SEO Lab

New Domain Search Performance Analysis

Perfect technical setup, yet only 42 clicks. A cross-tool diagnostic using GA4, SEO_CHECK, and GSC to unpack the impression/click gap on a new domain.

10 min read2026-04-25
GA4

Step 1: Discovering the Gap

Ninety days after launching seo.codequest.work. GA4 shows 15 to 20 daily active users, with 652 cumulative users on /ja pages. The tool itself works fine. But organic search traffic was far below expectations.

Compared to another site I run (codequest.work), sessions per page were strikingly low. Content and tools were in place. Structured data and llms.txt were implemented. So why wasn't traffic coming?

The first thing to isolate: is this a technical setup issue? Is the site failing to communicate with Google, or is Google receiving the signals and still not rewarding it? The two require fundamentally different responses.

SEO_CHECK

Step 2: Checking for Miscommunication

I ran the key pages through SEO_CHECK. Every page scored high. Title, meta description, structured data, internal links, OGP tags -- all aligned with Google's official guidelines. No technical miscommunication.

This is a critical distinction. Communication with Google is working correctly. The cause lies outside technical setup. Fixing title tags or adding more structured data won't solve this. We need to look from a different angle.

GSC

Step 3: Verifying Results -- Where the Real Cause Appears

Opening GSC was a surprise. There were 17,000 impressions. Only 42 clicks. CTR 0.25%.

Google recognizes this site. It's showing pages in search results. The problem isn't 'not being found' -- it's 'being found but not clicked.' If I had only looked at GA4, the conclusion would have been 'no traffic.' GSC's impression data completely reframed the problem.

From here, I dug into the page-level data. Three hypotheses emerged.

Hypothesis 1: The Ranking Ceiling of a New Domain

Observation

The overall average position is 50.7. Most pages appear on page 2 or beyond in search results. Even with 17,000 impressions, pages are effectively invisible to users.

Hypothesis

A new domain lacks domain authority, so there's a ranking ceiling regardless of technical setup quality. SEO_CHECK measures 'are you communicating correctly with Google,' but 'how much Google trusts your domain' is a separate variable.

Verification

But page-level data reveals exceptions. meta-tag-check ranks at 7.7 with 20% CTR -- breaking through the ceiling on a niche query. Meanwhile, en/seo-check has 9,204 impressions but ranks at 61.2. On competitive queries, the ceiling is clearly visible.

Analysis

For new domains, a 'raise everything' strategy doesn't work. The correct approach is to identify niche queries where the ceiling can be broken and concentrate resources there. Use SEO_CHECK to ensure the technical foundation is solid, then use GSC to find queries where rankings can climb -- niches with less competition. The success of meta-tag-check supports this hypothesis.

Hypothesis 2: Ranking 4.7 with Zero Clicks -- A SERP Structure Problem

Observation

llms-txt-check ranks at position 4.7. SEO_CHECK score is high. Yet clicks are zero. GA4 confirms zero traffic to this page.

Hypothesis

This query either has extremely low search volume, or SERP features like rich snippets or AI Overviews are answering the query without requiring a click. High ranking does not equal success.

Verification

SEO_CHECK confirms no miscommunication. GSC shows the impression count itself is low. This means very few people search for this query at all. llms.txt is still a new concept, and search demand hasn't matured.

Analysis

High ranking does not mean success. You need to verify search volume and SERP structure. SEO_CHECK verifies 'is the message reaching Google correctly,' but 'does demand exist for this query' can only be answered by GSC. A textbook case for cross-tool analysis. That said, interest in llms.txt is growing with AI search adoption, and maintaining this ranking has first-mover value.

Hypothesis 3: 9,204 English Impressions -- Discovering a Hidden Market

Observation

en/seo-check has 9,204 impressions -- more than half of all impressions. But it ranks at 61.2. GA4 shows near-zero sessions from English-speaking users.

Hypothesis

SEO checker demand in the English-speaking market is orders of magnitude larger than in Japanese. Technical setup is fine (confirmed by SEO_CHECK), but competition density in English is too high for the ranking to climb.

Verification

This discovery was only possible through GSC. GA4 can't show it (no traffic arriving). SEO_CHECK can't show it (technical setup is correct). Only GSC's impression data visualized this 'market not yet captured.'

Analysis

The 9,204 impressions justify investing in English content and backlink acquisition. If ranking moves from 61.2 to within the top 30, clicks will follow since impressions are already there. However, Ahrefs, Moz, and Semrush dominate English rankings, so direct competition should be avoided. Differentiation through niche angles -- like auto-generated fix code, a feature competitors lack -- is essential.

Decision Frameworks from This Analysis

  1. The niche concentration principle -- On new domains, high SEO_CHECK scores don't guarantee rankings. Use GSC to identify breakable niches and concentrate resources there.
  2. SERP structure verification -- When rankings are high but clicks are zero, suspect SERP structure issues. After ruling out miscommunication with SEO_CHECK, the impression/click ratio in GSC identifies structural problems.
  3. Impressions are a market map -- GSC impressions map 'markets not yet captured.' Even when GA4 shows no traffic, impressions in GSC reveal where demand exists.

Each of these judgments requires crossing all three tools. GA4 alone ends at 'no traffic.' SEO_CHECK alone ends at 'setup is fine.' GSC alone can't build a hypothesis for why impressions don't convert to clicks. Combining the three allows accurate structural diagnosis and informed next actions.

Start by Checking Your Google Communication

Whether your domain is new or established, the first step is verifying technical setup. If there's a communication gap, rankings are a secondary concern. Enter a URL for a free 30-point SEO audit with auto-generated fix code.

FAQ

Which GSC metric should you check first on a new domain?
Start with impressions. Impressions confirm Google recognizes your pages as search result candidates, making it the primary indicator of index health. Next, check average position by page to understand which queries surface your content. Clicks and CTR become meaningful only after rankings stabilize, so deprioritize them in the early stage of a new domain.
How do you diagnose impressions with no clicks?
Test three hypotheses in order: (1) ranking too low (page 2+ drives CTR below 1%), (2) SERP structure issues (rich snippets or AI Overviews answering the query directly), (3) title/description failing to attract clicks. First use SEO_CHECK to rule out technical miscommunication with Google, then analyze GSC data by page and query to isolate the cause.
What does GA4 show that GSC doesn't, and vice versa?
GA4 visualizes post-arrival behavior — sessions, pageviews, engagement time, conversions. GSC visualizes pre-arrival performance — impressions, clicks, position, queries. Even when GA4 shows zero traffic, GSC impression data can reveal untapped demand. The two tools cover complementary blind spots.
What if SEO_CHECK scores are high but rankings are low?
SEO_CHECK evaluates technical communication quality with Google. High scores mean the problem is outside technical setup. For new domains, investigate domain authority gaps (backlinks, domain age), content depth versus competitors, and the balance between search volume and competition density for target queries.
What are the limitations of this analysis?
This article is based on 90 days of data from a single site (seo.codequest.work). Results will differ across niches, domain ages, and competitive landscapes. GSC impression counts exclude anonymized queries, so actual impressions may be higher. Treat findings as correlations, not causation, and run the same diagnostic process on your own site to validate.