There is a lot of talk about content, keywords and reputation. All of it matters, but it comes after a step people forget: reading. A crawler that cannot access a page will not rank it. An assistant that does not understand who you are will not quote you, or will quote you wrongly.
The good news: what search engines read on your site can be checked, often in a few minutes and without paid tools.
1. Indexability: is the page allowed to be read?
Three settings decide whether a page can appear in Google. The robots.txtfile, at the root of the site, which allows or blocks crawling. The noindextag, which asks for a page not to be indexed. And the canonical tag, which names the reference version when several URLs show the same content.
A classic mistake after a redesign: the noindex tag placed on the staging site goes live with it. The site is new, polished, and invisible. In Search Console, the page indexing report shows for each URL whether it is indexed, and if not, why.
2. The sitemap: the list you hand to search engines
The sitemap.xml file lists the pages you want indexed, with their last-modified date. It does not force Google to index them, but it helps Google find them and notice what has changed.
Three checks are enough: the sitemap exists and is submitted in Search Console; it only contains pages that are actually live, with no redirects or errors; its modification dates are accurate. A sitemap that lists deleted pages or leaves out new ones sends mixed signals.
3. Structured data: saying what the page is
A human understands a page at a glance: an article, a product page, an address. For a machine, nothing is obvious. Structured data, in the schema.org format, states it explicitly: this page is an Article published on a given date, this business is an Organization with a given logo, this product has a given price and given reviews.
Google uses it for some rich results, and it is the kind of signal that helps an engine connect a page to an identifiable entity. Google's Rich Results Test shows, page by page, what it understood and what is invalid.
4. The llms.txt file: an introduction written for AI
A recently proposed format, llms.txt is a text file placed at the root of the site, written in Markdown, that introduces your business and lists your key pages for language models. It is not an official standard, and nothing guarantees that a given assistant reads it. But it costs almost nothing, and it forces you to state clearly what you do — which helps your site too.
Ours is online at /llms.txt: it says who we are, where we work, and what should not be attributed to us.
What still escapes crawlers
- Text inside images. A heading or a price written into an image is not read as text.
- Content that only appears after a script runs. Google executes JavaScript, but later, and several AI crawlers do not execute it at all. What matters must be in the HTML.
- Contradictory information. A different address on the site, the Google Business Profile and directories blurs the entity that search engines are trying to piece together.
Checking your site
These checks take about an hour with Search Console and the Rich Results Test. Our audit page runs some of them automatically — indexability, sitemap, markup, readability by AI — and gives you a first free report in a few seconds.