What we found auditing 1,045 Tasmanian business websites

What we checked

We took every listing on this directory with a website address — 1,184 of them — fetched the homepage as a crawler would, and examined the source. We did not score design, speed or content quality. We asked one question: could a machine tell what this business is, where it works, and how to contact it?

Of those 1,184:

• 1,045 returned a real website we could parse.
• 77 did not respond at all — dead domains, timeouts, broken certificates.
• 62 were a social profile (Facebook, Instagram) standing in for a website.

The percentages below use 1,045 as the base, because counting a Facebook page as a failure to have structured data would be measuring the wrong thing.

The headline

304 of 1,045 sites — 29% — carry no structured data of any kind. No name, no address, no phone. Whatever a machine knows about those businesses, it inferred from photographs and paragraphs.

It gets worse among the sites that did bother. Of the 741 pages carrying a structured-data block:

• 475 (64%) do not include a telephone number.
• 461 (62%) do not include an address.
• 552 (75%) do not include opening hours.

That is the shape of the problem. A tag exists, but it says the site is a website, or names the business without saying where it is or how to ring it. Half the work was done, and the useful half was skipped.

The rest of the picture

• 95.8% have no real llms.txt. Only 44 of 1,045 do. Dozens more answer at that address with an error page wearing a 200 status, which is worse than nothing.
• 27.5% have no meta description — the one-line summary that appears under a search result. Machines write their own when it is missing, usually badly.
• 33% have no tap-to-call link. A third of Tasmanian business sites make a phone user copy a number into their dialler by hand.
• 27% have no clear main heading. The page begins with a logo, a menu and a slideshow, and never states the trade.
• 3.9% carry live template residue — 123-456-7890, [email protected], placeholder text. Forty-one sites are telling customers, and machines, a phone number that does not exist.
• 90% do have a robots.txt, which tells you the basics of a site being set up properly once and then never visited again.

What this actually costs

No single one of these loses a job. Together they produce the moment every tradesperson has had: a customer rings on Saturday, says "your website said you open Saturdays", and the site said nothing of the kind — the assistant guessed.

The businesses hurt most are the good ones. A plumber with thirty years of work and a site built in 2016 that still says "Welcome to our website" is invisible to the fastest-growing way people now search, and has no way of knowing it.

What we would fix first

1. Add the five fields to every page template — name, address, phone, hours, area served. One afternoon, once.
2. Delete the placeholders. Search your own site for 123-456-7890 and mysite.com. It takes a minute and it is embarrassing when a customer finds it first.
3. Make the phone number tappable and keep it identical everywhere — site, Google profile, invoices, ute.
4. Then look at llms.txt, and only then at anything paid.

How to check your own site

The audit is free. We read your pages the same way we read these 1,045, and send back the specific fixes — the missing fields, the dead links, the placeholder left in the footer. There is no charge and no obligation; if the list is short we will tell you that too.

If you would rather fix it yourself, start with the five-field block and hand it to whoever maintains your site.

About the method

Homepages only, fetched once, on 21 October 2026, with no JavaScript executed — the same view most crawlers and AI fetchers get. A site that builds its content in the browser may be undercounted here, and a site behind a bot wall may be counted as unreachable. Sites that did not return were excluded from the percentages rather than counted as failures. The underlying scan is kept, so any figure in this article can be checked.