Glimana shows only measured data: what we read from your own Google accounts and what our own crawler sees. We never present another company's estimate as our measurement. If something could not be measured, the dashboard says "not measured" — never zero, never "clean".
What comes from where
| Data | Source | Refresh | Limits |
|---|---|---|---|
| Organic clicks, impressions, average position, queries | Google Search Console API (read-only) | Daily | 16 months of history. Google reports with a 2–3 day delay. Queries Google anonymises for privacy are not in the query list, so total clicks can be higher than the sum of query rows. |
| Visits, sources, device, country | Google Analytics 4 Data API (read-only) | Daily | 400 days of history. Visitors who decline cookies or use ad blockers do not reach GA4. |
| Page issues (102 rules) | Our own crawler (GlimanaBot) | At your plan's crawl frequency and on demand | Respects robots.txt. JavaScript-rendered content is processed only for a sample: the home page, the 50 most-clicked pages and pages with noindex/canonical in their raw HTML. A rule that cannot run shows as "not running". |
| Speed (LCP, INP, CLS) | PageSpeed Insights (lab) and Chrome UX Report (real users) | Weekly, Saturday 04:00 (Istanbul time) | Lab and real-user data are shown separately. Chrome UX Report only has data for sites with enough traffic and covers the last 28 days. |
| AI bot access | robots.txt and a live request from our server | Weekly, Sunday 03:00 (Istanbul time) | The robots.txt verdict is definitive. The live request does not come from the bot's own IP range, so it is an indicator: some firewalls block our request and still let the real bot through. |
| Competitors | Chrome UX Report (the competitor's real-user speed data) and a 25-page sample from the competitor's sitemap | On demand | No paid estimation service. The competitor's site is not fully crawled; the screen says how many pages were checked. |
| Broken external links | Our own checks | In hourly batches | Only certain breaks: 404/410, a domain that does not exist, or a server error twice in a row. Bot blocks (401/403/429) are not counted as broken. |
| Keyword ideas | Bing Webmaster Tools (if you connect it) | Weekly | Empty unless a Bing account is connected. |
| Article research | Web search by the AI provider you choose (OpenAI, Anthropic or Gemini, with your own key) | When writing | Glimana opens and verifies every source it finds; URLs that don't open, social media/forum links and URLs the search did not return are dropped. Unverified figures are marked "to verify" in the text. |
What we don't use
- "Domain authority"-style scores — another company's estimate from its own formula; not a number Google uses.
- Estimated traffic databases — your real traffic is already in Search Console and GA4; no need to guess.
- Backlink indexes — we don't buy third-party link databases; the link data we show comes from our own crawl.
What we don't measure
An "AI answer visibility" score. ChatGPT's or Gemini's answer changes with the person, the location and the moment; a score from a handful of sample prompts is not a reliable measurement. We measure what can be measured instead: whether AI bots can read your site.
Privacy
Google connections are read-only; Glimana cannot change anything on your site or in your Google account. Details: Privacy policy.