<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Lucas Fernández Díaz</title>
    <description>The latest articles on DEV Community by Lucas Fernández Díaz (@lucasfernandezdiaz).</description>
    <link>https://dev.to/lucasfernandezdiaz</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4128336%2F5d2bc3aa-dc65-452e-88f9-76c2bed1461e.png</url>
      <title>DEV Community: Lucas Fernández Díaz</title>
      <link>https://dev.to/lucasfernandezdiaz</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/lucasfernandezdiaz"/>
    <language>en</language>
    <item>
      <title>What Does "Response Time" Actually Measure? TTFB With curl and the Navigation Timing API</title>
      <dc:creator>Lucas Fernández Díaz</dc:creator>
      <pubDate>Thu, 19 Nov 2026 13:00:00 +0000</pubDate>
      <link>https://dev.to/lucasfernandezdiaz/what-does-response-time-actually-measure-ttfb-with-curl-and-the-navigation-timing-api-422</link>
      <guid>https://dev.to/lucasfernandezdiaz/what-does-response-time-actually-measure-ttfb-with-curl-and-the-navigation-timing-api-422</guid>
      <description>&lt;p&gt;&lt;em&gt;Written with help from an AI assistant for research and editing; the commands were run and checked by hand, and the opinions are ours.&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;"Your site is slow" is not a diagnosis. Most of the time the number behind it is some flavor of response time, and response time on its own is too vague to fix anything. The first question is always the same: which metric, measured where, from which request?&lt;/p&gt;

&lt;p&gt;The metric that usually hides behind the phrase is Time to First Byte. Google's &lt;a href="https://web.dev/articles/ttfb" rel="noopener noreferrer"&gt;TTFB article on web.dev&lt;/a&gt; defines it as the elapsed time between the start of a navigation and &lt;code&gt;responseStart&lt;/code&gt;, the moment the first byte of the response begins to arrive. It includes redirects, service worker startup if there is one, DNS lookup, connection and TLS negotiation, and the request itself up to that first byte.&lt;/p&gt;

&lt;p&gt;It does not include anything the browser does after that. Parsing, styles, images, scripts and painting all come later.&lt;/p&gt;

&lt;h2&gt;
  
  
  What does website response time actually measure?
&lt;/h2&gt;

&lt;p&gt;It depends on the metric named in the report. If it's TTFB, it measures the trip to the first byte, connection steps included. It doesn't mean the page is loaded or that a visitor can use it.&lt;/p&gt;

&lt;p&gt;That distinction matters because a small business site can have a perfectly fine first byte and still show a blank hero for two seconds while a large image and three scripts arrive. Or the reverse: a slow first byte from a cold server, then a light page that renders immediately.&lt;/p&gt;

&lt;h2&gt;
  
  
  How do I measure TTFB from the terminal?
&lt;/h2&gt;

&lt;p&gt;curl can break a single request into phases with &lt;code&gt;-w&lt;/code&gt;:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;curl -s -o /dev/null -w &lt;span class="s1"&gt;'
dns      %{time_namelookup}
connect  %{time_connect}
tls      %{time_appconnect}
ttfb     %{time_starttransfer}
total    %{time_total}
'&lt;/span&gt; https://example.com/
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Every value is in seconds, measured from the start of the request, so they're cumulative, not separate durations. &lt;code&gt;time_starttransfer&lt;/code&gt; is the closest curl gets to TTFB. To see how much of it is the server thinking rather than the network, subtract &lt;code&gt;time_appconnect&lt;/code&gt; from &lt;code&gt;time_starttransfer&lt;/code&gt;.&lt;/p&gt;

&lt;p&gt;Add &lt;code&gt;-L&lt;/code&gt; if the URL redirects. Without it, curl stops at the first response, and you'll be timing the redirect instead of the page. The web.dev definition counts redirect time as part of TTFB, so a chain of two redirects before the real page is worth knowing about.&lt;/p&gt;

&lt;p&gt;Run it several times. One request is an anecdote: DNS caching, a CDN edge warming up or a busy connection can move it a lot.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;&lt;span class="k"&gt;for&lt;/span&gt; i &lt;span class="k"&gt;in&lt;/span&gt; 1 2 3 4 5&lt;span class="o"&gt;;&lt;/span&gt; &lt;span class="k"&gt;do&lt;/span&gt;
  curl -s -o /dev/null -w &lt;span class="s1"&gt;'%{time_starttransfer}\n'&lt;/span&gt; https://example.com/
&lt;span class="k"&gt;done&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h2&gt;
  
  
  How do I read it in the browser?
&lt;/h2&gt;

&lt;p&gt;The browser gives you the same number for the navigation a real visitor makes. Paste this in the console on the page you care about:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight javascript"&gt;&lt;code&gt;&lt;span class="kd"&gt;const&lt;/span&gt; &lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="n"&gt;nav&lt;/span&gt;&lt;span class="p"&gt;]&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;performance&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;getEntriesByType&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="s1"&gt;'navigation'&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;&lt;span class="p"&gt;;&lt;/span&gt;
&lt;span class="nb"&gt;console&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;log&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="s1"&gt;'TTFB (ms):'&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="nb"&gt;Math&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;round&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;nav&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="n"&gt;responseStart&lt;/span&gt; &lt;span class="o"&gt;-&lt;/span&gt; &lt;span class="n"&gt;nav&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="n"&gt;startTime&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;&lt;span class="p"&gt;;&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;web.dev's article shows the same approach with a &lt;code&gt;PerformanceObserver&lt;/code&gt;, and the &lt;code&gt;web-vitals&lt;/code&gt; library wraps it as &lt;code&gt;onTTFB&lt;/code&gt;. The Network panel in DevTools shows the timing breakdown of the main document request, which helps tell waiting for the server apart from downloading the rest.&lt;/p&gt;

&lt;h2&gt;
  
  
  What does a visitor see while the page loads?
&lt;/h2&gt;

&lt;p&gt;After the first byte, the browser reads the markup, requests styles, images and scripts, and paints what it can. A visitor may still be looking at a blank area or a partial layout, or at a page that looks ready and doesn't respond to a tap yet.&lt;/p&gt;

&lt;p&gt;Core Web Vitals cover those parts: loading, interactivity and visual stability, as Google's &lt;a href="https://web.dev/articles/vitals" rel="noopener noreferrer"&gt;Web Vitals overview&lt;/a&gt; explains. TTFB isn't one of them. web.dev says so directly and treats it as a diagnostic that sits in front of the loading metrics.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;What you notice&lt;/th&gt;
&lt;th&gt;What to inspect first&lt;/th&gt;
&lt;th&gt;The question it answers&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;The page takes a while to start&lt;/td&gt;
&lt;td&gt;The main document request and its timing&lt;/td&gt;
&lt;td&gt;Is the delay before the server replies, or somewhere else in the request?&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;The page starts but the main content appears late&lt;/td&gt;
&lt;td&gt;The largest visible element and its image or font requests&lt;/td&gt;
&lt;td&gt;Is a heavy asset or a render-blocking resource holding up what matters?&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;The page appears but feels hard to use&lt;/td&gt;
&lt;td&gt;Whether links, menus and forms respond, and whether the layout jumps&lt;/td&gt;
&lt;td&gt;Is the page ready for interaction and visually stable?&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  Is there a "good" TTFB?
&lt;/h2&gt;

&lt;p&gt;web.dev gives a rough guide: 0.8 seconds or less is good, more than 1.8 seconds is poor. It also says that because TTFB isn't a Core Web Vitals metric, sites don't strictly need to hit that threshold, as long as the slow first byte isn't keeping the loading metrics from being good.&lt;/p&gt;

&lt;p&gt;So treat the number as a pointer. A TTFB of 1.4 seconds on a hosted builder is a question about the platform and the CDN, and often not one you can answer with your own code.&lt;/p&gt;

&lt;h2&gt;
  
  
  Which speed number should I trust?
&lt;/h2&gt;

&lt;p&gt;One whose collection method you know. A lab test, like Lighthouse or a curl loop, describes a controlled run. Field data, which &lt;a href="https://developers.google.com/speed/docs/insights/faq" rel="noopener noreferrer"&gt;PageSpeed Insights&lt;/a&gt; shows when there's enough of it for the page or the site, reflects real visits. Neither label tells you what's wrong on its own. Connect the metric to what you saw on the page and to the change you're considering.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Can I compare my site with a competitor using one test?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
One test is a clue. Pages differ in content, caching, location and network conditions. Compare equivalent pages under similar conditions, and use the result to form a question, not to declare a winner.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Does a fast first byte mean the whole page is fast?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
No. Images, styles and scripts can still take a while to arrive or render, and the page may not respond to input yet.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;What should I ask when a report says my site is slow?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Which page and which resource were measured, whether the metric is TTFB or a user experience metric, and whether it's lab or field data. Then ask what the evidence says to fix first. A useful answer names the delay and links it to a change you can verify.&lt;/p&gt;

&lt;p&gt;If a speed report is pushing you or a client toward a rebuild, run the curl loop and the console snippet first. If you get numbers you can't explain, drop them in the comments with the platform you're on.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://contento.solutions/blog/what-website-response-time-measures?utm_source=devto&amp;amp;utm_medium=referral&amp;amp;utm_campaign=stock-2026-11" rel="noopener noreferrer"&gt;Contento Solutions&lt;/a&gt;, where I write about SEO and AI visibility for service businesses.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>webdev</category>
      <category>performance</category>
      <category>webperf</category>
      <category>beginners</category>
    </item>
    <item>
      <title>Why Can't Google Find My New Page? Finding Orphan Pages With curl and Your Sitemap</title>
      <dc:creator>Lucas Fernández Díaz</dc:creator>
      <pubDate>Thu, 12 Nov 2026 13:00:00 +0000</pubDate>
      <link>https://dev.to/lucasfernandezdiaz/why-cant-google-find-my-new-page-finding-orphan-pages-with-curl-and-your-sitemap-3b6n</link>
      <guid>https://dev.to/lucasfernandezdiaz/why-cant-google-find-my-new-page-finding-orphan-pages-with-curl-and-your-sitemap-3b6n</guid>
      <description>&lt;p&gt;&lt;em&gt;I wrote this with help from an AI assistant for research and editing; the checks, the script and the opinions are mine.&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;A page nobody links to depends on luck. Google's own &lt;a href="https://developers.google.com/search/docs/fundamentals/seo-starter-guide" rel="noopener noreferrer"&gt;SEO Starter Guide&lt;/a&gt; says it primarily finds pages through links from other pages it already crawled. Your homepage gets crawled. The service page you published last weekend and never linked from anywhere might not.&lt;/p&gt;

&lt;p&gt;That's an orphan page: it exists, it returns a 200, and zero internal links point to it. The fix is boring. Finding them is the part worth automating.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why can't Google find my new page?
&lt;/h2&gt;

&lt;p&gt;Because no other page on your site points to it. Google's &lt;a href="https://developers.google.com/search/docs/crawling-indexing/links-crawlable" rel="noopener noreferrer"&gt;link best practices&lt;/a&gt; put the standard in one line: every page you care about should have a link from at least one other page on your site.&lt;/p&gt;

&lt;p&gt;Orphans happen for dull reasons. Builders let you create a page without placing it anywhere. On Squarespace, pages in the Not Linked section of the Pages panel are live but sit outside the navigation, and people use that section as a drafts drawer. Ad landing pages and service pages that lost their menu spot in a redesign are the other classics.&lt;/p&gt;

&lt;h2&gt;
  
  
  Doesn't my sitemap already tell Google about it?
&lt;/h2&gt;

&lt;p&gt;It helps, but it doesn't replace links. Google's &lt;a href="https://developers.google.com/search/docs/crawling-indexing/sitemaps/overview" rel="noopener noreferrer"&gt;sitemaps overview&lt;/a&gt; puts the linking first: if your pages are properly linked, Google can usually discover most of the site. The same page says a sitemap doesn't guarantee that every item in it gets crawled and indexed.&lt;/p&gt;

&lt;p&gt;So a URL that's in &lt;code&gt;sitemap.xml&lt;/code&gt; and missing from every page's HTML sends a mixed message. The sitemap says the page exists. The rest of the site says it isn't worth pointing to. That mismatch is exactly what we can detect with a few commands.&lt;/p&gt;

&lt;h2&gt;
  
  
  How do I check a single page by hand?
&lt;/h2&gt;

&lt;p&gt;Three checks, in this order.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;1. Search Console, Links report.&lt;/strong&gt; Under Internal links, open Top linked pages and look for the URL. If it isn't there, Google hasn't seen internal links to it. The report lags behind new pages, so don't trust it alone for anything recent.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;2. The sitemap.&lt;/strong&gt; Confirm the URL is listed:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;curl -s https://example.com/sitemap.xml &lt;span class="o"&gt;|&lt;/span&gt; grep -o &lt;span class="s1"&gt;'&amp;lt;loc&amp;gt;[^&amp;lt;]*&amp;lt;/loc&amp;gt;'&lt;/span&gt; &lt;span class="o"&gt;|&lt;/span&gt; grep your-page-slug
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;&lt;strong&gt;3. The pages that should link to it.&lt;/strong&gt; Fetch the homepage, the services page and the posts closest to the topic, and count how many times the slug appears in the served HTML:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;&lt;span class="nv"&gt;SLUG&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;your-page-slug
&lt;span class="k"&gt;for&lt;/span&gt; p &lt;span class="k"&gt;in&lt;/span&gt; / /services /blog&lt;span class="o"&gt;;&lt;/span&gt; &lt;span class="k"&gt;do&lt;/span&gt;
  &lt;span class="nb"&gt;printf&lt;/span&gt; &lt;span class="s1"&gt;'%s '&lt;/span&gt; &lt;span class="s2"&gt;"$p"&lt;/span&gt;
  curl -s &lt;span class="s2"&gt;"https://example.com$p"&lt;/span&gt; &lt;span class="o"&gt;|&lt;/span&gt; grep -c &lt;span class="s2"&gt;"$SLUG"&lt;/span&gt;
&lt;span class="k"&gt;done&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;A &lt;code&gt;0&lt;/code&gt; on every line means no link. This check reads what the server actually sends, so it doesn't wait for Search Console to catch up. One caveat: if your menu is rendered by JavaScript, curl won't see those links, and neither will a crawler that doesn't run your scripts. Google says it can generally only crawl a link that is an &lt;code&gt;&amp;lt;a&amp;gt;&lt;/code&gt; element with an &lt;code&gt;href&lt;/code&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  How do I find every orphan at once?
&lt;/h2&gt;

