. Deductions: value depends heavily on prior skill data quality; revenue impact numbers tend to be inflated without real data and require human review before delivery; comparative benefit over manually writing a report is unevidenced — only the rubric, no verified representative outputs.
The rubric and the appendix requirement to disclose data sources, limitations, and dates constitute auditable self-disclosure. Deductions: no tests, no CI, no third-party execution evidence; README sample output is marketing demo, not reproducible evidence; scoring benchmarks lack cited sources; fact and inference (revenue estimates) are mixed with only a thin confidence-labeling mechanism.
- The revenue-impact module explicitly encourages generating large dollar figures as a sales tool; these are speculative, not measured — verify and disclose assumptions before client delivery.
- Output overwrites any existing MARKETING-REPORT.md without a confirmation step; back up before running.
- Scores rest on subjective LLM rubric judgments; repeated runs on the same site may diverge — do not treat scores as objective benchmarks.
- No Chinese-language support and core data acquisition depends on scraping overseas websites; mainland-China usability is unverified.
- Publisher is unverified; no versioning or changelog, so long-term maintenance is not assured.
What does this skill do, and when should you use it?
This skill lives at skills/market-report/ in zubair-trabzada's ai-marketing-claude suite. It reads Markdown data files left by upstream audit skills (site audit, SEO, brand, competitors, funnel, ads, social, email), computes a six-category weighted marketing scorecard, and writes a full MARKETING-REPORT.md with deep dives, competitor comparison, revenue impact estimates, tiered action items, and a 30-60-90 day roadmap. It performs no new site crawling itself — its inputs come from earlier skills; if none exist, it offers to run a quick audit or produce a template. Its positioning is turning audit output into a sales-grade deliverable for agencies and freelancers.
1) Scans the project directory for upstream audit files (MARKETING-AUDIT.md, SEO-AUDIT.md, LANDING-CRO.md, COMPETITOR-ANALYSIS.md, etc.); 2) Scores six categories (Website & Conversion 25%, SEO 20%, Content & Messaging 15%, Social 15%, Email & Automation 15%, Paid Ads 10%) via per-factor point tables, computes the weighted overall score, and maps it to ratings from Critical (0-39) to Excellent (85-100); 3) Writes per-category deep dives with findings, strengths, gaps, recommendations, and revenue impact estimates using a confidence-factor-discounted formula; 4) Builds a competitor positioning matrix, content quality assessment, conversion/funnel summary, and an SEO health checklist; 5) Produces a three-tier action plan (quick wins / this month / this quarter) with impact, effort, and revenue columns, plus a week-by-week 30-60-90 roadmap; 6) Adds a methodology appendix, tools used, and a glossary, and writes MARKETING-REPORT.md.
- An agency consultant who has run /market audit and related skills needs a formal, scored report with revenue estimates before a client meeting.
- A freelancer uses the 'quick wins' list to show a prospect exactly what can be improved this week, as a deal-opening sales artifact.
- A solopreneur who has run several audit skills on their own site wants one scorecard showing which channel to invest in first.
- A team with competitor data from /market competitors needs a positioning matrix ranking the client against three rivals.
- Anyone who must justify methodology to a non-marketer client — the appendix documents scoring approach, benchmarks used, and a glossary.
What are this skill's strengths and limitations?
- The scorecard has explicit per-factor weights, point bands, and a total formula — results are reproducible and explainable.
- Revenue impact estimates are discounted by a confidence factor with a full worked example, curbing inflated claims.
- Action items specify impact level, effort in hours, and expected result — executable by a junior marketer as written.
- Methodology appendix and glossary make the report suitable for direct client delivery to non-marketers.
- Plain Markdown output with no runtime dependencies; low barrier to customizing the format.
- Limited standalone value — score quality depends entirely on whether upstream audit files exist; without them output is largely estimation.
- Scoring bands are the skill's own heuristics (e.g. 'under 2s = 15 points'); the source cites no external benchmark validation.
- Revenue figures are hypothetical extrapolations; the skill itself instructs transparency about analysis vs. estimation, so numbers must not be promised as fact.
- No test suite or sample output is provided for this skill; report quality depends on the model running it.
- Markdown only — a client-demanded PDF requires the suite's separate report-pdf skill and a reportlab install.
How do you install this skill?
The skill ships with the whole suite, not standalone: run curl -fsSL https://raw.githubusercontent.com/zubair-trabzada/ai-marketing-claude/main/install.sh | bash, or git clone https://github.com/zubair-trabzada/ai-marketing-claude.git && cd ai-marketing-claude && ./install.sh. Uninstall with ./uninstall.sh. Optional PDF support requires pip install reportlab (only needed by the suite's PDF report skill, not this Markdown one).
How do you use this skill?
First generate data with an audit command (e.g. /market audit <url>), then in Claude Code type /market report or /market report <domain>. The skill reads files like MARKETING-AUDIT.md and SEO-AUDIT.md from the project directory and writes MARKETING-REPORT.md. With no prior data it offers three options: run a quick audit first (recommended), generate from a URL and whatever you provide, or emit a fill-in template.