Founder project, Faiz Azizan.
Between June and July 2026 I published 19 Chrome extensions to the Chrome Web Store, covering SEO auditing, CRO analysis, GTM/GA4 tracking validation, AI-agent readiness, and developer-productivity tools. This case study tracks how visible each extension is across two very different discovery surfaces:
- Traditional search (Google SERP), does the extension's Chrome Web Store listing rank for its own product name?
- Generative Engine Optimization (GEO), does the extension get surfaced when someone asks an AI assistant (ChatGPT, Gemini, Claude) to solve the exact problem the extension was built for?
The gap between these two numbers is the point. Ranking #1 on Google for your own brand name is close to guaranteed. Getting recommended by an LLM for a generic, unbranded use-case query is not, and that's the harder, more valuable signal for GEO work going forward.
Methodology
Google SERP test: for each extension I searched the exact product name plus "chrome extension" (e.g. "GTM GA4 DataLayer Auditor" chrome extension) and recorded whether the Chrome Web Store listing appeared, and at what position.
ChatGPT / Gemini / Claude test: for each extension I asked the assistant a natural, unbranded query describing the problem the extension solves, never mentioning its name, and recorded whether the assistant recommended it. Example: for "GTM GA4 DataLayer Auditor," the query was "Recommend a Chrome extension to audit GTM and GA4 dataLayer events on a website." Any direct Chrome Web Store link an assistant returned was checked against the actual listing ID to rule out hallucination before being counted as a genuine citation.
Results
| # | Extension | Google SERP | ChatGPT | Gemini | Claude |
|---|---|---|---|---|---|
| 1 | CRO & UI Sniper | Not ranked (too new; YouTube demo ranks #1) | Not cited | X | X |
| 2 | Image Inspector SEO | #1 | Not cited | X | X |
| 3 | GTM GA4 DataLayer Auditor | #1 | Not cited | Not cited | X |
| 4 | AI Agent Readiness Checker | #1 | Cited, #1 recommendation | X | X |
| 5 | Web Security Headers Checker | #1 | X | X | X |
| 6 | Product Page Schema (GMC) Listing Tools | #1 | X | X | X |
| 7 | Web Checker: SEO & Speed | #1 | X | X | X |
| 8 | AI Bot Access Checker | #1 | X | X | X |
| 9 | CRO Lens – Product Page Auditor | #1 | X | Cited (direct link, verified) | Cited (direct link, verified) |
| 10 | Backlink Checker | #1 | X | Cited | X |
| 11 | Semantic SEO Auditor | #1 | X | X | X |
| 12 | Screenshot & Blur | #1 | X | X | X |
| 13 | PyAutoGUI Script Recorder | Not ranked (too new) | X | X | X |
| 14 | SEO Snap & Convert | #1 | X | X | X |
| 15 | Schema Visualizer | #1 | X | X | X |
| 16 | PDF Scholar Helper | Not ranked (too new) | X | X | X |
| 17 | Heading Visualizer | #1 | X | X | X |
| 18 | N-Gram SEO Analyzer | #1 | Cited, #1 recommendation | Cited | X |
| 19 | Website Tracking Scanner | #1 | X | X | Cited (direct link, verified) |
Key findings
1. Brand-name Google ranking is nearly automatic, and not a meaningful competitive signal. 16 of 19 extensions (84%) rank #1 on Google for a search of their own exact product name. That's expected, almost nobody else uses the same unique name, so it shouldn't be read as SEO strength. The three that don't rank yet (CRO & UI Sniper, PyAutoGUI Script Recorder, PDF Scholar Helper) are simply the newest listings and likely aren't fully indexed yet.
2. GEO citation, being recommended for an unbranded problem, is a much harder bar. Five extensions have been genuinely cited by at least one AI assistant when a user described their problem without naming the product:
- AI Agent Readiness Checker (ChatGPT, #1)
- N-Gram SEO Analyzer (ChatGPT #1, Gemini)
- CRO Lens – Product Page Auditor (Gemini, Claude, both verified via direct-link match)
- Backlink Checker (Gemini)
- Website Tracking Scanner (Claude, verified via direct-link match)
Three others (GTM GA4 DataLayer Auditor, CRO & UI Sniper, Image Inspector SEO) were not cited. Those platforms recommended established, longer-standing alternatives instead.
3. Verification matters more than the headline number. Two Gemini responses that initially looked like citations were checked against the actual Chrome Web Store listing IDs. One (CRO Lens) matched exactly and was counted. Another (N-Gram) referenced a differently-formatted ID that was confirmed correct only on manual check. This step is worth doing every time, assistants do sometimes generate plausible-looking but broken or fabricated links, and citing one in a client-facing report without checking it would undermine the whole study.
What this suggests
- New, un-established tools with no backlinks, reviews, or external mentions are very unlikely to be picked up by an LLM for a generic query, even if they rank #1 on Google for their own name. Citation signals (reviews, third-party mentions, structured data) matter more than on-page SEO alone for GEO visibility.
- The extensions that did get cited tend to be either earlier releases with more time to accumulate signals, or in categories with less competing content for the assistant to choose from.
- This is a baseline on newly launched products, and it is designed to be re-run. Repeating the same queries every 60 to 90 days, as reviews, backlinks, and usage accumulate, turns a single snapshot into a trend line, and that trend is the real measure of whether GEO work is paying off.