&lt;p&gt;Compare two lists: every URL in the sitemap, and every internal URL linked from any sitemap page. Whatever is in the first list and not in the second is an orphan candidate.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;&lt;span class="c"&gt;#!/usr/bin/env bash&lt;/span&gt;
&lt;span class="c"&gt;# Usage: ./orphans.sh https://example.com&lt;/span&gt;
&lt;span class="nv"&gt;SITE&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s2"&gt;"${1%/}"&lt;/span&gt;
&lt;span class="nv"&gt;HOST&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s2"&gt;"${SITE#*://}"&lt;/span&gt;
curl -s &lt;span class="s2"&gt;"$SITE/sitemap.xml"&lt;/span&gt; &lt;span class="o"&gt;|&lt;/span&gt; grep -o &lt;span class="s1"&gt;'&amp;lt;loc&amp;gt;[^&amp;lt;]*&amp;lt;/loc&amp;gt;'&lt;/span&gt; &lt;span class="o"&gt;|&lt;/span&gt; sed &lt;span class="s1"&gt;'s/&amp;lt;[^&amp;gt;]*&amp;gt;//g'&lt;/span&gt; &lt;span class="o"&gt;|&lt;/span&gt; sort -u &lt;span class="o"&gt;&amp;gt;&lt;/span&gt; sitemap.txt
: &lt;span class="o"&gt;&amp;gt;&lt;/span&gt; linked.txt
&lt;span class="k"&gt;while&lt;/span&gt; &lt;span class="nb"&gt;read&lt;/span&gt; -r url&lt;span class="o"&gt;;&lt;/span&gt; &lt;span class="k"&gt;do&lt;/span&gt;
  curl -s &lt;span class="s2"&gt;"$url"&lt;/span&gt; &lt;span class="o"&gt;|&lt;/span&gt; grep -o &lt;span class="s1"&gt;'href="[^"#?]*'&lt;/span&gt; &lt;span class="o"&gt;|&lt;/span&gt; sed &lt;span class="s1"&gt;'s/href="//'&lt;/span&gt; \
    &lt;span class="o"&gt;|&lt;/span&gt; sed &lt;span class="s2"&gt;"s#^/#$SITE/#"&lt;/span&gt; &lt;span class="o"&gt;|&lt;/span&gt; grep &lt;span class="s2"&gt;"$HOST"&lt;/span&gt; &lt;span class="o"&gt;|&lt;/span&gt; sed &lt;span class="s1"&gt;'s#/$##'&lt;/span&gt; &lt;span class="o"&gt;&amp;gt;&amp;gt;&lt;/span&gt; linked.txt
  sleep 1
&lt;span class="k"&gt;done&lt;/span&gt; &lt;span class="o"&gt;&amp;lt;&lt;/span&gt; sitemap.txt
sort -u linked.txt -o linked.txt
sed &lt;span class="s1"&gt;'s#/$##'&lt;/span&gt; sitemap.txt &lt;span class="o"&gt;|&lt;/span&gt; sort -u &lt;span class="o"&gt;|&lt;/span&gt; comm -23 - linked.txt
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;What it does, line by line:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Pulls every &lt;code&gt;&amp;lt;loc&amp;gt;&lt;/code&gt; from the sitemap into &lt;code&gt;sitemap.txt&lt;/code&gt;. If your sitemap is an index of other sitemaps, run it once per child sitemap.&lt;/li&gt;
&lt;li&gt;Fetches each of those pages and extracts every &lt;code&gt;href&lt;/code&gt;, dropping fragments and query strings.&lt;/li&gt;
&lt;li&gt;Turns relative links (&lt;code&gt;/services&lt;/code&gt;) into absolute ones and keeps only links to your own host.&lt;/li&gt;
&lt;li&gt;Strips trailing slashes on both sides so &lt;code&gt;/blog/&lt;/code&gt; and &lt;code&gt;/blog&lt;/code&gt; match.&lt;/li&gt;
&lt;li&gt;Prints the sitemap URLs that never appear as a link target. &lt;code&gt;comm -23&lt;/code&gt; needs both inputs sorted, which is why everything goes through &lt;code&gt;sort -u&lt;/code&gt;.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The &lt;code&gt;sleep 1&lt;/code&gt; is there on purpose. It's your own server, but there's no reason to hammer it.&lt;/p&gt;

&lt;p&gt;Read the output with judgment. The homepage often shows up because it links to itself with a full URL that differs from the sitemap version. Category and tag archives may show up too. The ones to act on are the pages that sell something.&lt;/p&gt;

&lt;h2&gt;
  
  
  Where should the links come from?
&lt;/h2&gt;

&lt;p&gt;From pages that already get attention, and from places where a reader would want the next step. This is the order I work through:&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Link from&lt;/th&gt;
&lt;th&gt;Why&lt;/th&gt;
&lt;th&gt;Example anchor text&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Homepage&lt;/td&gt;
&lt;td&gt;Usually the most crawled page you have&lt;/td&gt;
&lt;td&gt;Emergency plumbing in Austin&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Main services page&lt;/td&gt;
&lt;td&gt;Where people compare what you offer&lt;/td&gt;
&lt;td&gt;Couples therapy for new parents&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Navigation menu&lt;/td&gt;
&lt;td&gt;One click from everywhere&lt;/td&gt;
&lt;td&gt;Couples therapy&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Related posts&lt;/td&gt;
&lt;td&gt;Catches readers when they want help&lt;/td&gt;
&lt;td&gt;book a couples therapy session&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Contact or booking page&lt;/td&gt;
&lt;td&gt;Reassures people about to decide&lt;/td&gt;
&lt;td&gt;see what couples therapy includes&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;The homepage and the services page alone move a page from hidden to findable.&lt;/p&gt;

&lt;h2&gt;
  
  
  Does the anchor text matter?
&lt;/h2&gt;

&lt;p&gt;Yes. Google uses it to understand what the target page is about, and its link guidance lists "Click here to learn more" and "Read more" as examples of text that's too generic. The test from the same document works well in code review too: read only the anchor text, out of context, and ask whether it makes sense by itself.&lt;/p&gt;

&lt;p&gt;If the link is an image, its &lt;code&gt;alt&lt;/code&gt; attribute becomes the anchor text. That's one more reason not to leave it empty on a linked image.&lt;/p&gt;

&lt;h2&gt;
  
  
  What do I do after adding the links?
&lt;/h2&gt;

&lt;p&gt;Run URL Inspection in Search Console on the orphan and click Request indexing. Do the same for the page where you added the link, because Google has to recrawl that page to see the new path. Then rerun the script. An empty output is the goal.&lt;/p&gt;

&lt;p&gt;If the page gets crawled and still doesn't show up for searches, the problem has moved from discovery to something else, and that needs a different set of checks.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;How many internal links does a page need?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Google says there's no magical ideal number, and one link is the minimum it asks for. For a page that sells something, I want the homepage, the services page and at least one related post pointing at it.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Is a footer link enough?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
It counts, and it beats nothing. A link inside a paragraph, with descriptive words, tells people and Google much more than one of twenty footer links.&lt;/p&gt;

&lt;p&gt;There's no sample of mine behind this piece, so it has no numbers of mine. The Google pages cited were read on 10 November 2026. If you run the script on your own site, tell me in the comments what kind of pages came out orphaned. I'd like to know whether builders or redesigns cause more of them.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://contento.solutions/blog/why-cant-google-find-my-page-orphan-pages?utm_source=devto&amp;amp;utm_medium=referral&amp;amp;utm_campaign=stock-2026-11" rel="noopener noreferrer"&gt;Contento Solutions&lt;/a&gt;, where I write about SEO and AI visibility for service businesses.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>seo</category>
      <category>webdev</category>
      <category>beginners</category>
      <category>bash</category>
    </item>
    <item>
      <title>WWW or non-WWW: does it matter for your website?</title>
      <dc:creator>Lucas Fernández Díaz</dc:creator>
      <pubDate>Tue, 10 Nov 2026 20:59:01 +0000</pubDate>
      <link>https://dev.to/lucasfernandezdiaz/www-or-non-www-does-it-matter-for-your-website-528e</link>
      <guid>https://dev.to/lucasfernandezdiaz/www-or-non-www-does-it-matter-for-your-website-528e</guid>
      <description>&lt;p&gt;&lt;em&gt;I wrote this with help from an AI assistant for research and editing; the measurements and the opinions are mine.&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;It doesn't matter whether your site uses www or not. What matters is that it picks one, and that every other version of the address sends people there with a permanent redirect. In September I checked 54 therapy practice websites in the US, and 8 of them didn't.&lt;/p&gt;

&lt;h2&gt;
  
  
  Does it matter if my website uses www or not?
&lt;/h2&gt;

&lt;p&gt;No. Google has said for years that either is fine, and the sites that do this well split almost evenly. Of the 46 sites in my sample that handled it correctly, 26 chose the www version and 20 chose the bare domain. Nobody won, because there is nothing to win.&lt;/p&gt;

&lt;p&gt;The part that matters is quieter. Your domain can be typed four ways: with or without www, and with http or https. All four should end at the same address, and three of them should get there with a permanent redirect. When that is true, Google sees one site. When it isn't, it can see two.&lt;/p&gt;

&lt;h2&gt;
  
  
  Which should I choose, www or non-www?
&lt;/h2&gt;

&lt;p&gt;If your site already works, keep what you have. Switching buys you nothing and costs a few weeks of Google recrawling. If you are starting fresh, the choice comes down to plumbing, not rankings.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;The bare domain&lt;/strong&gt; (yoursite.com) is shorter, it is how most people say an address out loud, and it reads cleaner on a card or a van. For a small business site that is usually the deciding argument.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;The www version&lt;/strong&gt; (&lt;a href="http://www.yoursite.com" rel="noopener noreferrer"&gt;www.yoursite.com&lt;/a&gt;) is technically a subdomain, and that helps on bigger setups. DNS rules don't allow a CNAME record on the bare domain, so pointing it at a host or a CDN needs a workaround your registrar may call ALIAS, ANAME or CNAME flattening. And cookies scoped to the bare domain are sent to every subdomain you add later, which a www site keeps contained.&lt;/p&gt;

&lt;p&gt;Neither of those is a ranking factor. Pick the one your platform sets up with less friction, then make the other three addresses redirect to it.&lt;/p&gt;

&lt;h2&gt;
  
  
  What happens when both versions of my site load?
&lt;/h2&gt;

&lt;p&gt;Google has to guess which one is the real site. It usually guesses well, but it is still a guess, and links that point at the other version count less clearly toward the one it picked. Your analytics can split visits between two addresses, and an AI assistant that reads your site can quote whichever address it found first.&lt;/p&gt;

&lt;p&gt;In my sample, the 8 failures came in three shapes. Three sites loaded the same page at two or four different addresses, with no redirect at all, one of them still serving plain http. Three had one version that didn't load: a typo away from an error page. Two redirected, but different versions ended at different addresses.&lt;/p&gt;

&lt;p&gt;None of these owners did anything wrong on purpose. The setting usually came with the domain or the hosting, and nobody went back to check it.&lt;/p&gt;

&lt;h2&gt;
  
  
  How do I check which version my site redirects to?
&lt;/h2&gt;

&lt;p&gt;It takes two minutes. Open a private browser window and type each of the four versions of your address, one at a time. Watch the address bar after the page loads. All four should land on exactly the same address, with the padlock.&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;code&gt;http://yoursite.com&lt;/code&gt; should end at your chosen address through a 301 or 308 redirect.&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;http://www.yoursite.com&lt;/code&gt; should end at your chosen address through a 301 or 308 redirect.&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;https://yoursite.com&lt;/code&gt; should end at your chosen address: it loads directly if it is the one you chose, otherwise through a 301 or 308.&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;https://www.yoursite.com&lt;/code&gt; should end at your chosen address: it loads directly if it is the one you chose, otherwise through a 301 or 308.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The browser hides the type of redirect, so to check how each one gets there use a free redirect checker such as httpstatus.io, or run &lt;code&gt;curl -I http://yoursite.com&lt;/code&gt; in a terminal and read the first line. You are looking for 301 or 308 on every version except the one you chose, and a 200 on that one.&lt;/p&gt;

&lt;p&gt;If you want to see what else a machine notices about your site in the same sitting, I wrote a &lt;a href="https://contento.solutions/blog/twenty-minutes-what-a-machine-knows-about-your-business" rel="noopener noreferrer"&gt;twenty minute version of this check&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  Is a 302 redirect enough, or does it have to be a 301?
&lt;/h2&gt;

&lt;p&gt;Use a 301 or a 308. Both say the move is permanent. A 302 or a 307 says the move is temporary, and Google treats a temporary redirect only as a weak hint about which address is the real one. It may still end up choosing your preferred version, but you are asking it to work that out when you could simply tell it.&lt;/p&gt;

&lt;p&gt;The good news from the sample: none of the 46 sites that passed relied on a 302 anywhere in the chain. When this goes wrong on a small site, it is almost never the redirect type. It is the missing redirect.&lt;/p&gt;

&lt;h2&gt;
  
  
  How do I fix it on Squarespace, Wix or WordPress?
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Squarespace.&lt;/strong&gt; In your domain settings, set the domain you want as the primary one. Squarespace redirects the other versions to it, and with SSL turned on it sends http to https. If a version still fails, it is usually a DNS record at the registrar that points somewhere else.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Wix.&lt;/strong&gt; The domain you connect is the primary one, and Wix redirects to it. Check that the www record and the bare domain both point to Wix, following its connection instructions, or one of the two versions will not load.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;WordPress.&lt;/strong&gt; In Settings, General, make the WordPress Address and the Site Address the same, with https and your chosen www choice. The redirect itself usually lives with your host: most hosting panels have a force https option and a preferred domain option. If yours doesn't, your host can add the rule for you.&lt;/p&gt;

&lt;p&gt;Whatever the platform, run the four-address check again after the change. A redirect you didn't test is a redirect you hope exists, and if your pages still aren't showing up after that, the next place to look is &lt;a href="https://contento.solutions/academy/indexed-but-not-showing-up-on-google" rel="noopener noreferrer"&gt;why a site can be indexed and still not appear&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  Questions I get about this
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Will switching from www to non-www hurt my rankings?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Not if every old address redirects permanently to the new one. Expect some movement for a few weeks while Google recrawls. If your site already works, there is no reason to switch.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Do I need to tell Google which version I chose?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
The redirects tell it. A canonical tag on each page pointing to the chosen address and a sitemap that only lists that version reinforce the same message.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Does Search Console need both versions?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Not if you add your site as a Domain property: it covers www, non-www, http and https in one place. A URL-prefix property only sees the exact version you typed, which is a common reason the numbers look too low.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;My site loads on both versions and nothing bad has happened. Should I still fix it?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Yes, because it costs ten minutes and removes a guess. Nothing bad happening is not the same as nothing being split.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Is this worth fixing before I spend on ads?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
It is one of the first things, along with the rest of &lt;a href="https://contento.solutions/blog/fix-your-website-before-you-spend-on-ads" rel="noopener noreferrer"&gt;what to fix before paying for traffic&lt;/a&gt;. Paid clicks should land on one address you can measure.&lt;/p&gt;

&lt;h2&gt;
  
  
  What I can and can't claim
&lt;/h2&gt;

&lt;p&gt;The numbers come from 54 private therapy practice websites in ten US metros, checked in September 2026 by requesting all four versions of each domain and recording every redirect in the chain. Across 345 sites in six samples, including five in Latin America, the failure rate was similar: 60 of 345. What I can't tell you is how much a split site costs in rankings, because the honest answer depends on the site and nobody has a clean number for it.&lt;/p&gt;

&lt;p&gt;If you'd rather I run the four addresses and the rest of the technical basics on your site, that is part of &lt;a href="https://contento.solutions/services" rel="noopener noreferrer"&gt;the audit&lt;/a&gt;, or you can &lt;a href="https://contento.solutions/book-a-call" rel="noopener noreferrer"&gt;book a strategic fit call&lt;/a&gt; and we'll look at it together.&lt;/p&gt;

