What we observed
| Area | Evidence | What it means |
|---|---|---|
| Crawl files | robots.txt and sitemap.xml returned homepage HTML with HTTP 200. | The server was returning the wrong resource for those URLs. |
| Missing pages | A nonexistent URL returned the homepage with HTTP 200. | Missing pages were not being signalled with a proper 404 response. |
| Indexing | Search Console recorded “Crawled — currently not indexed”; a live test found the homepage eligible for indexing. | Accessibility and actual inclusion were different observations. The cause of non-indexing was not established. |
| Mobile lab performance | Lighthouse performance 100; LCP 0.9 s, total blocking time 0 ms, CLS 0. | This test did not identify speed as the main implementation priority. |
| Real-user performance | No CrUX field data was available. | A real-user Core Web Vitals outcome could not be assessed. |
| Content and tracking | One landing page, no dedicated service/About/privacy pages, and no Google analytics tag. | The site needed clearer service information and a measurable enquiry path. |
Test context
The mobile PageSpeed Insights test used Lighthouse 13.6.0 with an emulated Moto G Power and slow 4G. The reported accessibility score was 96 and SEO score 92. These are lab results from that test, not permanent site ratings or proof of rankings.
The stored Search Console result reflected an earlier crawl; the live test was performed on 14 November. No manual action or security issue was shown during that review.
What this implementation addresses
- Serve real robots and sitemap files, canonical redirects and an actual 404 page.
- Explain the audit, implementation and AI-search services on distinct pages.
- Add business, methodology and privacy information using verified facts.
- Use consistent Organization and Service entities in structured data.
- Connect consent-based analytics and count a lead only after a saved request.
What has not been demonstrated
This baseline does not show a ranking improvement, traffic increase, conversion uplift or AI citation gain. Those outcomes require post-release collection and sufficient observation. The initial AI prompt panel has not been run, so there is no measured citation baseline yet.
We will add dated results when evidence is available. A fixed technical defect is valuable, but it should not be described as a proven business outcome.
Read the measurement methodology or explore how an implementation sprint works.