Clampdown on TV 'Dodgy Boxes'

I am not sure of how MUSO goes about "detecting" websites. There is little information available.

It is possible to survey all gTLD websites and it is possible to get access to the zone files (the lists of domain names and their nameservers) if the gTLD registries approve your requests for access. The country code (ccTLD) TLDs do not generally provide access to ther zone files. It is possible to detect and categorise gTLD websites. While there were approximately 196 million gTLD websites in January 2026, not all of them are actively developed websites. For .COM, approximately 40% of the domain names can be immediately excluded as they have no website, are parked on holding pages or are listed for sale. Surveying the websites of the remaining domain names would require a crawler operation that would emulate a human user and would execute Javascript. This would load the ads and other content. Extracting the raw text and links from that survey would take some time. Web Usage measurement is done all the time by registries and others to see how websites are being used. Looking for specific indicators in file names is relatively easy. Removing false positives is not.

Identifying the owner (registrant) of a gTLD domain name got much more complex in 2018 after the GDPR fiasco. This meant that personal data was removed from the ownership (WHOIS) records. Some registrars took the opportunity to completely remove all data from these records.

The biggest problem with MUSO is the lack of explanation of its methodologies and data sources. Identifying ordinary websites in gTLDs is easy. Identifying them in ccTLDs is less so.

There is a major question mark over how MUSO "detects" these website visits. If a website is using ads, it might be possible to advertise on it and use this as a proxy metric. The Social Science numbers about visits may appeal to the media but the lack of explanations create a serious credibility issue.

boards fcuk up!

Well they are a private third party. You can tell that the instant you read any of their reports. More self promotion than anything.

They can only see aggregated web traffic in the same way advertisers do.

What MUSO can do is restricted by legal and technical constraints.

If you read their methodology and strip out the hyperbole it appears they manually create a list of 'piracy sites'. Then they use aggregated web traffic and cross reference it with their own manually curated list. Muso use phrases like 'AI Powered Crawling'. Sounds like an extremely sophisticated and deep level of inspection. It isn't. It's a fancy way of saying 'we use ai to perform a vlookup style search and match'.

So here's what they can gather: The number of hits on a site that they've deemed illegal and what country the ip originated from.

What they also do is delist sites from search engines, identify hosting providers and send them DMCA notices.

A few weeks ago a very popular ebook site called Annas archive was taken down but the site has about a half dozen mirrors all of which are listed on their wikipedia page. So it's really just like scooping water out of the titanic with a handbasin.

Here's what they don't know from all their site crawling: No ip tracking, no behavioural profiling, what the user does on a site, how long content was viewed, what was watched if anything, how many files were downloaded, how big the files are.

They also don't have any method to track any data pertaining to IPTV apps, apps like stremio and kodi, cracked software.

Add these to the list they don't know anything about - private torrent trackers, what's app groups, telegram, discord channels, usenet, file sharing sites etc.

Theres huge scope exaggeration here that probably is rooted in MUSO justifying their fees and private funding.

Building a list of "pirate domain names" is not a very good approach. There is a considerable churn in domain names especially with the cheaper $1 or so promtional offer new gTLDs. What happens is that they are registered for a year and then deleted because the renewal fee jumps to about $30 or so. This means that the websites may move to other domain names. The Chinese websites in the new gTLDs have very low renewal rates and in terms of Web Usage, millions of them were used for video style websites. It is not unusual to see 80% or more of domain names in some of these new gTLDs being replaced in a year. Without proper crawling of websites, that list generation can provide very misleading results if these sites in the list are being deleted and replaced. Some of the domain names in those gTLDs use a domain generation algorithn (DGA) to generate thousands of domain names with pseudo-random names. In .COM, there are 5,776 domain names beginning with the letter 'a' that contain the word 'movie'.Some of those will be legitimate.

Identifying the hoster of a pirate domain name may not be easy. If it is using a large web hoster like Amazon etc, there are published abuse contacts that can be used. On some smaller and non-EU/non-US hosting, there is no abuse contact and the ownership details may be false or intetionally misleading. Approximatley 99.5% of the gTLD websites are on just 33,837 web hosters. The problem is in identifying the hosters of the remaining 0.5% because it is often small web hosters and even websites hosted on ISP ranges. There is also a trade in IP address ranges so the ownership can change. Then there is the problem of stale ownership data where the data has been updated for years or even decades.

Estimating web traffic from aggregated sources is not reliable. Some of these sites and operations will not want to be traked. If their methodology is based on building their own list of "pirate domain names" then it has all the flaws of a reporting based approach. It limits the number to the domain names of which they are aware. (It is the detection versus reporitng problem.) Even trying to supplement it with a bit of crawling would need some means of classifying websites. And with multiple languages, that is an exremely difficult problem.

It is reasonable to be cynical about claims like that 229.4 billion visits to websites. Unfortunately, some people will believe it without question. Others will use it because it promotes their agenda.

Not sure what your point is.

My point is the point that I made in the post.

It may be a app to the user but under the hood it's accessing a website/url.

Probably more to do with the increasing use of VPN's due to their increased packaging with IPTV apps and significiant investment in YouTube advertising……..than some media campaign………or any one of several other potential factors.

It's the website that's stopped card payments to get visa/mastercard off their backs. Workaround is easy and only a small bit more in cost. The irony is that anyone buying off those websites knows their stuff rather than going through resellers on facebook.

But…..what if someone has an IPTV sub but only uses it for say BBC/RTE etc. never uses it for the 45000 other channels. Then are they doing something illegal? Does a court in theory have to prove you watched pay streams? The IP is the same regardless I assume.

Missed the part where MUSO was on about covering 192 countries?

The Web isn't all in English. There are other languages used. I've had to write parsers for dealing with non-English website content in crawlers for Web Usage measurement surveys. There was a lot of Chinese video sharing websites and the surveys involved millions of domain names and websites. Do you want an example of a non-English language website or can you simply ask Google to provide you with one?

I know a few users who are on the same service I am with, had issues over weekend as they're still on old apps and them apps shut off basically. All are back up and running now on newer apps. Just a quick app install and login was the fix.