&lt;p&gt;Originally published at &lt;a href="https://contento.solutions/blog/www-or-non-www-does-it-matter" rel="noopener noreferrer"&gt;contento.solutions&lt;/a&gt;.&lt;/p&gt;

</description>
      <category>seo</category>
      <category>webdev</category>
      <category>beginners</category>
      <category>smallbusiness</category>
    </item>
    <item>
      <title>Google Business Profile Suspended? What to Check Before You Appeal</title>
      <dc:creator>Lucas Fernández Díaz</dc:creator>
      <pubDate>Wed, 04 Nov 2026 13:00:00 +0000</pubDate>
      <link>https://dev.to/lucasfernandezdiaz/google-business-profile-suspended-what-to-check-before-you-appeal-7jj</link>
      <guid>https://dev.to/lucasfernandezdiaz/google-business-profile-suspended-what-to-check-before-you-appeal-7jj</guid>
      <description>&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://contento.solutions/academy/google-business-profile-suspended-appeal" rel="noopener noreferrer"&gt;Contento Solutions&lt;/a&gt;, where I write about SEO and AI visibility for service businesses.&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;If your Google Business Profile is suspended or disabled, the appeal goes through Google's appeals tool, with documents whose name and address match the profile. Fix what broke the guidelines before you submit, and don't create a new profile while the appeal is open. Nobody outside Google can promise a reinstatement.&lt;/p&gt;

&lt;p&gt;The panic is always the same. The phone stops ringing, the pin is gone from Maps, and the email from Google names a policy in one line.&lt;/p&gt;

&lt;p&gt;I went through 26 public questions about this from lawyers, dentists, veterinary clinics and a lawn care business. Eight turned out to be about something else or too thin to count: fake reviews, a category change, a disapproved ad, a worried question about stock photos, and a one-line title about a duplicate profile. The other 18 are the real thing. Seventeen suspended or disabled profiles, and one listing taken over by someone else.&lt;/p&gt;

&lt;p&gt;Two patterns stood out. In eight of those 18, the owner had already been turned down at least once. In six, the trouble started around an address: a move, a new office, or a suite shared with other businesses. One attorney wrote that a second appeal was denied "with only a general reference to" the guidelines. That's the part people find hardest. The rejection rarely says what to fix.&lt;/p&gt;

&lt;h2&gt;
  
  
  What Google actually says about suspensions
&lt;/h2&gt;

&lt;p&gt;Google's help page on &lt;a href="https://support.google.com/business/answer/4569145?hl=en" rel="noopener noreferrer"&gt;suspended or disabled profiles&lt;/a&gt; is short. It says Google may suspend or disable profiles that don't follow its guidelines, and that an account can be restricted too, in which case every profile it manages is suspended and it can't create or claim new ones.&lt;/p&gt;

&lt;p&gt;The appeals tool shows three things for each restricted profile: the profile, the reason for the action, and a link to the policy it broke. Read that link before you write anything. It's the only direct clue you'll get.&lt;/p&gt;

&lt;p&gt;The rules themselves live in the &lt;a href="https://support.google.com/business/answer/3038177?hl=en" rel="noopener noreferrer"&gt;guidelines for representing your business&lt;/a&gt; and in the &lt;a href="https://support.google.com/business/answer/13763036?hl=en" rel="noopener noreferrer"&gt;eligibility and ownership guidelines&lt;/a&gt;. The second one sets the base condition: the business has to make in-person contact with customers during its stated hours.&lt;/p&gt;

&lt;h2&gt;
  
  
  The usual causes, and what to check for each
&lt;/h2&gt;

&lt;p&gt;The left column comes from the threads I read. The rest comes from Google's own guidelines and its list of evidence.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Frequent cause&lt;/th&gt;
&lt;th&gt;What tends to set it off&lt;/th&gt;
&lt;th&gt;What to check before appealing&lt;/th&gt;
&lt;th&gt;Evidence that helps&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;An address Google doesn't accept&lt;/td&gt;
&lt;td&gt;Moving into a shared suite, law chambers or a co-working space&lt;/td&gt;
&lt;td&gt;Google allows a co-working office only with clear signage, customers received during business hours, and your own staff there during those hours. Virtual offices and P.O. boxes aren't eligible&lt;/td&gt;
&lt;td&gt;A business license or utility bill showing that exact address&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;A relocation&lt;/td&gt;
&lt;td&gt;Editing the address on a live profile, sometimes right after a move&lt;/td&gt;
&lt;td&gt;That the new address meets the same rules and that your website shows it&lt;/td&gt;
&lt;td&gt;Business registration or a utility bill at the new address&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Extra words in the name&lt;/td&gt;
&lt;td&gt;Adding a city, a service or a slogan to the business name&lt;/td&gt;
&lt;td&gt;That the name matches your sign, website and stationery, word for word&lt;/td&gt;
&lt;td&gt;Registration or license in that exact name&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;A visible address that should be hidden&lt;/td&gt;
&lt;td&gt;A home-based business that serves customers at their location&lt;/td&gt;
&lt;td&gt;Google says a service-area business should hide its address and keep its area within about two hours of driving&lt;/td&gt;
&lt;td&gt;Registration and license, with the address you keep on file&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;More than one profile for the same place&lt;/td&gt;
&lt;td&gt;A practitioner profile next to the practice's, or a second listing for one location&lt;/td&gt;
&lt;td&gt;One profile per location, and Google's rules for solo practitioners at a branded practice&lt;/td&gt;
&lt;td&gt;Registration showing which entity operates there&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;The evidence column isn't my invention. Google lists what can strengthen an appeal: an official business registration, a business license, tax certificates, and utility bills such as electricity, phone, water or internet. It also says to check that the name and address on each document match the profile.&lt;/p&gt;

&lt;h2&gt;
  
  
  First, read the reason in the appeals tool
&lt;/h2&gt;

&lt;p&gt;Open the &lt;a href="https://support.google.com/business/?p=manage_appeals" rel="noopener noreferrer"&gt;Google Business Profile appeals tool&lt;/a&gt;, signed in with the account that manages the profile, and select it. Write down the reason and open the linked policy. If the status says it can't be appealed, stop there and read the rest of this page before you do anything else.&lt;/p&gt;

&lt;h2&gt;
  
  
  Second, check the address against the guidelines
&lt;/h2&gt;

&lt;p&gt;Compare where you actually receive clients with what the profile says. A shared office needs your own sign, your own staff and clients received there during your hours. If you work from home and go to your clients, Google says to hide the address and show a service area instead.&lt;/p&gt;

&lt;h2&gt;
  
  
  Third, check the name and how the profile is set up
&lt;/h2&gt;

&lt;p&gt;Your business name on the profile should be the one on your door, your website and your letterhead, with nothing added. Google names taglines, locations and service descriptions among the things that don't belong there. Doctors, dentists and lawyers count as individual practitioners, and Google has specific rules for when they get their own profile and when a solo practitioner should share one with the practice.&lt;/p&gt;

&lt;h2&gt;
  
  
  Fourth, gather documents that match the profile exactly
&lt;/h2&gt;

&lt;p&gt;Pick from Google's list: registration, license, tax certificate, utility bill. Then check each one. Same legal or trading name as the profile, same street address, same suite number. A valid license that still shows your old address doesn't match the profile you're defending. So does a document in a different legal name, which is the mismatch one agency in the threads was still fighting.&lt;/p&gt;

&lt;h2&gt;
  
  
  Fifth, fix the profile, then submit once with the evidence
&lt;/h2&gt;

&lt;p&gt;Correct whatever the guidelines flagged, then submit the appeal from the tool. When you open the evidence form, you have 60 minutes to submit it, or it won't be attached, so have the files ready before you click. And don't create a new profile for the same business while the appeal is under review. Google says so in plain words.&lt;/p&gt;

&lt;h2&gt;
  
  
  Sixth, if it's denied, ask for the additional review with new evidence
&lt;/h2&gt;

&lt;p&gt;Google says that only if a reinstatement request is denied, it may be able to do an &lt;a href="https://support.google.com/business/contact/local_appeals" rel="noopener noreferrer"&gt;additional review&lt;/a&gt;, and that you can add evidence you didn't include the first time. One lawyer in the sample sent the same documents a second time and got the same answer. Change what you send, or change the profile first.&lt;/p&gt;

&lt;h2&gt;
  
  
  When the problem is video verification
&lt;/h2&gt;

&lt;p&gt;One dental practice in the sample was suspended for misrepresentation right after verifying by video. Google's &lt;a href="https://support.google.com/business/answer/14271705?hl=en" rel="noopener noreferrer"&gt;video verification page&lt;/a&gt; asks for a live, unedited recording with no breaks, at least 30 seconds long, recorded and uploaded from a mobile device through the profile.&lt;/p&gt;

&lt;p&gt;For a storefront it wants the location, signage that matches the profile name, and proof you manage the place, like opening an employee-only area. For a service-area business, tools, equipment or a branded vehicle. Keep tax IDs, bank details and other people's faces out of the frame. Review can take up to five business days, and a video that falls short can be recorded again with what it missed.&lt;/p&gt;

&lt;h2&gt;
  
  
  When someone else took over the listing
&lt;/h2&gt;

&lt;p&gt;This isn't a suspension, and the path is different.&lt;/p&gt;

&lt;p&gt;If you still have access, open People and access in the profile settings. Google's page on &lt;a href="https://support.google.com/business/answer/3403100?hl=en" rel="noopener noreferrer"&gt;owners and managers&lt;/a&gt; says only owners can remove other owners and managers. A manager can't. If you can't select someone for removal, your own role is probably manager.&lt;/p&gt;

&lt;p&gt;If you've lost access, Google's &lt;a href="https://support.google.com/business/answer/4566671?hl=en" rel="noopener noreferrer"&gt;ownership request process&lt;/a&gt; starts at business.google.com/add with Request access. The current owner gets an email and has 3 days to respond. If they deny it, you can suggest an edit or try to appeal. If they don't answer, you may get the option to claim the profile, though Google says that option isn't always there.&lt;/p&gt;

&lt;p&gt;If the listing shows a fake phone number, name or website, Google Maps points to a &lt;a href="https://support.google.com/maps/answer/16109801?hl=en" rel="noopener noreferrer"&gt;Business Redressal Complaint form&lt;/a&gt; for suspected fraud. The ownership rules help you here too: Google says an authorized representative must never claim a profile without the owner's express consent.&lt;/p&gt;

&lt;h2&gt;
  
  
  What nobody can promise you
&lt;/h2&gt;

&lt;p&gt;Nobody outside Google can guarantee a reinstatement. Not an agency, not a forum expert, not me. What someone outside can do is read the guidelines against your real setup, catch the mismatch you stopped seeing, and make sure the documents say the same thing as the profile. That improves the odds of a clean appeal. The decision stays with Google.&lt;/p&gt;

&lt;p&gt;If you're hiring help for this, the questions in &lt;a href="https://contento.solutions/blog/how-to-tell-if-your-seo-consultant-is-working" rel="noopener noreferrer"&gt;how to tell if your SEO consultant is working&lt;/a&gt; apply here too. And once the profile is back, your website, profile and documents should describe the same business in the same words. That's what &lt;a href="https://contento.solutions/academy/entity-map-small-brands-content-architecture" rel="noopener noreferrer"&gt;an entity map for a small brand&lt;/a&gt; is for. If calls dropped and you aren't sure the profile is the only cause, &lt;a href="https://contento.solutions/academy/when-seo-traffic-falls-diagnose-demand-indexing-conversion" rel="noopener noreferrer"&gt;diagnosing a traffic drop&lt;/a&gt; walks through the other suspects.&lt;/p&gt;

&lt;h2&gt;
  
  
  Questions I get about this
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;How do I appeal a suspended Google Business Profile?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Open the Google Business Profile appeals tool while signed in to the managing account, select the profile, read the reason and the linked policy, and select Submit Appeal. You can attach evidence such as a business license or utility bill, and Google replies by email.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Should I create a new profile while the appeal is pending?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
No. Google's help page says not to create a new Business Profile for the same business while the appeal is under review.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;What documents does Google accept as evidence?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Google lists an official business registration, a business license, tax certificates and utility bills such as electricity, phone, water or internet. The name and address on each one should match the profile you're appealing for.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Can I list my practice at a shared office or co-working space?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Only if that office has clear signage, receives customers during business hours and is staffed by your own people during those hours. Google says virtual offices aren't eligible.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;My appeal was denied. Is that the end?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Not always. Google says that if a reinstatement request is denied, it may be able to do an additional review, and you can add evidence you didn't send the first time. Some profiles show a status that says they can't be appealed.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Can someone guarantee my profile will be reinstated?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
No. The decision belongs to Google. Anyone outside it can only help you fix the profile and send better evidence.&lt;/p&gt;

&lt;p&gt;If you have read the reason and still cannot see what broke the guidelines, that is a reasonable thing to bring to a call. What the practice does and what gets reported is on the &lt;a href="https://contento.solutions/services" rel="noopener noreferrer"&gt;services page&lt;/a&gt;, with more detail on the &lt;a href="https://contento.solutions/faq" rel="noopener noreferrer"&gt;FAQ page&lt;/a&gt;. The call is &lt;a href="https://contento.solutions/book-a-call" rel="noopener noreferrer"&gt;thirty minutes&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;If you have been through a suspension, what did the rejection say, and what finally worked? Tell me in the comments. The rejection emails are the part nobody publishes.&lt;/p&gt;

</description>
      <category>seo</category>
      <category>google</category>
      <category>marketing</category>
      <category>smallbusiness</category>
    </item>
    <item>
      <title>Your robots.txt Allows AI Crawlers. Your Server May Not.</title>
      <dc:creator>Lucas Fernández Díaz</dc:creator>
      <pubDate>Wed, 28 Oct 2026 13:00:00 +0000</pubDate>
      <link>https://dev.to/lucasfernandezdiaz/your-robotstxt-allows-ai-crawlers-your-server-may-not-3dmg</link>
      <guid>https://dev.to/lucasfernandezdiaz/your-robotstxt-allows-ai-crawlers-your-server-may-not-3dmg</guid>
      <description>&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://contento.solutions/academy/the-block-nobody-declared" rel="noopener noreferrer"&gt;Contento Solutions&lt;/a&gt;, where I write about SEO and AI visibility for service businesses.&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;Nine of 18 agency sites turned an AI crawler away. None of the nine said so in robots.txt. I tested 12 AI crawlers against 18 marketing agencies that sell search visibility on 16 September 2026, and every refusal came from the server or the CDN, the layer most audits never check.&lt;/p&gt;

&lt;p&gt;A week earlier the count was one of nine. One site declared its block in robots.txt, and its server ignored the declaration. Eight days later that site had swapped sides: the file welcomed everyone and the server did the refusing. As far as I can tell, nobody at the agency touched a thing.&lt;/p&gt;

&lt;p&gt;That swap is the reason this piece exists.&lt;/p&gt;

&lt;h2&gt;
  
  
  Two layers, and why almost everyone reads only one
&lt;/h2&gt;

