Digital transformation baseline · Article 10
One Site, Three Addresses
The slate for this piece was written before the site had even moved to its current domain, when the finding was one set of duplicate builder pages on an old .xyz family. That finding no longer applies — the project doesn't run that infrastructure any more. What replaced it is a plainer, more current question — how many addresses answer for this brand right now, and does each one do the right thing? Checked one by one, the count was five, not three, and four of them needed fixing.
Tenth in the digital transformation baseline series.
A domain migration tells DNS where to send new visitors. It does not tell every other address a business has ever owned to stop answering. Nothing about an old subdomain still resolving looks broken from the inside — no error, no downtime, no signal of any kind. It just quietly keeps being there, for as long as nobody checks.
Five addresses, one of them the actual site
The canonical address is jshwsolutions.net — production, correct, the only one meant to be found. The other four were each doing something different, and none of it was intentional:
- The
wwwsubdomain served the entire site directly — a full, independent 200 response with its own copy of every asset, not a redirect. The only thing claiming otherwise was a<link rel="canonical">tag in the page's own head, which is a hint a search engine may choose to honour. It is not a redirect a browser obeys, and nothing was enforcing one. - A retired domain from before this project's current build was still fully live: the entire pre-migration site, its own working contact form, its own consent checkbox, its own CAPTCHA — completely disconnected from the current form handler and the current privacy policy governing it.
- A second address from that same earlier era — a distinct build, per this project's own prior audit finding, with different copy and different navigation from the first — was also still live and unredirected.
- A domain this project's own documentation already calls retired was resolving to a bare Apache directory listing — folder names and modification dates, openly browsable, on a domain nobody had reason to be watching.
The www duplicate and the two older builder sites are the same failure: a live, independently-crawlable copy of real content. The directory listing on the retired domain is a different one — nothing meaningful was being duplicated there, but a server left to answer with no page in place will show whatever it finds on disk by default. Both failures share one cause: an address kept resolving after the reason to maintain it was gone.
The through-line so far
A migration moves the site. It does not tell every other address the business ever owned to stop answering.
Why this accumulates, and why nobody notices
Every domain and subdomain a business has ever pointed at a server keeps answering requests indefinitely, by default, until someone explicitly tells it to stop. A platform migration changes where the current site lives. It does nothing to the DNS records, the old hosting account, or the old builder export that were never touched by the move — those simply keep existing, exactly as they were left, with no expiry and no warning attached.
This is also precisely why it goes unnoticed. None of the four addresses above threw an error, returned a broken page, or behaved in any way that would surface in a routine check of the current site. Each one worked perfectly — just not as the thing it should have stopped being.
What actually closed it
Each of the four addresses now issues a permanent redirect to jshwsolutions.net. The www subdomain is handled in this project's own .htaccess, via the same canonical-host rule that has existed, written and ready, since before this domain went live — enabled once production had been live long enough that flipping it carried no risk of a redirect loop to nowhere. The other three sit on separate, older hosting entirely, outside this repository's reach, and were closed directly at that host's own domain-redirect settings — permanent, matching with or without www, covering every path rather than only the homepage.
Verified afterward the same way the finding was made: by visiting each address directly, with cache disabled, and confirming it lands on the real site rather than trusting that the redirect configuration looked correct.
What this doesn't settle
- Anything that already links to an old address keeps doing so. A redirect fixes what happens when that link is followed; it does not find or update the link itself. Old business listings, directories, or a search result cached before today may still show a superseded URL.
- The redirect was verified at the homepage and at a small sample of paths, not exhaustively. A path with no equivalent on the current site correctly reaches a real 404 rather than the old page — checked directly — but not every possible legacy path has been walked one by one.
- This covers what a browser can reach from outside. It says nothing about whichever other services, if any, those four addresses were ever connected to beyond serving pages.
The through-line
The slate called this "one site, three addresses" before the actual count was ever checked. It was five, and the gap between the working title and the real number is itself the finding: nobody had counted, because nothing about an old address quietly still answering ever asks to be counted. A site is only as trustworthy as its least maintained address — and the only way to know how many of those exist is to go and check, the same way everything else in this series got checked.