Weekly report

Watch a site's GEO score, week by week

The scanner answers whether AI search engines can reach, read and cite a page right now. This answers the other question: whether that changed. Enter a domain and an address, confirm it, and the same 40 checks run once a week.

Start the weekly report

Free, no account. The unsubscribe link in every message works without signing in.

What the report contains

Each report carries the current score and grade, then the difference from the week before rather than the whole picture again. The interesting line is almost always the same one - a check that passed last week and fails now - and a report that buries it under twelve dimension tables is a report nobody reads after the third week.

A failing check is named, along with the evidence behind the verdict and a link to the page that explains what it wants. The scanner publishes every rule it applies, so a report can be argued with rather than taken on trust: if a verdict looks wrong, the rule, its wording and its weight are all readable at methodology.

Nothing in the weekly run is scored differently from a manual scan. Both call the same engine over the same four files, which is the only way the two numbers can be compared to each other at all.

How the weekly run works

  1. One batch a week. Every confirmed subscription is scanned in the same run, oldest-scan-first, so a slow week cannot starve one subscriber while another is served twice.
  2. The four files a GEO audit needs. robots.txt, the homepage, llms.txt and the sitemap, requested over whichever scheme answered. A site that refuses one of them is reported as refusing it rather than scored as if the file were absent - the two mean different things.
  3. Diffed against the last run. The previous score and the previous list of failing checks are stored precisely so the email can say what moved.
  4. Emailed, with a one-click unsubscribe. No confirmation step on the way out, and no reason asked for.

What is collected, and what happens to it

A subscription stores three things: the domain, the address to send to, and the score history for that domain. It is not sold, not shared, and not used to build a profile of anything. The withdrawal is recorded rather than the row deleted, so a request to know what was held can be answered after the fact - the detail is in the privacy policy, which names the retention period rather than describing one in general terms.

If the scanner reports something about a site that looks wrong, the same address that receives the report is the one that reaches the person who wrote the rule: services@llmention-geo.com.

Questions about the weekly report

What arrives in the email?

The score, the grade, and what changed since the previous run: which checks newly pass, which newly fail, and the evidence the failing ones saw. A report with no change says so in one line rather than padding itself out.

How often does it run, and when?

Once a week, in a single batch, rather than on demand. The scan reads robots.txt, the homepage, llms.txt and the sitemap the same way the free scanner does, so a weekly report and a manual scan of the same site cannot disagree.

Do I need an account?

No. The email address is the whole subscription: confirming it starts the report, and the unsubscribe link in every message ends it. There is no password to set, no dashboard to log in to, and nothing to cancel.

What is stored about me?

The domain, the email address, and the score history for that domain. Nothing else, and nothing is sold or shared. The unsubscribe link is in every message and works with one click, months later, without logging in.

Can I watch more than one domain?

Yes, by signing up once per domain with the same address. Note that unsubscribing stops every report to that address rather than only the one the link came from, because a link that leaves other messages arriving is the thing recipients report as spam.

Why does it need confirming?

Because anyone can type an address that is not theirs. The confirmation proves the address asked for the report, which protects the person whose address it is and keeps the sending domain out of spam filters.

Evidence and sources

Adding source citations produced the largest measured visibility gain for low-ranking sites, at +115%, ahead of the addition of expert quotations at +41% and statistics at +30-40%, across the strategies tested on generative engines. — Generative Engine Optimization, KDD 2024

The weekly run applies the same weightings this research produced, which is why a change in citability moves the score further than a change in a missing viewport tag.

Primary sources