&lt;p&gt;The question I asked about each site is narrow on purpose. Not whether an assistant recommends it. Whether an AI crawler can fetch its home page at all.&lt;/p&gt;

&lt;p&gt;The first layer is robots.txt, the public text file at the root of a domain that tells crawlers which paths they may request. The &lt;a href="https://www.rfc-editor.org/rfc/rfc9309.html" rel="noopener noreferrer"&gt;standard that defines it&lt;/a&gt; says plainly that its rules aren't a form of access authorization. It's a request, it's free to read, and it's where nearly every audit stops.&lt;/p&gt;

&lt;p&gt;The second layer is the response. A server can hand your browser the page and refuse the same page to GPTBot with a 403, before any of the site's own code runs. Nothing about that shows up in robots.txt. It comes from a CDN setting, a security plugin or a host firewall.&lt;/p&gt;

&lt;p&gt;So for each crawler I checked both: the robots.txt rule that applies to it, and a real request for the home page carrying its user agent. Two of the 12, Google-Extended and Applebot-Extended, have no user agent of their own, so for those two robots.txt is the only signal there is.&lt;/p&gt;

&lt;h2&gt;
  
  
  The count on 16 September 2026
&lt;/h2&gt;

&lt;p&gt;I left out one of the 19 domains I track because it refused even a desktop browser. That leaves 18.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;What I measured&lt;/th&gt;
&lt;th&gt;8 September&lt;/th&gt;
&lt;th&gt;16 September&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Agency sites measured&lt;/td&gt;
&lt;td&gt;18&lt;/td&gt;
&lt;td&gt;18&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Turn away at least one AI crawler&lt;/td&gt;
&lt;td&gt;9&lt;/td&gt;
&lt;td&gt;9&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Of those, declare it in robots.txt&lt;/td&gt;
&lt;td&gt;1&lt;/td&gt;
&lt;td&gt;0&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Of those, only turn away Bytespider&lt;/td&gt;
&lt;td&gt;5&lt;/td&gt;
&lt;td&gt;5&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Of those, turn away a crawler from OpenAI, Anthropic or Perplexity&lt;/td&gt;
&lt;td&gt;3&lt;/td&gt;
&lt;td&gt;3&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Of the 18, bounce every bot, so the AI tests prove nothing&lt;/td&gt;
&lt;td&gt;1&lt;/td&gt;
&lt;td&gt;1&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Publish a real llms.txt file&lt;/td&gt;
&lt;td&gt;not comparable&lt;/td&gt;
&lt;td&gt;10&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;The honest reading starts with the fourth row. Five of the nine only refuse Bytespider, the crawler from ByteDance, and that block is common enough to be close to a default. My own site refuses it too. The ninth site refuses CCBot and Meta's crawler. The three that turn away crawlers from OpenAI, Anthropic or Perplexity are the ones that matter most for assistants, and none of them says so in its file.&lt;/p&gt;

&lt;p&gt;The llms.txt figure isn't comparable because I fixed a flaw in my own script in between. On 8 September it counted a page that answered with HTML as if it were a real llms.txt file.&lt;/p&gt;

&lt;h2&gt;
  
  
  The case that changed in eight days
&lt;/h2&gt;

&lt;p&gt;On 8 September, one agency's robots.txt opened with a block marked "BEGIN Cloudflare Managed content". It disallowed GPTBot, ClaudeBot, CCBot, Bytespider, Google-Extended, Applebot-Extended and meta-externalagent. Under it sat the site's own rule, which lets every crawler in. The server let them in too.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Same site, home page&lt;/th&gt;
&lt;th&gt;8 September&lt;/th&gt;
&lt;th&gt;16 September&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;What robots.txt says&lt;/td&gt;
&lt;td&gt;A Cloudflare block that disallows seven AI crawlers, then "allow everyone"&lt;/td&gt;
&lt;td&gt;Only "allow everyone" and the sitemap&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;GPTBot, ClaudeBot, CCBot, Bytespider&lt;/td&gt;
&lt;td&gt;200&lt;/td&gt;
&lt;td&gt;403&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;OAI-SearchBot, ChatGPT-User, Claude-SearchBot, PerplexityBot, Perplexity-User&lt;/td&gt;
&lt;td&gt;200&lt;/td&gt;
&lt;td&gt;200&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;A made-up bot, as a control&lt;/td&gt;
&lt;td&gt;200&lt;/td&gt;
&lt;td&gt;200&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;On 16 September the refusal was a 25-byte line of plain text, "Your request was blocked." A separate check with Amazon's crawler that night got the same answer. The search and assistant crawlers still got the full page.&lt;/p&gt;

&lt;p&gt;Look at what flipped. On the 8th the file said no and the door said yes. On the 16th the file said yes and the door said no. Anyone who audited this site by reading robots.txt got the wrong answer both times.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why a provider, and not the agency, probably did it
&lt;/h2&gt;

&lt;p&gt;On 15 September, Cloudflare &lt;a href="https://blog.cloudflare.com/accountable-mixed-use-ai-crawlers/" rel="noopener noreferrer"&gt;announced&lt;/a&gt; it was retiring two features: "Block AI Bots" and the managed robots.txt that wrote blocks like the one above into the file. AI traffic now runs through separate Search, Training and Agent controls per domain. Training got a new option, Disallow AI Training, which Cloudflare describes as keeping Googlebot, Bingbot and Applebot crawling for search while it blocks crawlers that only train, including those from Amazon, Anthropic, Meta and OpenAI. Cloudflare also says existing settings carry over and most customers don't need to do anything.&lt;/p&gt;

&lt;p&gt;What I saw at the edge fits that option closely. Training crawlers got a 403, search and assistant crawlers got in. There's a second clue. The domain Cloudflare uses as the example in its managed robots.txt documentation had lost its managed block by 16 September too, and it refused GPTBot with the same 25-byte line.&lt;/p&gt;

&lt;p&gt;This is an inference, and I'd rather label it than oversell it. From outside I can't see the site's settings, who changed what, or when. One piece doesn't fit, either. Disallow AI Training is supposed to publish a no-training line in robots.txt, and neither file had one on the 16th.&lt;/p&gt;

&lt;p&gt;Still, it changes the story. When I first read this site, the point was that a vendor had written the block, not the owner. Now the vendor has rewritten it, moved it to a place the owner can't see from the file, and told customers they didn't need to act.&lt;/p&gt;

&lt;h2&gt;
  
  
  The case with no provider in sight
&lt;/h2&gt;

&lt;p&gt;A second agency publishes the most permissive robots.txt there is, generated by an SEO plugin: one rule for every crawler, with an empty disallow. On both dates its server answered a desktop browser and my made-up bot with a 200, and answered seven named crawlers with a 403: GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-SearchBot, PerplexityBot and Perplexity-User.&lt;/p&gt;

&lt;p&gt;This one is more serious than the first case. It shuts out the crawlers that fetch a page while an assistant answers a live question, not only the training ones. And its response headers point to a plain nginx server, not Cloudflare, so the provider story above doesn't cover it. The rule could live in a firewall, a plugin or a host preset. I can't tell which from outside.&lt;/p&gt;

&lt;p&gt;Its three other crawlers, CCBot, Bytespider and meta-externalagent, got a 429 both times. I didn't count those as blocks.&lt;/p&gt;

&lt;h2&gt;
  
  
  What this does not prove
&lt;/h2&gt;

&lt;p&gt;A 403 to a crawler I impersonated isn't proof that the real crawler gets a 403. My requests come from a home connection in Argentina, not from OpenAI's or Anthropic's addresses, and a CDN can treat a verified crawler differently. That cuts both ways, so I only report what the server said to the user agent I sent.&lt;/p&gt;

&lt;p&gt;It doesn't prove anyone chose the block. In the first case the evidence points at a provider migration. In the second I can't see the cause at all.&lt;/p&gt;

&lt;p&gt;It doesn't cover the whole market. The 18 sites are agencies I track, picked because they compete for the same kind of work I do. They aren't a random sample.&lt;/p&gt;

&lt;p&gt;And being readable isn't being cited. An open door doesn't make a site the answer, the same way being indexed never made a page rank. Access is the first of the three questions on &lt;a href="https://contento.solutions/method" rel="noopener noreferrer"&gt;my method page&lt;/a&gt;. Whether a machine can quote you once it's inside is a different job, covered in &lt;a href="https://contento.solutions/academy/make-website-citable-ai-search" rel="noopener noreferrer"&gt;making a site citable&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  My own site, measured the same way
&lt;/h2&gt;

&lt;p&gt;On 16 September, 11 of the 12 crawlers could read Contento. The twelfth was Bytespider, refused by my platform at the server, with nothing in my robots.txt. I didn't choose that. I know about it because I measure it, and that's the only difference I'd claim.&lt;/p&gt;

&lt;p&gt;If you want the numbers for a whole niche rather than one site, I ran the same test on &lt;a href="https://contento.solutions/academy/therapy-websites-ai-visibility-index" rel="noopener noreferrer"&gt;54 therapy practice websites&lt;/a&gt;, and &lt;a href="https://contento.solutions/academy/how-to-build-an-ai-visibility-index" rel="noopener noreferrer"&gt;the method behind that index&lt;/a&gt; is written so you can repeat it.&lt;/p&gt;

&lt;h2&gt;
  
  
  How to check your own site
&lt;/h2&gt;

&lt;p&gt;You don't need a tool or a vendor. A terminal, ten minutes and some patience will do. Six steps, and the ones people skip are the second, the third and the sixth.&lt;/p&gt;

&lt;h2&gt;
  
  
  First, read your robots.txt, including what's missing
&lt;/h2&gt;

&lt;p&gt;Open yourdomain.com/robots.txt and read it top to bottom. Look for any block you didn't write, especially one wrapped in comments naming a vendor. If you're on Cloudflare and a managed block used to sit at the top, it may be gone since 15 September. A clean file doesn't mean crawlers get in. It means the file no longer tells you.&lt;/p&gt;

&lt;h2&gt;
  
  
  Second, get a baseline as a browser
&lt;/h2&gt;

&lt;p&gt;Request your home page with a normal desktop browser user agent and note the status code and the size of the page. Every result after this is a comparison against it. Without a baseline, a 403 or a short page means nothing.&lt;/p&gt;

&lt;h2&gt;
  
  
  Third, run the control with a bot that has nothing to do with AI
&lt;/h2&gt;

&lt;p&gt;Request the same page with a made-up bot name. If that gets a 403 too, your server refuses every bot, and the AI tests can't tell you anything until that changes. One site in my sample did exactly this, and I didn't count it among the nine.&lt;/p&gt;

&lt;h2&gt;
  
  
  Fourth, ask as each AI crawler, with the full user agent
&lt;/h2&gt;

&lt;p&gt;Request the page once per crawler, a few seconds apart, with each crawler's full user agent string and not just its name. On the first site in this piece, the bare word "GPTBot" got the page and the full string got the 403. A 403 counts only when the browser and the made-up bot both got a 200. A 200 with a fraction of the browser's bytes is a soft block, so write that down too.&lt;/p&gt;

&lt;p&gt;From a terminal, the GPTBot request looks like this:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;curl &lt;span class="nt"&gt;-s&lt;/span&gt; &lt;span class="nt"&gt;-o&lt;/span&gt; /dev/null &lt;span class="nt"&gt;-w&lt;/span&gt; &lt;span class="s2"&gt;"%{http_code} %{size_download}"&lt;/span&gt; &lt;span class="nt"&gt;-A&lt;/span&gt; &lt;span class="s2"&gt;"Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; GPTBot/1.1; +https://openai.com/gptbot"&lt;/span&gt; https://yourdomain.com/
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h2&gt;
  
  
  Fifth, treat every 429 as your own fault
&lt;/h2&gt;

&lt;p&gt;A 429 is a rate limit. It says how fast you asked, not what the site allows. Wait, retry once more slowly, and only record the result if it holds. Two agencies in my first run went from "blocks OpenAI and Anthropic" to open once I slowed down.&lt;/p&gt;

&lt;h2&gt;
  
  
  Sixth, if you're on Cloudflare, read the Training setting
&lt;/h2&gt;

&lt;p&gt;Open the domain's security settings and find the Search, Training and Agent controls. Disallow AI Training is meant to keep you in Google, Bing and Apple search. Block, since 15 September, reaches Googlebot, Bingbot and Applebot as well. If you picked Block on purpose, confirm in Search Console or your server logs that Google still crawls you. Pretending to be Googlebot from your laptop won't answer that.&lt;/p&gt;

&lt;h2&gt;
  
  
  What to do with what you find
&lt;/h2&gt;

&lt;p&gt;If you find a block you didn't put there, the fix is a setting, not a project. Usually it's in the CDN dashboard, a security plugin's bad bot list or the host's firewall.&lt;/p&gt;

&lt;p&gt;Decide on purpose before you flip it. Refusing training crawlers while letting search and assistant crawlers in is a coherent position. Refusing everything is coherent too, if your content is the product. Paying for AI visibility while a setting you never read does the opposite isn't.&lt;/p&gt;

&lt;h2&gt;
  
  
  Questions I get about this
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;If my robots.txt allows everything, can AI crawlers read my site?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Not necessarily. On 16 September 2026, all nine agency sites in my sample that turned an AI crawler away had a robots.txt that allowed it. The file tells you what a site asks for. The server's response tells you what it does.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Does blocking GPTBot remove me from ChatGPT?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
No. GPTBot is OpenAI's training crawler. The crawlers that fetch pages for ChatGPT search and for users who ask it to open a link are different ones, with their own names. Blocking one doesn't block the others.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Did Cloudflare change how AI crawlers are blocked in September 2026?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Yes. On 15 September 2026 Cloudflare announced it was retiring Block AI Bots and its managed robots.txt, and added a Disallow AI Training option to its Training control. On one site I track, the robots.txt block disappeared and the server started refusing training crawlers in the same week.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Is a 403 to a crawler always a decision about AI?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
No. Many sites refuse anything that isn't a browser. Unless a made-up bot unrelated to AI gets a 200 from the same page, a 403 to an AI crawler proves nothing about AI.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;How often should I check?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Whenever your CDN, host or security plugin changes a plan or ships an update, and otherwise once a quarter. The block in the first case moved without anyone at the agency typing a rule, on the provider's schedule.&lt;/p&gt;

&lt;p&gt;If you would rather read your server's answers with someone, that is a reasonable thing to bring to a call. What the practice does and what gets reported is on the &lt;a href="https://contento.solutions/services" rel="noopener noreferrer"&gt;services page&lt;/a&gt;, with more detail on the &lt;a href="https://contento.solutions/faq" rel="noopener noreferrer"&gt;FAQ page&lt;/a&gt;. The call is &lt;a href="https://contento.solutions/book-a-call" rel="noopener noreferrer"&gt;thirty minutes&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;If your site sits behind a CDN, run the six steps and tell me in the comments whether your robots.txt and your server agree. I would like to know how common the mismatch is outside the agencies I track.&lt;/p&gt;

