The MRC method: readable, understandable, citable
MRC stands for Machine Readable Content: it is the method I use on every site. A search engine or an AI assistant cites content only if it can read it, understand who it is about and attribute it to whoever wrote it. Here you will find the three pillars (readable, understandable, citable) and the three steps (I measure, fix it with your team, check it works), each with what I look at and what I deliver: this page is for anyone who wants to know how I work before getting in touch.
Updated on
The three pillars
-
Readable
Content is readable when a bot finds it in the HTML the server sends, without waiting for JavaScript. GPTBot, ClaudeBot and PerplexityBot download the HTML but do not run JavaScript: if the text only appears in the browser, the page is almost empty for them. Googlebot, on the other hand, does run it.
Readable also means reachable:
robots.txt, sitemap andllms.txtin order, and no bot excluded without someone having decided it.What I look at
- The raw HTML of every template, compared with the page a visitor sees.
- The
robots.txtfile, bot by bot: search crawlers, user-initiated agents, training crawlers. - Sitemap, canonicals and
llms.txt. - The server logs: which bots actually visit, and which pages.
What I deliver
- What bots do not read, template by template, with the evidence.
- The rules for bots to put in
robots.txt.
-
Understandable
Content is understandable when a machine understands who you are and what you talk about. It understands this from entities, structured data and hreflang across languages.
The finding that comes up most often in my audits is a brand fragmented across different names, domains and profiles: to a machine, they are not necessarily the same company.
What I look at
- The brand name in the title tag, structured data and sharing meta tags: it must be the same.
- The structured data of every template: valid, complete, matching the visible text.
- Reciprocal hreflang across languages, and canonicals that never point to a page in another language.
What I deliver
- The fixes to the structured data, template by template.
- The brand as an entity, put back together: one name, one domain, the linked profiles.
- The hreflang tags and canonicals to fix.
-
Citable
Content is citable when whoever reads it can attribute it to you: there is an author, there are sources, the profiles are linked with
sameAs.How it is written matters too. A page made to be cited opens with two or three sentences that say what it is and who it is for, and that stand on their own; it puts definitions before details, numbers with their source, and the date of the last update.
What I look at
- Who signs the content, and whether the structured data says so.
- The profiles linked with
sameAs. - How the pages that matter most open.
- The answers of ChatGPT, Gemini, Perplexity, Copilot, AI Overviews and AI Mode: whether they cite you, and for what.
What I deliver
- The fixes to author, sources and linked profiles.
- Guidance for the pages that matter most: opening, definitions, sources, date.
The three steps
-
I measure the site as it is today
Every engagement starts with a measurement: crawl, bot logs, and the raw HTML compared with the page a visitor sees.
From the measurement I separate template issues from scattered ones. A template issue sits in the model that many pages are built from: it is fixed once for all of them.
What I look at
- The crawl of the site, template by template.
- The server logs: visits from search engine bots and from AI bots.
- The raw HTML, compared with the page a person sees.
What I deliver
- The audit: every finding with its evidence and why it matters.
- Template issues separated from scattered ones.
- Priorities, with an estimate of the development work.
-
I fix it with your team
The fixes are made together with the people who develop the site: structured data, rules for bots and entities.
I work with the team you have, in-house or at an agency.
What I look at
- The structured data of every template.
- The rules for bots in
robots.txt. - The brand as an entity: name, profiles,
sameAs.
What I deliver
- The fixes written up for the developers, in order of priority.
-
I check the fixes work
After release I check AI bot visits in the logs, citations in answers, and traffic.
Every engagement ends with a check: a fix counts when it shows in the data.
What I look at
- The logs after release: whether AI bots read the fixed pages.
- Citations in the answers of AI assistants.
- Traffic.
What I deliver
- The checklist after release: item by item, what is fixed and what is not.
It starts with a measurement
This site applies the method to itself: the How this site is built page tells the story. For a first measurement of your site, try MRC Check, which is free. To measure all of it, get in touch: I’ll reply personally and, if needed, we’ll arrange a call.