I started tracking UK business domain changes — it turned into something much bigger

[MELIAM]

I've been working on a data experiment that originally started because of my interest in domains.

I wanted to see whether I could identify businesses that had changed from one domain to another by comparing successive monthly UK business/place datasets.

I compared July → August 2026 and found 14,044 own-domain → own-domain changes.

But once I'd built the comparison, I realised there was quite a bit more hiding in the monthly changes.

The same comparison found:

• 56,976 existing-business records with a meaningful change
• 90,234 newly observed places
• 11,497 genuine phone-number changes
• websites being added
• websites being removed
• places no longer observed

Rather than just recording that something changed, I've kept the useful business record with it — business name, category, location, postcode, phone, current website and the before/after values where applicable.

I've ended up building a small site around the experiment, ChangeCapture.co.uk/Samples, where I've put up 50-record samples of each change type so people can see what the data actually looks like.

I'm not really posting this as a sales pitch — the full files aren't even available yet. I'm interested in whether other people can see uses for this data that I haven't thought of.

Obviously, the 14,044 domain changes caught my attention as a domainer, but I'm starting to think the newly observed businesses and website-added/removed data could be just as interesting.

What would you look for in data like this?

 

[nicenicnicenic is verified member.]

The actual before-and-after domain swaps are honestly what I find most fascinating. Watching an established brand upgrade from a clunky multi-word URL to an exact-match, or move from a .co.uk over to the global .com, adds so much context that raw sale numbers completely miss. It really highlights the commercial justification behind why that specific buyer pulled the trigger.

 

[MELIAM]

That's exactly the sort of thing that got me started on this in the first place.

The before/after is what I found interesting too, because instead of just seeing that a domain is now being used by a business, you can see what it replaced.

With 14,044 own-domain to own-domain changes in one monthly comparison, I'm wondering whether there are patterns in there — .co.uk to .com, long domains to shorter ones, rebrands, exact-match upgrades, etc.

I hadn't thought about matching those changes against historical domain sales though. That could add another interesting layer to it.

Once I've got several months rather than just July → August, I think the history could become more useful than any individual monthly snapshot.

 

[Mr. DeletedMr. Deleted is verified member.]

Could be interesting, not just for the uk but other countries as well if you ever expand but I’d focus on the uk first to refine the process.

 

[MELIAM]

Thanks, that's pretty much how I'm looking at it. The comparison process should work with other countries as well, but I think there's enough to learn from the UK data first.

I'm particularly interested to see what becomes more useful once I've got several consecutive months rather than just the July → August comparison. A proper history of changes could be more interesting than any single monthly dataset.

I'm also interested in what happens to the old domains when businesses move to a new one - particularly whether any of those old domains eventually expire and become available to register. With 14,044 own-domain → own-domain changes in this comparison alone, that could be an interesting dataset in itself.

 

[Mr. DeletedMr. Deleted is verified member.]

https://shoplocator.williamhill/ That’s a sponsored tld.

WilliamHill.com was registered in 1995. Another question you could look at is when did the “new” website address get registered, was it recently sold, etc. because obviously a company that chooses to have a STLD is large enough that it has a large amount of its ip locked up.

 

[MELIAM]

https://shoplocator.williamhill/ That’s a sponsored tld.

WilliamHill.com was registered in 1995. Another question you could look at is when did the “new” website address get registered, was it recently sold, etc. because obviously a company that chooses to have a STLD is large enough that it has a large amount of its ip locked up.

That's a very good point. Domain registration age hadn't occurred to me as another layer to add to the change data. It could be particularly interesting to distinguish a move to a newly registered domain from a move to a long-established domain, and also check what subsequently happens to the domain they've moved away from. I'll have a look at whether I can add that to the analysis.