</description>
      <category>seo</category>
      <category>ai</category>
      <category>webdev</category>
      <category>cloudflare</category>
    </item>
    <item>
      <title>Your Website Is Built to Be Contacted, Not Chosen</title>
      <dc:creator>Lucas Fernández Díaz</dc:creator>
      <pubDate>Wed, 21 Oct 2026 13:00:00 +0000</pubDate>
      <link>https://dev.to/lucasfernandezdiaz/your-website-is-built-to-be-contacted-not-chosen-3dim</link>
      <guid>https://dev.to/lucasfernandezdiaz/your-website-is-built-to-be-contacted-not-chosen-3dim</guid>
      <description>&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://contento.solutions/blog/your-site-is-built-to-be-contacted" rel="noopener noreferrer"&gt;Contento Solutions&lt;/a&gt;, where I write about SEO and AI visibility for service businesses.&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;Most small sites make contacting you easy and choosing you hard. Across the 54 sites we measured, the gap between those two things ran fifteen points. That's a website doing the second half of the job well and the first half barely at all.&lt;/p&gt;

&lt;h2&gt;
  
  
  Eleven at night, three tabs open
&lt;/h2&gt;

&lt;p&gt;Picture the person you actually want. They have a problem, a budget they haven't said out loud, and three tabs open with three businesses that look roughly the same. They are not going to contact all three. They are going to close two.&lt;/p&gt;

&lt;p&gt;Now open your own site and try to be them. What do you learn in ninety seconds that helps you close the other two tabs? For most sites the answer is: what the business does, that it cares, and how to get in touch. All three of those are also true of the competitors.&lt;/p&gt;

&lt;h2&gt;
  
  
  The fifteen point gap
&lt;/h2&gt;

&lt;p&gt;We scored those 54 sites on two things that sound similar and aren't.&lt;/p&gt;

&lt;p&gt;Friction to contact, meaning how hard it is to actually reach you once someone has decided: 75.6 out of 100. Decision served, meaning whether the page gives someone enough to decide at all: 60.5.&lt;/p&gt;

&lt;p&gt;Nearly everyone had the contact form working. Far fewer had anything that helps a stranger choose. We have spent about fifteen years as an industry optimising the click, on pages that never earned it.&lt;/p&gt;

&lt;h2&gt;
  
  
  What a decision actually needs
&lt;/h2&gt;

&lt;p&gt;Here is what people are trying to settle at eleven at night, and what most pages offer instead.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;What they need to settle&lt;/th&gt;
&lt;th&gt;What the page usually says&lt;/th&gt;
&lt;th&gt;What settles it&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Can I afford this&lt;/td&gt;
&lt;td&gt;"Contact us for a quote"&lt;/td&gt;
&lt;td&gt;A number, a range, or an honest starting point&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Is this for someone like me&lt;/td&gt;
&lt;td&gt;"We work with businesses of all sizes"&lt;/td&gt;
&lt;td&gt;Who you're a bad fit for, in your own words&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;What happens if I say yes&lt;/td&gt;
&lt;td&gt;"Get in touch to learn more"&lt;/td&gt;
&lt;td&gt;The first three steps and roughly how long they take&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;What if it doesn't work&lt;/td&gt;
&lt;td&gt;Nothing at all&lt;/td&gt;
&lt;td&gt;What you do about it, stated before they ask&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Look at the middle column. Every one of those lines is written to avoid losing anybody. That's the problem. A page that could apply to anyone helps nobody choose, and the person closing tabs is choosing whether they can tell or not.&lt;/p&gt;

&lt;h2&gt;
  
  
  The price question, honestly
&lt;/h2&gt;

&lt;p&gt;Plenty of businesses can't publish a flat price, and pretending otherwise is bad advice. Scope varies, and a wrong number costs you more than no number.&lt;/p&gt;

&lt;p&gt;But "it depends" and silence are not the same thing. You can publish what drives the price up or down. You can publish a floor, the smallest sensible engagement, and say plainly that most work costs more. You can publish what you charged for a real recent project with the client's name removed. All three let someone rule themselves in or out, which is the actual job of the page.&lt;/p&gt;

&lt;p&gt;We publish our own numbers for the same reason, and we assume it costs us some enquiries. We can't prove that either way yet, which is the honest version. The bet is that the enquiries a price scares off were heading for a bad fourth call.&lt;/p&gt;

&lt;h2&gt;
  
  
  The bad fit paragraph
&lt;/h2&gt;

&lt;p&gt;If you only change one thing, write the paragraph that says who shouldn't hire you.&lt;/p&gt;

&lt;p&gt;Two or three sentences. Not false modesty, not a humblebrag about being too thorough. The real thing: the budget you can't work under, the timeline you can't meet, the kind of project you turn down. It costs an afternoon, and it's the fastest way to stop being interchangeable with the other two tabs.&lt;/p&gt;

&lt;p&gt;It also does something for the machines. An assistant comparing options needs contrast to recommend anyone, and a page that refuses to distinguish itself gives it nothing to work with. That's the same underlying issue as &lt;a href="https://contento.solutions/blog/machine-cannot-confirm-you-are-real" rel="noopener noreferrer"&gt;not being confirmable&lt;/a&gt;, arriving from a different direction.&lt;/p&gt;

&lt;h2&gt;
  
  
  Questions we get
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Won't publishing prices let competitors undercut me?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
They can already call and ask, and most have. The people your silence actually stops are the buyers.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;My work genuinely varies. What number do I put?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
The floor, plus the two or three things that move it. "From X, and it goes up with Y and Z" is a real answer and rules out the wrong people just as well as a fixed price.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Isn't a bad fit paragraph just turning away business?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
It turns away the business you were going to turn away anyway, earlier and for free, instead of after two calls and a proposal.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Where does this go on the site?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
On the page that sells the service, not on a separate pricing page nobody reaches. If you have a pricing page, it should repeat rather than replace.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;How do I know if my page has this problem?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Give it to someone outside your industry and ask them one question: would you hire this business, and why not. The "why not" is the part worth writing down.&lt;/p&gt;

&lt;h2&gt;
  
  
  What we can and can't claim
&lt;/h2&gt;

&lt;p&gt;Those two scores come from 54 private therapy practices in ten US metros, measured on one day in September 2026, with our own instrument. We sell to that market. The four sites we couldn't read are named in &lt;a href="https://contento.solutions/academy/therapy-websites-ai-visibility-index" rel="noopener noreferrer"&gt;the published index&lt;/a&gt; instead of dropped, and &lt;a href="https://contento.solutions/academy/how-to-build-an-ai-visibility-index" rel="noopener noreferrer"&gt;the method is written down&lt;/a&gt; so you can run it on your own niche and disagree with us.&lt;/p&gt;

&lt;p&gt;What we can't tell you is whether fixing this raises your revenue, because we haven't measured that and nobody has shown us a study that has.&lt;/p&gt;

&lt;p&gt;If you want a second pair of eyes on the four questions above, that is a reasonable thing to bring to a call. What the practice does and what gets reported is on the &lt;a href="https://contento.solutions/services" rel="noopener noreferrer"&gt;services page&lt;/a&gt;, with more detail on the &lt;a href="https://contento.solutions/faq" rel="noopener noreferrer"&gt;FAQ page&lt;/a&gt;. The call is &lt;a href="https://contento.solutions/book-a-call" rel="noopener noreferrer"&gt;thirty minutes&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;If you build sites for clients, does your own site publish a price or a bad fit paragraph? Tell me in the comments what stopped you, or what happened when you did.&lt;/p&gt;

</description>
      <category>seo</category>
      <category>webdev</category>
      <category>ux</category>
      <category>marketing</category>
    </item>
    <item>
      <title>The Machine Can Find You. It Can't Confirm You're Real.</title>
      <dc:creator>Lucas Fernández Díaz</dc:creator>
      <pubDate>Wed, 14 Oct 2026 13:00:00 +0000</pubDate>
      <link>https://dev.to/lucasfernandezdiaz/the-machine-can-find-you-it-cant-confirm-youre-real-528o</link>
      <guid>https://dev.to/lucasfernandezdiaz/the-machine-can-find-you-it-cant-confirm-youre-real-528o</guid>
      <description>&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://contento.solutions/blog/machine-cannot-confirm-you-are-real" rel="noopener noreferrer"&gt;Contento Solutions&lt;/a&gt;, where I write about SEO and AI visibility for service businesses.&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;Machines can reach almost every site we measured. They can confirm barely half of them. That gap is where small businesses lose recommendations they never knew they were up for, and closing it takes filing work, not marketing work.&lt;/p&gt;

&lt;h2&gt;
  
  
  Two numbers from the same 54 sites
&lt;/h2&gt;

&lt;p&gt;We measured 54 private therapy practice websites across ten US metros. Two of the numbers sit right next to each other and tell opposite stories.&lt;/p&gt;

&lt;p&gt;On reachability, the cohort scored 97 out of 100. Robots fetch the pages, the servers answer, no wall stops them. That problem has already solved itself, and most owners still worry about it.&lt;/p&gt;

&lt;p&gt;On verifiable identity, the same cohort scored 57.8.&lt;/p&gt;

&lt;p&gt;So the machine arrives, reads the page, and then cannot establish who wrote it, whether that person is licensed, where they actually work, or whether the business on this page is the same business it saw somewhere else. It has your content. It does not have you.&lt;/p&gt;

&lt;h2&gt;
  
  
  What "confirm" actually means to a machine
&lt;/h2&gt;

&lt;p&gt;An assistant recommending a local service is doing something closer to a background check than a search. It wants the same fact to show up in more than one place, attached to a name, without contradicting itself. When it can't get that, the safe move is to recommend somebody else. You are not rejected. You are skipped.&lt;/p&gt;

&lt;p&gt;Here is what moved the score in the cohort, and what it costs to fix.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Signal&lt;/th&gt;
&lt;th&gt;What we looked for&lt;/th&gt;
&lt;th&gt;Where most sites lost points&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;A named human&lt;/td&gt;
&lt;td&gt;A real person's full name on the page that sells the service&lt;/td&gt;
&lt;td&gt;"Our team of caring professionals" and no names anywhere&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Credentials in text&lt;/td&gt;
&lt;td&gt;Licence type and number written as text, not baked into an image&lt;/td&gt;
&lt;td&gt;A scanned certificate, invisible to anything that doesn't read pictures&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;One address, everywhere&lt;/td&gt;
&lt;td&gt;The same address and phone on the site, the map listing and the directories&lt;/td&gt;
&lt;td&gt;Suite number on one, missing on another, old number on a third&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Something outside your own site&lt;/td&gt;
&lt;td&gt;At least one independent page that names the business the same way&lt;/td&gt;
&lt;td&gt;Every claim living only on the domain that makes it&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;None of that is content strategy. It's closer to making sure the paperwork agrees with itself.&lt;/p&gt;

&lt;h2&gt;
  
  
  The part that surprises people
&lt;/h2&gt;

&lt;p&gt;Consistency scored 88.6 in the same cohort, well above verifiable identity. Sites were usually consistent about the little they said. The failure wasn't contradiction. It was silence.&lt;/p&gt;

&lt;p&gt;A practice that never states who works there is perfectly consistent and completely unconfirmable. Getting a bit more specific is what moves this number, and specificity is free.&lt;/p&gt;

&lt;p&gt;You can watch this happen in about a minute. Ask an assistant to name the owner of your business and cite where it found that. If it hedges, or names the wrong person, or cites a directory you've never updated, you just watched the check fail. Try it on the practice down the road too. That comparison tends to land harder than any score we could hand you.&lt;/p&gt;

&lt;p&gt;One caution about that test: assistants sometimes fill gaps with a confident guess. A wrong name is not proof the name exists somewhere wrong. It's proof the machine had to invent one, which is the same problem wearing a friendlier face.&lt;/p&gt;

&lt;h2&gt;
  
  
  What to do this week
&lt;/h2&gt;

&lt;p&gt;Open your services page and read it as a stranger who has to write one sentence about you, with a name in it. If you can't, the machine can't either.&lt;/p&gt;

&lt;p&gt;Then check three things: your name and role in text on that page, your licence or registration written out rather than pictured, and your address matching your map listing character for character. Suite numbers are the usual culprit. So are phone numbers from two offices ago.&lt;/p&gt;

&lt;p&gt;If you want to see how this fits with everything else a machine checks, we published &lt;a href="https://contento.solutions/academy/therapy-websites-ai-visibility-index" rel="noopener noreferrer"&gt;the full index of those 54 sites&lt;/a&gt;, including the four we couldn't measure at all. And if your worry is the earlier step, whether your pages can be quoted in the first place, that's &lt;a href="https://contento.solutions/academy/make-website-citable-ai-search" rel="noopener noreferrer"&gt;a different problem with a different fix&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  Questions we get
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Does this mean I need a schema markup plugin?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
It helps, but it isn't the starting point. Structured data restates facts a machine can already read. If the fact isn't on the page in plain text, marking it up gives the machine a cleaner version of nothing.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;My credentials are on my About page. Isn't that enough?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Often not. The machine usually judges the page that sells the service, and assistants don't reliably carry facts across your site. Put the name and the licence where the offer is.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;I run a solo business and I don't want my address public. What then?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Use a service area and a business registration number instead of a street address, and keep that identical everywhere. The goal is a fact that can be checked against an outside source, not your front door.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;How long until this shows up?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
We don't know, and anyone who gives you a number is guessing. What we can say is that these are the cheapest fixes on the list, and unlike content they don't need to be redone every quarter.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Is 57.8 out of 100 bad?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
It's the average. The median came in lower, at 55, so half the cohort sits below that. The useful read isn't the grade, it's the company: this isn't a race against a standard, it's a race against the practice down the road, and most of them haven't done it either.&lt;/p&gt;

&lt;h2&gt;
  
  
  One thing worth saying out loud
&lt;/h2&gt;

&lt;p&gt;We sell to this market, so read our numbers with that in mind. We published the method so you can rerun it and argue with the result. The cohort held 58 sites and we could only measure 54, because two blocked us and two never answered. The index names those four rather than quietly dropping them.&lt;/p&gt;

&lt;p&gt;If you would rather have someone go through your own site with the four signals above, that is a reasonable thing to bring to a call. What the practice does and what gets reported is on the &lt;a href="https://contento.solutions/services" rel="noopener noreferrer"&gt;services page&lt;/a&gt;, with more detail on the &lt;a href="https://contento.solutions/faq" rel="noopener noreferrer"&gt;FAQ page&lt;/a&gt;. The call is &lt;a href="https://contento.solutions/book-a-call" rel="noopener noreferrer"&gt;thirty minutes&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;If you run the one minute test on your own business, what does the assistant say about who owns it? Tell me in the comments, wrong answers included. The wrong ones are the useful ones.&lt;/p&gt;

</description>
      <category>seo</category>
      <category>ai</category>
      <category>webdev</category>
      <category>marketing</category>
    </item>
    <item>
      <title>We Tested 56 Therapy Websites for AI Access</title>
      <dc:creator>Lucas Fernández Díaz</dc:creator>
      <pubDate>Wed, 07 Oct 2026 13:00:00 +0000</pubDate>
      <link>https://dev.to/lucasfernandezdiaz/we-tested-56-therapy-websites-for-ai-access-2mlk</link>
      <guid>https://dev.to/lucasfernandezdiaz/we-tested-56-therapy-websites-for-ai-access-2mlk</guid>
      <description>&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://contento.solutions/academy/therapy-websites-ai-access" rel="noopener noreferrer"&gt;Contento Solutions&lt;/a&gt;, where I write about SEO and AI visibility for service businesses.&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;We tested 12 AI crawlers against 56 private practice therapy websites. Eight refuse the crawlers behind ChatGPT, Claude or Perplexity, and seven of those eight never wrote that rule: a server or a CDN did. The bigger number, two thirds of the sample, we threw away, and this explains why.&lt;/p&gt;

