Two weeks of a new SEO site: every number, including the bad ones
Most SEO case studies start after the results arrived. This one starts at zero, because that's where the site currently is, and because a case study written only when the numbers look good isn't a case study — it's an advertisement.
Here is everything that happened to rankwises.com in its first two weeks: the technical work, the mistakes, and the actual Search Console data. No projections, no "expected results" — just what the dashboards say today.
Google Search Console, 24-hour view, 25 August 2026.
The uncomfortable number first
Zero clicks from 1,110 impressions. That looks like failure, and I understand why anyone would read it that way. The explanation is the third number: average position 82.3.
Google shows roughly ten results per page. Position 82 is around page nine. Realistic click-through rates that far down are a fraction of a percent, so at 1,110 impressions, zero clicks is not a bug — it's arithmetic. The pages are being shown to people; they're being shown somewhere nobody scrolls to.
Impressions vs clicks — first two weeks
What's actually encouraging in that data
The queries. Here are the top ten search terms the site appeared for:
| Query | Impressions |
|---|---|
| on page seo checker | 109 |
| on-page seo checker | 63 |
| onpage seo checker | 46 |
| onpage checker | 42 |
| on page checker | 38 |
| on page seo check | 37 |
| on page seo checker online | 35 |
| seo onpage checker | 28 |
| on page seo analyzer | 28 |
| onpage seo check | 21 |
Every one of those is someone looking for exactly the tool that exists on this site. Google has understood what the on-page analyzer is and is matching it to the right intent. That's the mechanism working correctly.
What's missing isn't relevance. It's trust in the domain — and that only accumulates through age and links, neither of which can be rushed.
What was built, and what it cost
Forty-six pages, twenty-seven articles, six free tools and one paid product, for the price of a domain. Everything is hand-written HTML with no CMS, which is why the hosting bill is zero and the pages load without waiting on JavaScript.
The technical work, measured
The site was crawled repeatedly with Screaming Frog throughout. These are real before-and-after figures, not estimates:
| Issue | Before | After |
|---|---|---|
| Non-indexable canonicals | 29 | 0 |
| Canonicalised URLs | 29 | 0 |
| Internal redirect hops | 30 | 1 |
| Invalid structured data blocks | 1 | 0 |
| Security headers configured | 0 | 6 |
Four things I broke
All four were found by something deliberately checking. None were visible during normal browsing, which is the entire point.
1. Twenty-nine canonical tags pointing at redirects
Migrating hosts introduced this. The new host strips .html from URLs, so /faq.html began 308-redirecting to /faq — while every canonical tag still pointed at the .html version. Canonicals must resolve to a live 200 page; pointing them at redirects means search engines can disregard them entirely. It affected 96.67% of URLs. The site looked completely normal.
2. One unescaped quotation mark
A page title containing quote marks went into a JSON-LD block unescaped, breaking the JSON. The structured data on that page wasn't degraded — it was unreadable.
3. A mobile menu that hid every link
Responsive CSS hid the navigation on narrow screens without providing a replacement menu. Since Google indexes mobile-first, that's an internal linking failure, not just a usability one. Worse: the toggle script was later found missing from 31 of 45 pages entirely.
4. No Open Graph tags, for weeks
Every link shared anywhere rendered as bare text with no image or preview. Found only because the site was audited specifically for it before a social post.
These are written up in full, with fixes, in technical SEO examples. The canonical checker exists because of the first one.
Indexing: 18 of 79 known URLs
Search Console reports 18 indexed and 61 not indexed. Broken down:
| Reason | Pages | Assessment |
|---|---|---|
| Crawled — currently not indexed | 25 | Domain age. Nothing to fix. |
| Alternate page with proper canonical | 16 | www duplicates — since 301'd to apex. |
| Redirect error | 10 | Redirect chains — since consolidated to one hop. |
| Page with redirect | 5 | Correct behaviour. No action needed. |
| Discovered — currently not indexed | 5 | Queued, not yet crawled. Normal. |
Only 26 of those 61 were genuine problems, and both have been fixed. The remaining 25 are the honest reality of a two-week-old domain: Google crawls the pages, evaluates them, and declines to store them yet.
Who's actually visiting
Cloudflare logged roughly 1,800 requests in 24 hours. The largest single source was 1,290 requests from Singapore, hitting paths like /wp-admin/ and /wp-content/plugins/ — automated bots probing for WordPress vulnerabilities on a site that has never run WordPress. Harmless, and every public site receives this.
Genuine Googlebot activity in the same window: 33 requests. Real human visitors: effectively zero, consistent with Search Console reporting no clicks.
What I'd tell a client in this position
- The content and technical work are done. Adding a seventh tool or a twenty-eighth article changes nothing at position 82.
- Backlinks are the constraint. The site has essentially none, and that is the single largest input into domain trust.
- Don't optimise click-through rate yet. Rewriting titles to be more clickable is pointless on page nine. Fix position first, then CTR.
- Measure in months. Meaningful movement on low-difficulty terms typically starts around month two or three for a new domain. Checking daily just generates anxiety.
Next update
Roughly 30 days from now, with the same metrics: impressions, clicks, average position, indexed pages, and backlinks acquired. Whether they improve or not.
Want the same analysis on your own site? The on-page analyzer runs about 40 checks free, or a free audit covers what a tool can't.