GIG Gulf · Content + GEO Audit · Pipeline & Tool Gaps
Every audit, in order
Four stages — Download → Index → Save to DB → Analyse. Status measured against work done this session. Benchmarked against the Visible proposal (the standard GIG is being pitched).
Benchmark note: Visible (Gilles Praet) is pitching GIG a $3.5k–$9.5k/month GEO+SEO retainer. Their moat is one thing we don't yet have: a 100-prompt × 4-LLM AI-visibility measurement. Everything else in their proposal, we can match from first-party data we now hold. The pipeline below closes the gap.
Stage 1 · Download — acquire raw data
| # | Audit / pull | Source | Status |
|---|
| 1 | Full site crawl + URL inventory | Firecrawl sitemap | DONE 1,765 URLs |
| 2 | Blog corpus scrape — 131 posts → records | Firecrawl | PARTIAL index ✓, 1/131 |
| 3 | Document-library inventory — 527 PDFs | Firecrawl + PDF tools | TODO |
| 4 | GSC export — pages · queries · dates (per market) | Search Console | PARTIAL UAE 16mo ✓ |
| 5 | GSC decay export — 3mo vs prior 3mo (fatigue input) | Search Console compare | TODO |
| 6 | GA4 export — page engagement + quote/conversion events | GA4 | PARTIAL live access + multi-market verified; per-page export pending |
| 7 | Google Ads export — paid keywords + landing pages | Google Ads | TODO |
| 8 | Competitor crawl — Sukoon · Tawuniya · Salama · Orient + Lemonade/Hippo | Firecrawl (no access needed) | TODO |
| 9 | AI-visibility capture — prompts × LLMs (ChatGPT·Claude·Google; Perplexity/Gemini when keyed) | LLM APIs (built in-house) | DONE tool built + baseline run (20 prompts ×3) |
| 10 | SERP + AI-Overview presence capture | Serper (UAE geo) | TODO |
| 11 | Backlink / authority profile | Ahrefs | BLOCKED plan tier |
Stage 2 · Index — structure & classify the raw
| # | Audit / step | Status |
|---|
| 12 | URL taxonomy — classify all 1,765 by type · LOB · language · funnel-stage · format | TODO |
| 13 | Blog record population — 22-field schema per post | schema proven |
| 14 | Date-stamping pass — publish/updated dates | TODO |
| 15 | Query classification — brand vs non-brand · intent · LOB · market | DONE 80/20 |
| 16 | Prompt → citation index — which LLM cited which page, per prompt (GEO) | TODO (needs #9) |
Stage 3 · Save to database — persist a queryable grid
| # | Audit / step | Status |
|---|
| 17 | Airtable master audit base — pages × metrics × scores × disposition | TODO (Airtable live) |
| 18 | Per-URL metrics join — GSC + GA4 + Ads onto the inventory | GSC in hand |
| 19 | Fatigue index compute — age + decay + cannibalisation + CTR-gap + repetition | inputs partial |
| 20 | Richness scoring — the 10-element /20 per asset | TODO (rubric ready) |
| 21 | GEO phase scoring — Category Formation → Attribute Recall → Competitive Selection → Trust, per LOB/market | TODO (needs #9) |
Stage 4 · Analyse — turn data into the plan
| # | Audit / analysis | Status |
|---|
| 22 | Traffic concentration + dead-inventory | DONE top5=68%, 37%<10 |
| 23 | Category ROI — clicks per post by lane | DONE Motor 897 vs Travel 86 |
| 24 | Brand vs non-brand demand gap | DONE 80/20 |
| 25 | CTR quick-wins — high-impression, low-CTR giants | STARTED |
| 26 | Cannibalisation + technical-junk (duplicate URLs, language-switcher index bloat) | STARTED |
| 27 | Schema / AI-answer readiness audit (JSON-LD, FAQPage, llms.txt) | TODO |
| 28 | Competitor gap — volume · format · topics · share-of-voice | TODO |
| 29 | AI-citation share — the GEO headline metric | DONE 83% overall / 92% category (baseline) |
| 30 | Multi-market + Arabic parity | STARTED Arabic dark |
| 31 | Gap → 6-month roadmap + needle-movers | DRAFTED |
Missing tools — what to buy or build
Everything else the audit needs is keyed and working: Firecrawl, Serper, Airtable, the LLM stack, document/visual production, Cloudflare/Vercel hosting.