&lt;h2&gt;
  
  
  The number we threw away, and why
&lt;/h2&gt;

&lt;p&gt;The first pass said 37 of 56 sites, two thirds of the sample, block at least one AI crawler. That headline was available, it was technically true, and it was useless.&lt;/p&gt;

&lt;p&gt;Thirty five of those 37 cases were the same crawler: a bulk collector that many hosts block by default because it requests aggressively. Blocking it is a decision about server load. It is not a decision about whether a person asking an assistant for a therapist will ever see your practice.&lt;/p&gt;

&lt;p&gt;So we split the twelve crawlers into two groups and asked two different questions. The first group is the assistants people actually use: the crawlers behind ChatGPT, Claude, Perplexity and Google's AI surfaces. Blocking those has a consequence you can describe in one sentence. The second group collects at scale for training. Blocking those is defensible and says nothing about visibility.&lt;/p&gt;

&lt;p&gt;Everything below counts the first group only.&lt;/p&gt;

&lt;h2&gt;
  
  
  The two layers, because one of them is invisible
&lt;/h2&gt;

&lt;p&gt;Two places decide whether a crawler can read your site, and only one of them is public.&lt;/p&gt;

&lt;p&gt;The first is robots.txt, a text file at the root of your domain that names which crawlers may request which paths. Anyone can read it, including you, right now. It is where every audit stops.&lt;/p&gt;

&lt;p&gt;The second is the response itself. A site can hand a page to a browser and refuse the identical page to a named crawler, with a 403, at the edge, before any of your website's own code runs. That block appears nowhere in robots.txt. It arrives from a CDN setting, a security plugin, or a managed rule somebody enabled once and never revisited.&lt;/p&gt;

&lt;p&gt;We checked both layers for each of the 12 crawlers: the longest matching rule in robots.txt, and then a real request carrying that crawler's user agent.&lt;/p&gt;

&lt;h2&gt;
  
  
  Two controls, because without them the test accuses innocent sites
&lt;/h2&gt;

&lt;p&gt;The first control is a generic bot. Before recording a block, we requested the same page with a bot user agent that has nothing to do with AI. A site that refuses every non browser is not making a decision about AI, it is making a decision about bots, and it cannot count. Two sites failed this control, so we dropped them from the sample entirely.&lt;/p&gt;

&lt;p&gt;The second control is patience. We treated every 429 as our own rate limit rather than as their policy: we waited, then asked again more slowly. Five sites produced a 429 at some point, and we counted none of them as blocking.&lt;/p&gt;

&lt;p&gt;That is why the denominator below is 56 and not 58.&lt;/p&gt;

&lt;h2&gt;
  
  
  What we found
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Assistant crawler&lt;/th&gt;
&lt;th&gt;Sites that refuse it&lt;/th&gt;
&lt;th&gt;Of&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;ClaudeBot&lt;/td&gt;
&lt;td&gt;8&lt;/td&gt;
&lt;td&gt;56&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;GPTBot&lt;/td&gt;
&lt;td&gt;7&lt;/td&gt;
&lt;td&gt;56&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;OAI-SearchBot&lt;/td&gt;
&lt;td&gt;6&lt;/td&gt;
&lt;td&gt;56&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;ChatGPT-User&lt;/td&gt;
&lt;td&gt;6&lt;/td&gt;
&lt;td&gt;56&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;PerplexityBot&lt;/td&gt;
&lt;td&gt;6&lt;/td&gt;
&lt;td&gt;56&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Claude-SearchBot&lt;/td&gt;
&lt;td&gt;5&lt;/td&gt;
&lt;td&gt;56&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Perplexity-User&lt;/td&gt;
&lt;td&gt;5&lt;/td&gt;
&lt;td&gt;56&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Google-Extended&lt;/td&gt;
&lt;td&gt;1&lt;/td&gt;
&lt;td&gt;56&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Eight distinct practices out of 56 close the door to at least one assistant crawler. That is roughly one in seven.&lt;/p&gt;

&lt;p&gt;The shape of those eight matters more than the count. Five of them refuse seven of the eight assistant crawlers at once, which is not what a considered policy looks like. A practice that decided to keep ChatGPT out would write one line. Refusing almost every assistant simultaneously, with nothing in robots.txt, is the signature of a single switch somewhere in front of the website.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Seven of the eight never declared it.&lt;/strong&gt; Only one site has the rule written in its own robots.txt, where the owner could find it. In the other seven the block is invisible from inside the website.&lt;/p&gt;

&lt;h2&gt;
  
  
  The contrast nobody expects
&lt;/h2&gt;

&lt;p&gt;Fourteen of the 56 practices publish an llms.txt, the newer file meant to help assistants read a site. That is almost twice as many sites doing the optional new thing as sites accidentally locked out of the assistants entirely.&lt;/p&gt;

&lt;p&gt;Read that in order. Fourteen practices went looking for the current best practice and added a file. Eight have a door closed that they did not close and cannot see. The work went to the part that is visible in a blog post, not to the part that decides whether a crawler gets a page at all.&lt;/p&gt;

&lt;p&gt;One more measurement, because it corrects something we expected to find: not a single site in the sample uses the noai meta tag. Whatever is happening here, almost nobody chose it deliberately.&lt;/p&gt;

&lt;h2&gt;
  
  
  How to check your own, in four minutes
&lt;/h2&gt;

&lt;p&gt;None of this needs a tool or a consultant.&lt;/p&gt;

&lt;h2&gt;
  
  
  First, open your robots.txt
&lt;/h2&gt;

&lt;p&gt;Type your domain followed by /robots.txt in a browser. If you see rules naming GPTBot, ClaudeBot, PerplexityBot or Google-Extended with Disallow, that is a declared block. If the file is empty or missing, that layer is clear, which is not the same as open.&lt;/p&gt;

&lt;h2&gt;
  
  
  Second, look at what sits in front of your site
&lt;/h2&gt;

&lt;p&gt;Most invisible blocks come from a CDN or a security layer, not from your website builder. If your domain runs through a service that filters traffic, find its bot settings. Rules that promise to stop scrapers or AI bots do exactly that, including the assistants you want.&lt;/p&gt;

&lt;h2&gt;
  
  
  Third, ask an assistant for your own page
&lt;/h2&gt;

&lt;p&gt;Give ChatGPT or Claude the URL of your practice page and ask it to tell you what the page says. If it cannot fetch it while the page loads fine in your browser, you have the second layer problem, and you just proved it in thirty seconds.&lt;/p&gt;

&lt;h2&gt;
  
  
  Fourth, ask your host the one useful question
&lt;/h2&gt;

&lt;p&gt;The question is not whether they block AI. It is: does anything in front of my site return 403 to a named crawler? A support agent can answer that in a sentence.&lt;/p&gt;

&lt;h2&gt;
  
  
  What this does not prove, and it matters
&lt;/h2&gt;

&lt;p&gt;Being readable is not being cited. Forty eight of these 56 practices can be fetched perfectly well by every assistant crawler we tested, and most of them still will not come up when someone asks for a therapist in their city. Access is the floor, not the outcome. A crawler that can read a page nobody would ever quote has changed nothing.&lt;/p&gt;

&lt;p&gt;What the measurement does prove is narrower and worth knowing: &lt;strong&gt;for about one in seven practices, the work of being chosen has not started, because the page cannot be read at all.&lt;/strong&gt; And in seven of those eight cases, nobody made that decision. It was made for them, by something they are paying for.&lt;/p&gt;

&lt;p&gt;If you want the rest of the picture, the part that decides whether an assistant quotes you once it can read you, that is what &lt;a href="https://contento.solutions/method" rel="noopener noreferrer"&gt;our method&lt;/a&gt; measures, and &lt;a href="https://contento.solutions/academy/make-website-citable-ai-search" rel="noopener noreferrer"&gt;how to make a page worth citing&lt;/a&gt; is the practical version. The file everyone is adding is covered in &lt;a href="https://contento.solutions/academy/llms-txt-what-to-put-in-it" rel="noopener noreferrer"&gt;what to actually put in llms.txt&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;We did not name any of the 56 practices. They did not ask to be measured, and seven of the eight blocked sites are not doing anything wrong: they are the ones with the least visibility into their own setup. Naming them would punish the people this piece is meant to help.&lt;/p&gt;

&lt;h2&gt;
  
  
  Questions we get about this
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Does blocking an AI crawler hurt my Google ranking?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
No. These are separate crawlers from the one that indexes your site for classic search results. Blocking GPTBot does nothing to your position in Google's blue links. What it changes is whether an assistant can read your page when someone asks it a question.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;My robots.txt is empty, so I am fine, right?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Not necessarily. In seven of the eight blocked practices we found, robots.txt was clean and the refusal came from a layer in front of the website. An empty robots.txt means you did not write a block. It does not mean nobody else did.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Should I add an llms.txt file?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
It does not hurt, and fourteen of these 56 practices already have one. Just be aware of the order: a file that tells an assistant what your site contains is useless if the assistant cannot fetch your pages in the first place. Check the door before you add the sign.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Is being readable enough to get recommended?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
No, and this is the honest limit of everything above. Forty eight of these 56 practices are perfectly readable and most still will not be named when someone asks for a therapist in their city. Access is the floor, not the outcome.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Why did you not name the practices you tested?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
They did not ask to be measured, and seven of the eight blocked sites are not doing anything wrong. They are the ones with the least visibility into their own setup, and naming them would punish exactly the people the piece is meant to help.&lt;/p&gt;

&lt;p&gt;If you want to know which of the two problems your own site has, the door or the citation, that is a reasonable thing to bring to a call. What the practice does and what gets reported is on the &lt;a href="https://contento.solutions/services" rel="noopener noreferrer"&gt;services page&lt;/a&gt;, with more detail on the &lt;a href="https://contento.solutions/faq" rel="noopener noreferrer"&gt;FAQ page&lt;/a&gt;. The call is &lt;a href="https://contento.solutions/book-a-call" rel="noopener noreferrer"&gt;thirty minutes&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;Have you ever found a block on your own site that nobody on the team wrote? Tell me in the comments where it turned out to live: the CDN, a security plugin or the host. I would like to know which layer shows up most often outside this niche.&lt;/p&gt;

</description>
      <category>seo</category>
      <category>ai</category>
      <category>webdev</category>
      <category>datascience</category>
    </item>
    <item>
      <title>How I run a one-person company's website like a production system</title>
      <dc:creator>Lucas Fernández Díaz</dc:creator>
      <pubDate>Tue, 06 Oct 2026 13:02:20 +0000</pubDate>
      <link>https://dev.to/lucasfernandezdiaz/how-i-run-a-one-person-companys-website-like-a-production-system-2j0l</link>
      <guid>https://dev.to/lucasfernandezdiaz/how-i-run-a-one-person-companys-website-like-a-production-system-2j0l</guid>
      <description>&lt;p&gt;&lt;em&gt;A version of this story first appeared on &lt;a href="https://contento.solutions/blog/how-i-built-contento-as-a-system" rel="noopener noreferrer"&gt;Contento Solutions&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;My company's website runs on Squarespace. It also has 33 checks that run against it every morning, 118 scripts, a git history of 684 commits since September 1, two independent backups a day and a decision log with more than 900 rows.&lt;/p&gt;

&lt;p&gt;I'm the only person in the company. That's exactly why it's built this way.&lt;/p&gt;

&lt;p&gt;This is the architecture, what each part is for, and the two places where the platform said no and I had to find another way in.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why treat a small site like production?
&lt;/h2&gt;

&lt;p&gt;Because nobody else is going to notice when it breaks. On a team, someone sees the broken button on their phone. When you're alone, a regression can sit on the site for weeks. The fix isn't paying more attention. It's building something that pays attention for you.&lt;/p&gt;

&lt;p&gt;The site has 41 public pages: an Academy with 11 guides, a Blog with 7 posts, six landing pages for specific fields, plus the usual home, services, pricing and so on. Every page carries custom code. That's a lot of surface for one person to eyeball.&lt;/p&gt;

&lt;h2&gt;
  
  
  The front: Squarespace plus custom code
&lt;/h2&gt;

&lt;p&gt;Squarespace handles hosting, the CMS and the editor. Everything that makes the site look like itself is mine: custom HTML blocks and per-page code injections, all wrapped in one root class so the CSS is scoped and never leaks into the platform's own styles. One palette, three typefaces, the same button sizes everywhere.&lt;/p&gt;

&lt;p&gt;Writing to Squarespace is where it got interesting. There's no public API for page content, so I watched what the editor's own Save button sends and replayed those same requests from a logged-in session. Two rules came out of that:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;
&lt;strong&gt;Read the object, change one key, send the whole thing back.&lt;/strong&gt; Rebuilding a payload by hand, or cloning a section with fresh IDs, gets you a 400 with no error body. The same request on the object you just read gets a 200.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;The error message is the documentation.&lt;/strong&gt; When the layout grid rejects a change, the 400 tells you the exact limit it wanted. I raised the values until it stopped complaining.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;The code injection editor was the other wall. Setting its value programmatically looks perfect: the editor updates, the Save button lights up, you save. Reload the page and the change is gone. It only persists what arrives as real input. So the process is: put the cursor where it belongs, enter the text as a person would, and never call anything saved until a cold reload outside the editor proves it.&lt;/p&gt;

&lt;h2&gt;
  
  
  site.json: one source for every script
&lt;/h2&gt;

&lt;p&gt;The scripts don't keep their own lists. Everything structural lives in one machine-readable file, &lt;code&gt;site.json&lt;/code&gt;: the list of public URLs, the site IDs, every surface where custom code lives (with a SHA-256 of its last known content), and the baseline value for every gate.&lt;/p&gt;

&lt;p&gt;That rule exists because of a bug. At one point the list of URLs had been copied into six different scripts, and one gate spent several sprints checking 29 of the 33 pages that existed then. Nothing failed. It just wasn't looking. Now there's one list, and every script imports it.&lt;/p&gt;

&lt;p&gt;The baselines get rewritten only on purpose, with an explicit update mode, when a change is intended. Never as a side effect of a run.&lt;/p&gt;

&lt;h2&gt;
  
  
  33 checks before coffee
&lt;/h2&gt;

&lt;p&gt;&lt;code&gt;ct-gates.py&lt;/code&gt; runs all the gates against the public URLs, anonymously, with no session. Each check compares a number against its baseline in &lt;code&gt;site.json&lt;/code&gt; and prints PASS or FAIL. The exit code is 0 if everything passes, 1 if something fails, 2 if the network did.&lt;/p&gt;

