Host My Blog

Google stopped showing FAQ rich results from 7 May 2026. Your FAQ text still works for readers. Keep FAQPage markup only if it matches what is on the page, and do not add it to get a Google result. A short check: hostmy.blog/faq-schema-goo…

Analytics will never show you AI crawler traffic. Google Analytics runs JavaScript and crawlers do not execute it. The access log is the only honest record.

The obvious first command lies. Include the client IP field and every row is unique, so uniq -c collapses nothing and every count comes back as 1.

Crawlers arrive from large IP ranges and almost never repeat an address. Drop that one field and the same data becomes a report.

Read the user agent field explicitly rather than matching the whole line, and expect zero for Google-Extended and Applebot-Extended. Those are robots.txt control tokens, not fetcher names.

hostmy.blog/access-log-ai-…

Update your old content is not an actionable instruction when you have 400 posts.

Two numbers make it concrete. Stale pages were roughly 3 times more likely to lose citations, and the median cited page was 298 days old.

So age is not the filter. A post last touched 900 days ago that still names 2021 as the current state of something has gone stale. A 2023 post revised last month has not.

The script ranks a whole archive by staleness through the public WordPress REST API. No plugin, no database access, no packages.

hostmy.blog/find-stale-pos…

The mean cited sentence ran 9.27 words. The median was 10, and the 6 to 10 word band alone carried 45.2 percent of all citations.

Nothing longer than 18 words was cited even once across 153,425 citations from six AI platforms. That is a ceiling, not a preference.

The script measures any page or draft against that distribution on stock Python 3. Four numbers, no grade attached.

One more thing before you chase a readability score. Cited pages were bimodal: 22.9 percent very easy, 20.5 percent very confusing, and only 2.6 percent in the Flesch 50 to 59 middle.

hostmy.blog/measure-quotab…

Google's Rich Results Test checks one URL at a time and asks whether Google will draw a rich snippet. That is a narrower question than whether your structured data is valid at all.

curl and a grep count answer the first thing: does the page emit any JSON-LD. Zero happens more often than people expect on custom themes.

The Python script here checks required fields per type across a whole site. On one real page it found 14 typed nodes inside a single block.

That is the part hand-rolled checkers miss. Modern SEO plugins emit an @graph, so a checker reading only the top-level @type reports clean.

hostmy.blog/check-schema-w…