Flowpane scores each of the eight governance artefacts from 0 to 100, then combines them into one weighted site score. Every number is built from what was observed on the public web, so drafts, fixes you have not yet published and failed runs never move it. Scores are Flowpane's product models for prioritising work, not compliance certificates.
Artefact scores and bands
Each artefact gets its own score from 0 to 100, shown in one of three bands.
| Score | Band | Meaning |
|---|---|---|
| 80 to 100 | Good | Healthy within Flowpane's model |
| 60 to 79 | Needs attention | Notable findings to address |
| 0 to 59 | Failing | Material findings, or a required file is missing |
The six fixed text files share one finding model. Sitemaps use a layered model of structure, truthfulness, freshness, reliability and, when a Deep Scan crawl exists, coverage. Cookies and consent produces an observation-risk score from bounded homepage evidence.
Some conditions override the arithmetic and set an artefact to 0: a missing file, an empty file, an HTML page served where a text file should be, an expired security.txt, or an ads.txt with no seller records.
Findings
| Severity | Meaning | Effect on score |
|---|---|---|
| Issue | An objective defect | May deduct |
| Warning | A risk, or an intent to review | Never deducts |
| Recommendation | Improvement advice | May deduct where configured |
| Information | Context or provenance | Never deducts |
Repeated findings of the same kind are grouped and capped, so one mistake repeated across 400 lines is not counted 400 times.
The site score
The site score is a weighted average of the artefacts that are included, rounded to a whole number.
| Artefact | Weight |
|---|---|
| Sitemaps | 35 |
| robots.txt | 15 |
| Cookies and consent | 12 |
| security.txt | 12 |
| ads.txt | 8 |
| llms.txt | 7 |
| ai.txt | 6 |
| humans.txt | 5 |
Required, recommended, not applicable
robots.txt, sitemaps and security.txt are always required. humans.txt is always recommended. ads.txt is not applicable unless the site declares that it sells advertising. The rest depend on the capabilities declared in the site's settings.
| State | Required | Recommended | Not applicable |
|---|---|---|---|
| Present | Scored at its weight | Scored at its weight | Left out |
| Missing | Counts as 0 | Counts as 0 | Left out |
| Blocked | Site is unscorable | Left out | Left out |
| Not yet checked | Site score pending | Left out | Left out |
The difference between required and recommended is what happens when Flowpane cannot see the file. A recommended artefact never holds the site score hostage.
Last good result
The site score uses each artefact's last good result. If a run errors, through a timeout or a server error, the previous result stands, so a transient failure does not erase good evidence. Missing and Blocked are observations, not errors, so they do replace it.
Blocked means unscorable
If a firewall or bot challenge blocks a required artefact, the site shows no score rather than a partial one. A figure built only from the files Flowpane could see would look real while hiding that the most important evidence is absent. See How Flowpane fetches for how Blocked differs from Missing.
What moves a score
Only the next public check. Proposed files, Fix, Fix all, AI Assist drafts, validation and approval never change a score. Flowpane does not publish to your origin: you publish through your own host or CMS, and the next scheduled or on-demand check observes the live file.
Scheduled cadence depends on the plan: Free is on demand, plus one automatic first run after verification; Starter weekly; Pro daily; Scale every 12 hours; Agency every 6 hours; Enterprise hourly. Every check runs Fetch, Parse, Assess, Score and Store. Results are immutable and grouped under one parent run, so history cannot be rewritten.
AI Readiness
AI Readiness is a separate grade from 0 to 100 that does not feed the site score: A from 90, B from 75, C from 60, D from 40, and F below 40.
Seven core checks cover AI crawler rules and posture in robots.txt, llms.txt quality, sitemap health and security.txt presence. Four bonus checks cover an ai.txt training stance, Content-Signal directives, consistent AI crawler coverage and a clean coherence result, with the total capped at 100. Coherence conflicts carry a penalty.
AI Readiness uses current results only, not the last good result. If the latest robots.txt check was blocked or failed, no grade is shown rather than a stale one.
A model, not a certificate
A score summarises findings under a documented model so that you can compare sites and prioritise work. It is not a legal or regulatory determination. That matters most for cookies and consent, where the score reflects observable risk, not compliance with any law.