&lt;p&gt;The daily run reads the served HTML. Its 33 checks are grouped into families:&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Family&lt;/th&gt;
&lt;th&gt;What it catches&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;HTTP and sitemap&lt;/td&gt;
&lt;td&gt;URLs that don't answer, public pages missing from the list&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Structured data&lt;/td&gt;
&lt;td&gt;
&lt;code&gt;ld+json&lt;/code&gt; blocks that don't parse&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Links&lt;/td&gt;
&lt;td&gt;broken links, orphan pages that aren't orphans on purpose&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Typography&lt;/td&gt;
&lt;td&gt;a banned font sneaking back in, the wrong family on eyebrows&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Page weight&lt;/td&gt;
&lt;td&gt;pages outside their normalized size band&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Footer and CTAs&lt;/td&gt;
&lt;td&gt;missing global footer, drift in call to action labels&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;FAQ&lt;/td&gt;
&lt;td&gt;questions on a page without matching FAQ schema&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;llms.txt&lt;/td&gt;
&lt;td&gt;site pieces missing from the file AI assistants read&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Once a week, a heavier run renders the pages in headless Chrome and checks what HTML can't tell you: computed visibility after JavaScript, real keyboard focus order through the DevTools protocol, and alignment at three screen widths.&lt;/p&gt;

&lt;p&gt;Most of these families exist because something got through once. A banned font lived on the site for months in 183 declarations until I spotted it by accident on a form. Now it's a gate, and it can't come back without a FAIL.&lt;/p&gt;

&lt;h2&gt;
  
  
  The watcher: did something change that I didn't touch?
&lt;/h2&gt;

&lt;p&gt;A gate tells you when a number leaves its band. It doesn't tell you when a number moves inside the band, or when someone updated the baseline without looking. So a second script, the watcher, runs the gates on a schedule and compares every result against the &lt;strong&gt;previous run&lt;/strong&gt;, not only against the baseline. It also records the git HEAD of each run:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;number changed and HEAD changed: I did that. Report it, don't alarm.&lt;/li&gt;
&lt;li&gt;number changed and HEAD is the same: it moved on its own. That's the one to look at.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;That logic caught the strangest bug on the site. One page appeared to double in weight in a few days, with nothing published. I saved both versions and diffed them. The content was identical byte for byte. The difference was runs of blank space the platform injects depending on which cached copy you get, more than half the page. Raw page weight on Squarespace turned out to be useless, so the weight gate now measures the page with whitespace collapsed.&lt;/p&gt;

&lt;p&gt;The watcher runs from Windows Task Scheduler at 09:00 every day, with the full rendered run on Sundays. The runner holds a lock so two scheduled tasks firing at once after a cold boot don't overwrite each other's log, and it retries once after 90 seconds, because the first request after waking a laptop fails often.&lt;/p&gt;

&lt;p&gt;The honest gap: if the laptop is off at 09:00, nothing runs. Moving the watcher to a small always-on server is next on the list.&lt;/p&gt;

&lt;h2&gt;
  
  
  Backups that don't depend on remembering
&lt;/h2&gt;

&lt;p&gt;&lt;code&gt;ct-respaldo.py&lt;/code&gt; runs at the end of the same morning round and makes two copies of every repo:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;A push to a private GitHub remote. It never forces, and it never commits human work. The only files it commits on its own are the ones the checkers write.&lt;/li&gt;
&lt;li&gt;A &lt;code&gt;git bundle&lt;/code&gt; with the full history in a cloud-synced folder: one latest copy that gets overwritten and one per ISO week, keeping the last two.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;Anything it couldn't back up, uncommitted changes or unpushed commits, gets written to a status file, and the run exits 1. Restoring is one command: clone the bundle.&lt;/p&gt;

&lt;h2&gt;
  
  
  A decision log instead of a memory
&lt;/h2&gt;

&lt;p&gt;Every decision goes into a markdown table: ID, evidence, decision, status. It's past 900 rows. The ones that went wrong stay in, with what was learned.&lt;/p&gt;

&lt;p&gt;It's grepped by ID, never read whole. That matters for the agents too: they read the same record, so they don't propose something I already rejected last week. A daily brief, &lt;code&gt;HOY.md&lt;/code&gt;, is composed by a script from the gates, the backups and the queue. No AI writes it by hand, and every number in it says where it came from.&lt;/p&gt;

&lt;h2&gt;
  
  
  llms.txt and IndexNow, kept in sync by scripts
&lt;/h2&gt;

&lt;p&gt;&lt;code&gt;llms.txt&lt;/code&gt; was written by hand once, and a few weeks later an audit found it listed 7 of 12 pieces because nobody remembered to add the new ones. Now &lt;code&gt;ct-llms.py&lt;/code&gt; regenerates it from the sitemap. It keeps the hand-written sections intact, rewrites only the Academy and Blog lists in sitemap order, keeps the curated line for any piece already there, and builds new lines from each page's title and meta description. A check mode exits 1 if the live file is missing anything, and that check is one of the daily gates.&lt;/p&gt;

&lt;p&gt;For IndexNow, the problem was that Squarespace won't let you put a file at the domain root, and that's where the key has to live. The workaround: upload the key as a site file and add a URL mapping that redirects the root path to it. &lt;code&gt;ct-indexnow.py&lt;/code&gt; fetches the key from the root before every ping and refuses to send anything if it doesn't match, then notifies the IndexNow endpoints and reports which ones accepted.&lt;/p&gt;

&lt;h2&gt;
  
  
  Agents split by function
&lt;/h2&gt;

&lt;p&gt;I work with AI agents, and I split them the way you'd split a team. Several do research. One evaluates what all the others produce. One documents. One keeps improving the system itself. Others work on branding, others on code. Some days 22 run at once.&lt;/p&gt;

&lt;p&gt;What they share is the system above: the same &lt;code&gt;site.json&lt;/code&gt;, the same gates, the same decision log. The AI executes. The judgment stays with me.&lt;/p&gt;

&lt;h2&gt;
  
  
  What I'd tell anyone running a site alone
&lt;/h2&gt;

&lt;p&gt;Write the decision down before you build. Put every list in one file. Make each bug you find into a check that runs forever. Compare against yesterday, not only against the rule. And never call anything saved until a cold reload says so.&lt;/p&gt;

&lt;p&gt;This is what I build for businesses: the site, the measurement, the content and the automation behind it, end to end, with a strategy and a plan.&lt;/p&gt;

&lt;p&gt;Bring me a digital problem.&lt;/p&gt;

</description>
      <category>webdev</category>
      <category>python</category>
      <category>automation</category>
      <category>devops</category>
    </item>
    <item>
      <title>Why a Perfect Consistency Score Is a Warning Sign</title>
      <dc:creator>Lucas Fernández Díaz</dc:creator>
      <pubDate>Wed, 30 Sep 2026 13:00:00 +0000</pubDate>
      <link>https://dev.to/lucasfernandezdiaz/why-a-perfect-consistency-score-is-a-warning-sign-289</link>
      <guid>https://dev.to/lucasfernandezdiaz/why-a-perfect-consistency-score-is-a-warning-sign-289</guid>
      <description>&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://contento.solutions/academy/perfect-consistency-score-warning-sign" rel="noopener noreferrer"&gt;Contento Solutions&lt;/a&gt;, where I write about SEO and AI visibility for service businesses.&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;Consistency is the signal almost nobody fails. Fifty one of the 54 therapy websites we measured scored 85 or better, and 17 scored a perfect 100. Then we split the cohort by that score and it ran backwards: the flawless sites were lower on all five of the other signals we measured, and lower on the index overall.&lt;/p&gt;

&lt;h2&gt;
  
  
  What we actually measure when we say consistency
&lt;/h2&gt;

&lt;p&gt;The word covers a lot of ground in marketing writing, so here is the narrow version we score. Consistency asks whether a site contradicts itself in the places a machine reads. Not whether the brand feels coherent. Whether the served HTML says two different things.&lt;/p&gt;

&lt;p&gt;Three checks, and they start at 100 and come down from there. We deduct 15 points if any page title drops the brand name that the rest of the titles carry. We deduct 25 if the word placeholder is sitting in text a visitor can read, which is the fingerprint of a template that was filled in halfway. And we count the distinct prices published across the sampled pages, which we report as data and do not penalise at all.&lt;/p&gt;

&lt;p&gt;That last decision was a correction to our own tooling. An earlier version subtracted 20 points from any site publishing five or more price points, and that punishes a practice with seven services for having seven services. The real contradiction is the same item carrying two different prices, and this method cannot tell those apart from the outside, so it does not claim to. The &lt;a href="https://contento.solutions/academy/how-to-build-an-ai-visibility-index" rel="noopener noreferrer"&gt;method piece&lt;/a&gt; covers the other four times our instrument produced a confident wrong number.&lt;/p&gt;

&lt;h2&gt;
  
  
  Fifty one sites out of 54 pass it
&lt;/h2&gt;

&lt;p&gt;Here is the whole distribution. Seventeen sites at 100, thirty four at 85, one at 75, two at 60. Nothing below 60 in the entire cohort.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;What loses points&lt;/th&gt;
&lt;th&gt;Deduction&lt;/th&gt;
&lt;th&gt;Sites affected&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;A page title that drops the brand the other titles carry&lt;/td&gt;
&lt;td&gt;15&lt;/td&gt;
&lt;td&gt;36 of 54&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;The word placeholder visible in served text&lt;/td&gt;
&lt;td&gt;25&lt;/td&gt;
&lt;td&gt;3 of 54&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Several different prices published across pages&lt;/td&gt;
&lt;td&gt;0, reported as data&lt;/td&gt;
&lt;td&gt;34 publish at least one price&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Compare that to the rest of the index. Citability averages 50.5 across the same 54 sites. Verifiable identity, 57.8. Decision served, 60.5. Consistency sits at 88.6, which makes it the second highest of the six signals and the highest one that anybody has to work at, since retrievability at 97 is mostly the hosting platform doing its job.&lt;/p&gt;

&lt;p&gt;A signal that 94 percent of a niche passes cannot separate anyone inside that niche. That was the finding we expected to write up, and it would have been a short and slightly boring piece. Then we split the cohort.&lt;/p&gt;

&lt;h2&gt;
  
  
  Then we split the cohort, and the score ran backwards
&lt;/h2&gt;

&lt;p&gt;We grouped the 54 sites by their consistency score and averaged everything else. We expected the two groups to look roughly the same, because a signal nobody fails should not predict anything in either direction.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Signal&lt;/th&gt;
&lt;th&gt;17 sites scoring 100&lt;/th&gt;
&lt;th&gt;34 sites scoring 85&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Retrievability&lt;/td&gt;
&lt;td&gt;92.9&lt;/td&gt;
&lt;td&gt;98.8&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Citability&lt;/td&gt;
&lt;td&gt;38.4&lt;/td&gt;
&lt;td&gt;56.9&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Verifiable identity&lt;/td&gt;
&lt;td&gt;48.2&lt;/td&gt;
&lt;td&gt;62.5&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Decision served&lt;/td&gt;
&lt;td&gt;57.1&lt;/td&gt;
&lt;td&gt;62.9&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Friction to contact&lt;/td&gt;
&lt;td&gt;68.2&lt;/td&gt;
&lt;td&gt;77.5&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Overall index score&lt;/td&gt;
&lt;td&gt;67.5&lt;/td&gt;
&lt;td&gt;73.9&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Every row goes the wrong way. The sites that scored perfectly on consistency came out 6.4 points lower on the index than the sites that lost 15 points to a stray page title, and the gap is widest on citability, at 18.5 points.&lt;/p&gt;

&lt;p&gt;Our first move was to suspect the measurement rather than the finding, which is the habit that piece eight is mostly about. Three of the 17 perfect scorers had fewer than eight pages read, so we dropped them and ran it again on the 50 sites where the full sample came back. The gap survived: 63.2 against 71.7 on the five other signals. Smaller, still there, still pointing the same way.&lt;/p&gt;

&lt;p&gt;Say the size of this honestly. It is 17 sites against 34, in one cohort, in one niche, with no significance test run on it. That is a pattern worth explaining, not a law worth quoting.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why a thin site cannot contradict itself
&lt;/h2&gt;

&lt;p&gt;The mechanism falls out of the deduction, once you look at what it takes to trigger it.&lt;/p&gt;

&lt;p&gt;Losing those 15 points requires having a page whose title says something the other titles do not. A site with eight distinct pages, each about a specific problem, each titled for that problem, will eventually have a title that carries the topic and not the practice name. A site with four pages titled Home, About, Services and Contact, each with the brand pasted on the end, cannot possibly trip the check. There is nothing there to be inconsistent with.&lt;/p&gt;

&lt;p&gt;Citability is the row that makes this concrete, because it counts the opposite thing: headings, questions, lists, dated content, answers near the top of the page. A site built out of specific pages scores well there and is exposed to the title deduction. A site built out of four template slots scores 100 on consistency and 38 on citability, and those two numbers are the same fact described from two directions.&lt;/p&gt;

&lt;p&gt;So the perfect score is not measuring discipline. In this cohort it is mostly measuring how little the site says. That is a hypothesis that fits the data rather than something we proved, and it is the one we would test first on the next niche.&lt;/p&gt;

&lt;h2&gt;
  
  
  What this instrument cannot see
&lt;/h2&gt;

&lt;p&gt;The uncomfortable part is that consistency is also the signal where our measurement is furthest from the thing it is named after.&lt;/p&gt;

&lt;p&gt;Real inconsistency is a fee that says 150 on the services page and 180 on the FAQ. It is an intake process described one way in the about page and another way in the booking flow. It is a therapist listed as accepting a payer that the contact page says is not accepted. None of that is visible to a script reading eight pages, because catching it requires knowing which two statements are supposed to match.&lt;/p&gt;

&lt;p&gt;What we can see are two proxies, brand in the title and template leftovers, plus a count of prices we deliberately do not score. Calling that combination consistency is convenient shorthand. It is also the weakest name to score mapping in the whole index, and we would rather write that down than let the 88.6 sit there looking like good news.&lt;/p&gt;

&lt;h2&gt;
  
  
  Twenty sites publish no price at all
&lt;/h2&gt;

&lt;p&gt;One number from this signal is worth pulling out on its own, even though it costs nobody a point. Twenty of the 54 sites publish no price anywhere in the sampled pages. No fee, no range, no sliding scale.&lt;/p&gt;

&lt;p&gt;Those sites cannot be caught contradicting themselves on price, because they never say anything to contradict. Their consistency score is clean by absence. Their decision served score, which is the signal that asks whether a visitor can work out fit and cost without contacting anyone, is where that silence gets paid for, and the cohort average there is 60.5.&lt;/p&gt;

&lt;p&gt;This is the pattern the whole piece keeps circling. Publishing less is the cheapest way to look consistent, and every signal that measures usefulness moves in the other direction.&lt;/p&gt;

&lt;h2&gt;
  
  
  What to do with a signal that runs backwards
&lt;/h2&gt;

&lt;p&gt;Three things, and only the first one is work.&lt;/p&gt;

&lt;p&gt;Check the template leftovers, because that one is real and it is fast. Three sites out of 54 are serving the word placeholder to visitors, and there is no argument for leaving it there. Search your own site for it, along with lorem ipsum and any stock copy your platform shipped with. It takes ten minutes and it is the only part of this signal we would ask anyone to fix on a deadline.&lt;/p&gt;

