MACHINE 404 / The index

We scanned 42 Greek business websites. 81.5% publish no product or service as structured data.

Hotels, retail, clinics, law firms and food brands, all well known. The average score was 57.8 out of 100, the range ran from 21 to 84, and 63% scored under 65. A further 13 refused an automated reader and could not be scored, and 2 did not answer at all.

By sector, worst first

What fails, and how often

How the sample was chosen, and how it was measured

This is a convenience sample of well-known Greek brands across five sectors, not a random draw from the business register. It says nothing about smaller businesses, which we did not measure.

Each site was scanned three times on 2026-09-18, with the same eighteen checks we sell, from one address, one site at a time with pauses between them. The number counted for each site is the median of the passes that scored it. A single pass can record a firewall hiccup as a finding; the median does not. The rubric is published in full with the weight of every check, so any number here can be re-derived.

The refusals, and what we are not claiming

13 of the 40 sites that answered refused an automated reader in most passes: a challenge page, a firewall error or an empty response. We are not claiming those sites block ChatGPT. A firewall may treat a verified AI crawler differently from a generic reader, and we did not impersonate one to find out. What we measured is only that their homepage refused a plain automated request.

The three heaviest gaps

81.5% publish no product or service as structured data: no Product or Service schema on the homepage or on the inner page we read. It is the heaviest check in the scan, at 12 points. Scored sites that do publish one: 5 of 27, and 3 of those state a price in it.

37% do not declare the company as data. And 92.6% name no AI crawler in robots.txt, so they have stated no policy either way.

The table below lists all eighteen checks by how often they failed.

Limits, stated plainly

One moment in time. A convenience sample, not a random one. Five sectors, not the economy. The homepage, one inner page and a handful of standard addresses per site, not whole sites. And measured by the company that sells the fix, which is why the rubric, the weights and the method are published rather than described. Re-run it yourself on any domain, including ours.

If one of these is yours

No business is named here and none will be. If you want to know whether you were in the sample, ask: you get the same thing the free form on the product page gives anybody, your score out of 100, your band and the three checks costing you the most, at no charge and whether or not you ever buy anything.

The other fifteen checks with their findings, and the instructions written for your own platform, are the report, and that is the part that is paid for.

Corrections

18 September 2026. The first version of this page, measured on 13 September, reported an average of 51.5, a best score of 70 and 96.6% with no offer as structured data. Two things changed. The scanner had never checked /feed/, the standard feed address of every WordPress site, so nine sites in the sample with a valid feed there lost eight points they had earned. And the first version counted a single pass per site; this one counts the median of three, and the three passes agreed on every site. Four sites also gained points on other checks. Re-scanning them with the scanner as it was before the /feed/ fix gives the same new results, so that change is not from the fix, and two dates cannot tell us whether the sites changed or the first pass missed something. Every number on this page is from the re-run.