Public reports sometimes show 'better than X%'. That is a national (dataset-wide) percentile, not a marketing claim and not a sector ranking.
Each comparison uses the newest completed scan per domain. Older repeat scans of the same domain are excluded so frequent re-audits do not skew the set.
We compare the overall score (average of the four pillars) and, separately, each pillar score. Sector splits stay off until the annual-report dataset has classification and a minimum group size.
Only once there are at least 10 distinct scanned domains in the dataset. Below that the percentage is too thin statistically; we show nothing instead of a misleading figure.
There are no public leaderboards or 'top 10' lists on small samples. The mark remains a hard threshold (average ≥ 80, every pillar ≥ 60), independent of percentiles.
A high percentile does not prove full WCAG conformance or safety against every threat. It only situates your score among sites we have scanned so far.
Engine updates (new seo-cli version) can shift scores. On a paid report timeline we flag methodology changes when the engine version changes.
Jaarrapport in de maak. In de maak: het jaarrapport over de staat van het Nederlandse web. Lees meer →