&lt;p&gt;Do not chase the other 15 points. If your consistency score is 85 because a page title carries a topic rather than your practice name, that title is very likely doing its job. Rewriting eight titles to include the brand will lift one signal out of six by 15 points and will not touch citability, identity or whether a visitor can decide. Spending an afternoon there is spending it on the scoreboard.&lt;/p&gt;

&lt;p&gt;And treat a perfect score as a prompt rather than a result. If you scored 100, the question is not what you did right. It is whether there is enough on the site to be inconsistent about. That question is answered by the citability and decision served numbers sitting next to it, which is the argument for scoring the three axes separately instead of blending them, and it is the whole reason the &lt;a href="https://contento.solutions/method" rel="noopener noreferrer"&gt;method&lt;/a&gt; is built the way it is.&lt;/p&gt;

&lt;h2&gt;
  
  
  Questions
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Does a perfect consistency score mean my site is bad?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
No. It means the score on its own tells you almost nothing, so read it next to citability and decision served. In this cohort the sites at 100 averaged 38.4 on citability, and that is the number that carries the information.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Should I put my brand name in every page title?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Only if it fits. Our check looks for the brand somewhere in the title rather than at the end, and dropping it on one page costs 15 points on one signal of six. A title that describes what the page is about is worth more than the 15 points.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Why do you report prices without scoring them?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Because publishing seven prices is what a practice with seven services looks like, not a contradiction. The contradiction we would want to catch is the same service priced twice, and reading eight pages from the outside cannot distinguish the two, so we report the count and leave the judgement to the reader.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Is this finding strong enough to act on?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Act on the template leftovers. Treat the inversion as a reason not to spend time on the remaining 15 points, which is a decision about where not to work rather than a claim about what will happen if you do. It is 17 sites against 34 in a single niche with no significance test.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;How would I check this on my own site?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Read the served HTML of your eight most important pages, not the rendered page. Count how many carry your brand in the title, search all eight for placeholder text, and list every price you find. The &lt;a href="https://contento.solutions/academy/how-to-build-an-ai-visibility-index" rel="noopener noreferrer"&gt;method piece&lt;/a&gt; has the sampling and exclusion rules that make the result repeatable.&lt;/p&gt;

&lt;h2&gt;
  
  
  The part that transfers
&lt;/h2&gt;

&lt;p&gt;A signal everybody passes is not a standard, it is a floor, and a floor tells you who fell through rather than who is doing well. The useful move is to check which of your signals separate the field and which of them only separate the broken from everyone else. In this cohort, retrievability at 97 and consistency at 88.6 are floors. Citability at 50.5 is where the field actually spreads out.&lt;/p&gt;

&lt;p&gt;Run the same split on your own index and expect at least one signal to run backwards. Ours did, on the axis we would have been happiest to report, which is roughly how these things go. The full cohort results are in &lt;a href="https://contento.solutions/academy/therapy-websites-ai-visibility-index" rel="noopener noreferrer"&gt;the index itself&lt;/a&gt;, and we should say plainly that we sell into the niche we measured.&lt;/p&gt;

&lt;p&gt;If you want a read on what a machine can currently retrieve, trust and quote from your own site, that is a reasonable thing to bring to a call. What the practice does and what gets reported is on the &lt;a href="https://contento.solutions/services" rel="noopener noreferrer"&gt;services page&lt;/a&gt;, with more detail on the &lt;a href="https://contento.solutions/faq" rel="noopener noreferrer"&gt;FAQ page&lt;/a&gt;. The call is &lt;a href="https://contento.solutions/book-a-call" rel="noopener noreferrer"&gt;thirty minutes&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;If you score sites, or your own site, on more than one signal: which one would you expect to run backwards if you split by the signal everybody passes? Tell me in the comments. I would like to know whether ours is a quirk of one niche or something that shows up everywhere.&lt;/p&gt;

</description>
      <category>seo</category>
      <category>webdev</category>
      <category>ai</category>
      <category>datascience</category>
    </item>
    <item>
      <title>What Actually Goes In an llms.txt</title>
      <dc:creator>Lucas Fernández Díaz</dc:creator>
      <pubDate>Wed, 16 Sep 2026 20:05:37 +0000</pubDate>
      <link>https://dev.to/lucasfernandezdiaz/what-actually-goes-in-an-llmstxt-2me5</link>
      <guid>https://dev.to/lucasfernandezdiaz/what-actually-goes-in-an-llmstxt-2me5</guid>
      <description>&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://contento.solutions/academy/llms-txt-what-to-put-in-it" rel="noopener noreferrer"&gt;Contento Solutions&lt;/a&gt;, where I write about SEO and AI visibility for service businesses.&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;An llms.txt file is a plain text summary of your site. It is written for language models, not for crawlers. It is also a proposal rather than a standard (&lt;a href="https://llmstxt.org/" rel="noopener noreferrer"&gt;Jeremy Howard published it&lt;/a&gt; in September 2024), and no major AI company has confirmed that it reads one. Write it anyway. The reason is not the file.&lt;/p&gt;

&lt;h2&gt;
  
  
  The honest version of what this file is
&lt;/h2&gt;

&lt;p&gt;Somebody proposes a convention. A few thousand sites adopt it. Nobody who runs a large model says whether they use it. That is roughly where llms.txt sits today, and any article that tells you otherwise is selling something.&lt;/p&gt;

&lt;p&gt;So the useful question is not whether assistants read it. It is whether writing one is worth an hour of your time given that they might not.&lt;/p&gt;

&lt;p&gt;It is. Here is the part that gets left out of most explanations: the file is a forcing function. To write it you have to state, in fewer than a thousand words and without marketing cover, what your business does, who it is for, who it is not for, what it charges, and what it refuses to claim. Most businesses have never written that down anywhere. The ones that have tend to have a clearer website, because the same clarity leaks into every page.&lt;/p&gt;

&lt;p&gt;The file is the cheap part. The decision is the asset.&lt;/p&gt;

&lt;h2&gt;
  
  
  Where it sits next to the files you already have
&lt;/h2&gt;

&lt;p&gt;Four files describe your site to machines, and they answer four different questions. Confusing them is the most common mistake in this area.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;File&lt;/th&gt;
&lt;th&gt;Who reads it&lt;/th&gt;
&lt;th&gt;The question it answers&lt;/th&gt;
&lt;th&gt;Honored?&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;robots.txt&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;Any crawler that chooses to&lt;/td&gt;
&lt;td&gt;May you fetch this path?&lt;/td&gt;
&lt;td&gt;Yes, by long convention&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;sitemap.xml&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;Search crawlers&lt;/td&gt;
&lt;td&gt;What exists here, and when did it change?&lt;/td&gt;
&lt;td&gt;Yes&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Schema, as JSON-LD&lt;/td&gt;
&lt;td&gt;Search engines and assistants&lt;/td&gt;
&lt;td&gt;What is this page, expressed as data?&lt;/td&gt;
&lt;td&gt;Yes, where the type is supported&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;llms.txt&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;Proposed for language models&lt;/td&gt;
&lt;td&gt;What is this business, in prose?&lt;/td&gt;
&lt;td&gt;Unconfirmed&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Notice that only one of the four is prose. That is the whole idea. Schema is excellent at facts with a fixed shape, and useless for the things that do not have one: judgement, scope, refusals, the difference between two services that look identical on a price list. An &lt;a href="https://contento.solutions/academy/entity-map-small-brands-content-architecture" rel="noopener noreferrer"&gt;entity map&lt;/a&gt; handles the structured half. This file handles the half that only survives in sentences.&lt;/p&gt;

&lt;h2&gt;
  
  
  What actually goes in it
&lt;/h2&gt;

&lt;p&gt;The convention asks for Markdown at &lt;code&gt;/llms.txt&lt;/code&gt;, opening with an H1 that names the business and a blockquote that summarises it. After that the format is loose, which is where most files go wrong. Loose does not mean write whatever you want. It means the discipline has to come from you.&lt;/p&gt;

&lt;p&gt;Six things earn their place.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;One sentence that survives paraphrase.&lt;/strong&gt; Assume a model will compress everything you wrote into two lines and hand those to somebody who has never heard of you. Write that compression yourself. If your summary only makes sense with the rest of the page attached, it is not a summary.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;The boundary of what you do.&lt;/strong&gt; Not the list of services. The edge. A sentence like "this is not a general agency and does not sell rankings, packages or volume" tells a model more than six bullet points of capability, because it can be used to rule you out. Being ruled out correctly is not a loss. It is the only way being ruled in means anything.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Who the buyer is, in their words.&lt;/strong&gt; Sector, size, situation. A therapist opening a second practice is a different reader from a firm of nine accountants, and if your file cannot tell them apart, neither can anything reading it.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Names, spelled the way people get them wrong.&lt;/strong&gt; If the founder's name carries an accent, write the unaccented spelling too and say it refers to the same person. This sounds small. It is the difference between a model resolving you to one entity and resolving you to two half entities that never accumulate.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Your citation terms.&lt;/strong&gt; Say plainly whether an assistant may quote you, at what length, and with what attribution. Most sites are silent here, and silence is the least useful answer. If you want to be cited, saying so costs nothing.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;A contact route that is not a form.&lt;/strong&gt; An address a system can read and repeat. Forms are drawn by JavaScript on most platforms, which means a machine reading your page may find no way to reach you at all.&lt;/p&gt;

&lt;h2&gt;
  
  
  What to leave out, and this is the harder half
&lt;/h2&gt;

&lt;p&gt;The failure mode is not omission. It is writing the file as a pitch.&lt;/p&gt;

&lt;p&gt;Leave out adjectives that cannot fail. Leading, innovative, trusted, results driven. Every one of them is true of every competitor, which means none of them separate you, which means a model summarising ten businesses in your category will produce ten identical sentences and pick on some other basis.&lt;/p&gt;

&lt;p&gt;Leave out numbers you cannot stand behind in public. A percentage with no method attached is worse than no percentage, because the first person to ask where it came from finds out that nobody knows. If a number matters, say how it was measured. If it cannot be measured honestly with the tools you have, say that instead. We do exactly this on our own file, and it is not modesty, it is that the alternative is a claim we would have to defend.&lt;/p&gt;

&lt;p&gt;Leave out anything the site itself contradicts. A file that says pricing is transparent, on a site where pricing is on request, has told a reader something more damaging than either statement alone. This is the cheapest audit in the whole exercise: read the file, then read the site, and fix whichever one is lying.&lt;/p&gt;

&lt;p&gt;And leave out the entire sitemap. A list of every URL is what &lt;code&gt;sitemap.xml&lt;/code&gt; is for. Six or eight pages that matter, each with a line saying why it exists, is worth more than eighty links with no explanation.&lt;/p&gt;

&lt;h2&gt;
  
  
  How to check it did anything
&lt;/h2&gt;

&lt;p&gt;You mostly cannot, and pretending otherwise is how this field earns its reputation.&lt;/p&gt;

&lt;p&gt;There is no report for this. No console panel, no impression count, no line in analytics that says an assistant read your file. Anyone offering you a dashboard for llms.txt performance has invented the numbers on it.&lt;/p&gt;

&lt;p&gt;What you can do is sample. Ask four or five assistants the questions a buyer would actually ask, in a fresh session with no history: who does this kind of work for this kind of business, what does it cost, who is it not for. Record the answers verbatim, with the date. Repeat in a month. You are not measuring a metric, you are watching a description of your business drift toward or away from the one you wrote. That is a slower and less satisfying instrument than a chart, and it is the honest one. The same sampling discipline is what makes &lt;a href="https://contento.solutions/academy/search-console-ai-overviews-visibility-measurement" rel="noopener noreferrer"&gt;Search Console useful after AI Overviews&lt;/a&gt;, where the temptation to read a number that does not mean what it looks like is even stronger.&lt;/p&gt;

&lt;p&gt;One more check, and it takes a minute. Fetch your own file with a plain request and read it end to end as a stranger would. If you finish it and could not say who this business turns away, the file is not finished.&lt;/p&gt;

&lt;h2&gt;
  
  
  The order to do this in
&lt;/h2&gt;

&lt;p&gt;Write the file last, not first.&lt;/p&gt;

&lt;p&gt;An llms.txt that describes a site a model cannot fetch, or cannot parse, or finds contradicts itself page to page, is a well written summary of a problem. The sequence that works is unglamorous: make the pages readable without JavaScript, make the structured data say what the page says, make the site agree with itself, and then write the file that summarises all of it. We wrote about the first three as &lt;a href="https://contento.solutions/academy/make-website-citable-ai-search" rel="noopener noreferrer"&gt;making a site citable&lt;/a&gt;, and the order matters more than any individual step.&lt;/p&gt;

&lt;p&gt;Then the hour you spend on the file is spent describing something true, which is the only version of this work that survives contact with a reader who checks.&lt;/p&gt;

&lt;h2&gt;
  
  
  Questions people ask about this
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Do ChatGPT, Claude, Gemini or Perplexity actually read llms.txt?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
None has confirmed it. Treat any claim that they do, or that they do not, as unverified. The case for writing one rests on the clarity it forces and its near zero cost, not on a confirmed behaviour.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Is llms.txt a replacement for schema markup?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
No. They do different jobs. Schema states facts in a form machines parse reliably. llms.txt states judgement, scope and boundaries in prose, which schema has no way to express. A site that needs one usually needs both.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Does adding llms.txt affect Google rankings?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
There is no evidence that it does, and no reason to expect it. It is not a ranking file. Anybody selling it as one is describing a mechanism that has never been shown to exist.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;How long should it be?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
Long enough to state what you do, who for, who not for, and how to reach you. Ours runs about seven thousand characters and that is on the generous side. Under a thousand words is a reasonable ceiling for most businesses.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Where does the file go, and how do I add one on a hosted platform?&lt;/strong&gt;&lt;br&gt;&lt;br&gt;
It belongs at the root, as &lt;code&gt;yourdomain.com/llms.txt&lt;/code&gt;. Platforms that do not let you place arbitrary files at the root make this awkward, and the workaround varies by platform. If yours will not allow it, the schema and page level work still stands on its own.&lt;/p&gt;

&lt;h2&gt;
  
  
  The short version
&lt;/h2&gt;

&lt;p&gt;Write the file because writing it makes you decide what your business is. Keep it true, keep it specific, keep it consistent with the site it describes. If assistants turn out to read it, you were ready. If they never do, you still ended up with the clearest paragraph anyone in your company has written about what you sell.&lt;/p&gt;

&lt;p&gt;If you want a second read on whether your site says the same thing to a person, a crawler and an assistant, that is a reasonable thing to bring to a call. How this practice works and what gets reported is on the &lt;a href="https://contento.solutions/services" rel="noopener noreferrer"&gt;services page&lt;/a&gt;, with more detail on the &lt;a href="https://contento.solutions/faq" rel="noopener noreferrer"&gt;FAQ page&lt;/a&gt;. The call is &lt;a href="https://contento.solutions/book-a-call" rel="noopener noreferrer"&gt;thirty minutes&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;If you already have an llms.txt, drop the URL in the comments and tell me what you decided to leave out. That choice says more about a business than anything that made it into the file.&lt;/p&gt;

</description>
      <category>seo</category>
      <category>ai</category>
      <category>webdev</category>
      <category>llm</category>
    </item>
  </channel>
</rss>
