Our own data

How we measure

If a number about our own site is on this page, you can repeat the check that produced it. Each row says when it was measured, with what, what the result was and how to run it yourself.

Last updated 18 October 2026

The rules

What we hold ourselves to.

Our own data only. Every row is about devmarde.com or a tool we run, and is labelled that way.
Every row has a date. New results are added as new rows. Old rows are not edited, so you can see what changed.
We show what a tool returned, not a rounded-up version of it, and we say what a check cannot tell you.
We do not claim that an AI assistant mentions us, or that AI answers sent us leads, until we have measured it. The last section lists what is still unmeasured.

The numbers are put together, with dates and what we changed, in the case study of devmarde.com. The same checks sit behind Crawl Audit and our guide to AI crawler user agents.

The log

Measurements, newest first.

The same home-page lab scores, re-run after we limited the moving card borders to cards near the screen and delayed analytics.

18 October 2026

Result: Mobile: performance 59 (was 58), LCP 3.4 s, blocking time 1,990 ms. Desktop: performance 96 (was 61), LCP 1.0 s, blocking time 80 ms. Accessibility 94, best practices 73, SEO 100 in both.
Tool: Lighthouse 13.5.0, run locally against the live home page, after the card-border and analytics change
Repeat it: npx lighthouse https://devmarde.com --only-categories=performance,accessibility,best-practices,seo
Limit: A lab run, not field data. Mobile is still low and we are working on it.

Lab speed, accessibility, best-practice and SEO scores for the home page.

18 October 2026

Result: Mobile: performance 58, accessibility 94, best practices 73, SEO 100, LCP 3.8 s, CLS 0, blocking time 1,530 ms. Desktop: performance 61, accessibility 94, best practices 73, SEO 100, LCP 1.9 s, CLS 0, blocking time 460 ms.
Tool: Lighthouse 13.5.0, run locally against the live home page
Repeat it: npx lighthouse https://devmarde.com --only-categories=performance,accessibility,best-practices,seo
Limit: A lab run, not field data from real visitors. Scores move a few points between runs.

The home page requested with 11 different user-agent names: a normal browser, OAI-SearchBot, GPTBot, ChatGPT-User, ClaudeBot, Claude-SearchBot, Claude-User, PerplexityBot, Perplexity-User, Applebot and Googlebot.

18 October 2026

Result: 11 of 11 returned HTTP 200 with the same page, 372,809 bytes each, in 0.8 to 1.7 seconds.
Tool: curl, from our own computer, 01:59 UTC
Repeat it: curl -s -o /dev/null -A "<name>" -w "%{http_code} %{size_download}" https://devmarde.com/
Limit: This sends each bot's name from our computer. It does not send the bot's network address, so a firewall that checks both could treat the real bot differently.

The same 11 user-agent names against the home page.

14 October 2026

Result: 11 of 11 returned HTTP 200, 372,444 bytes each.
Tool: curl, from our own computer, 21:19 UTC
Repeat it: The same command as above.
Limit: Same limit as above.

Which crawlers the file names, and what it allows.

18 October 2026

Result: 20 groups: the default group plus 19 named crawlers (GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-SearchBot, Claude-User, PerplexityBot, Perplexity-User, Google-Extended, Applebot, Applebot-Extended and 8 others). Every named group allows / and disallows /api/, /admin/ and one query-string pattern. One sitemap line.
Tool: Reading devmarde.com/robots.txt
Repeat it: Open https://devmarde.com/robots.txt in a browser and count the User-agent lines.

The structured-data nodes on the home page.

18 October 2026

Result: 9 nodes: Organization (also typed ProfessionalService), WebSite, 3 Person, WebPage, BreadcrumbList, Article and FAQPage.
Tool: Reading the JSON-LD in the home page source
Repeat it: View the page source and search for application/ld+json, or paste the URL into Google's Rich Results Test.
Limit: This lists what is on the page. It does not say how any search engine or assistant uses it.

How many URLs the sitemap lists.

18 October 2026

Result: 63 URLs.
Tool: Reading devmarde.com/sitemap.xml
Repeat it: Open https://devmarde.com/sitemap.xml and count the <loc> entries.

Whether the file exists and how large it is.

18 October 2026

Result: HTTP 200, 11,330 bytes.
Tool: Reading devmarde.com/llms.txt
Repeat it: Open https://devmarde.com/llms.txt.
Limit: We have no evidence that any AI company reads this file. We publish it because it costs little.

Impressions and clicks in the site's first four weeks in Google.

16 September to 14 October 2026 (pulled 17 October)

Result: 3,027 impressions and 10 clicks. The page with the most impressions was the guide to AI search optimization, 2,320 impressions at an average position of 47.4.
Tool: Google Search Console, performance report for devmarde.com, web search
Repeat it: In Search Console, open Performance and set the dates to 16 September to 14 October 2026.
Limit: A young site. These numbers describe the first weeks, not a trend.

Not measured yet

What we do not claim.

Whether AI assistants name DevMarDe in their answers

We have not run a baseline. Until we have one, we make no claim in either direction.

Leads that came from AI answers

Nothing is tracked for this yet, so we show no number.

Our own Crawl Audit result

We will add it here, with its date and the tool version, once the scoring method page is published.

Check your own site the same way.

Crawl Audit runs these kinds of checks on your domain. The free check needs no sign-up.

Run the free check About DevMarDe