Chinese Domain Bubble Bursting?

Q3/Q4 2015 Domains dropping in .com

[jmccormac]

Looks like the Chinese bubble in .com is beginning to burst. The last few days have seen hundreds of thousands of domain names deleted in .com alone and the .com zone count is now at 127,114,123 domains as of 21/November/2016. The zone count for 01/November/2016 was 127,543,305. It might recover enough registrations for a net positive result on 01/December/2016 but there are more Chinese speculative registrations coming up for renewal/deletion.

Regards...jmcc

[engine]

Domains are cheap, but making something out of them is a lot tougher. I wonder if that issue is finally hitting home.

[jmccormac]

No. This is the Chinese speculative bubble of Q3/Q4 2015 when millions of highly speculative four character, five character and numerical domains were added to the main TLDs and some of the new gTLDs by Chinese speculators. The downside on .com alone is an exposure of at least 3.6 million domain names that are unlikely to be renewed.

Regards...jmcc

[keyplyr]

I wonder how many clones of my site will be lost?

[ken_b]

Were these domains just sitting idle, or were they loaded up with spam stc and dumped onto the active internet?

If the latter, are they too toxic to use?

[jmccormac]

@keyplyr Very few. These were speculative registrations for the most part.

@ken b
From web usage surveys, many of them had no functioning website (PPC or otherwise) and a lot of them were on parked on Chinese domain auction websites. In Western terms, they would not have been useful domain names (four character letter/number combinations, five character letter/number combinations) and also Chinese market specific terms. This isn't limited to the legacy TLDs (com/net/org/info - biz being very low profile in this). Some of the new gTLDs have an exposure on this too.

Regards...jmcc

[accurate]

@jmccormac

Unsurprising as this was predicted.

If .com/.net are going to experience massive drops, new TLDs are going to get massacred.

[jmccormac]

Yes but most people don't seem to realise that the bubble started bursting this month and they are expecting it to burst in January. Some new gTLDs are badly exposed to this bubble but most of them are not.

Regards...jmcc

[accurate]

According to NtldStats China has the most new domain regs. New Gs across the board are going to experience massive drops John. Looking forward to it. :)

[keyplyr]

New Gs across the board are going to experience massive drops

What's a New G?

[accurate]

@keyplyr

New gTLDs, new TLD, new domain, etc. :)

[jmccormac]

Some new gTLDs do not have a major Chinese market exposure, accurate. The ntldstats figures are skewed by taking all new gTLDs as a single set. Some of the larger new gTLDs like xyz and others have gone CN/HK/JP market dominant. The .top is stuffing its zone and seems to want to compete directly with .xyz. The .xyz has based its business model on the Chinese market and is doing promotional deals with registrars to keep the new registrations ahead of the deletions. The renewal rates on heavily discounted registrations are dire and it has already caused a few major implosions as heavily discounted gTLDs had large numbers of these registrations up for renewal. Think of the Network Solutions promotion on .xyz but played out on a number of gTLDs but worse. The Network Solutions promotion wasn't as stupid as some people might think. It had a better than expected renewal rate for some of these domain names. However many of the newer registrars don't have this kind of bluechip customer base. The Chinese market has a very high turnover and some of the registrars there seem to have build their business model on discounting.

The problem is with the new gTLDs, typically the low cost and heavily discounted ones, that are dominated by the Chinese market. The US/CA/EU focused ones are less exposed. The Chinese domain speculation activity is not limited to a short window. What are being deleted in .com and .net at the moment are domains from August/September/October 2015. The big drops, if these are any indication, are yet to come. There are even spikes in 2016 new gTLD registrations that look rather iffy And there's the serious problem with .xyz renewals/drops that will occur in July-September 2017 due to the 1 cent promotion. That was approximately a 3.3 million boost to .xyz.. There's approximately an eight month window of instability with .com, .net, .org, ,biz, .info, .mobi, .pro and some of the new gTLDs.

The second and third renewal anniversaries are always the tough ones for new gTLDs. That's when domainers have to make the hold'em or fold'em decision. What a lot of people expecting large drops on the first year didn't realise is that the high prices on some of the new gTLDs actually encouraged renewals. That's why the renewal rates for some of the new gTLDs were well over 50%.

Regards...jmcc

[RedBar]

Good info as always jmccormac

Insofar as .biz/info are concerned I have now released all of mine except for mycompanyname and I'll continue to retain those at current prices.

I do have a couple of .cn which I have had for 10 years now, mycompanyname and a keyword.

I had a surprise recently when I was able to register a couple of 4 letter .co.uk names never before registered therefore I had a general search around for "other" names and was amazed just how many com/net/org have been released.

I'm glad it's not my money involved in these new ones.

[accurate]

The new domains are going to fail together. The public and press, that don't care about domains, isn't going to differentiate. When most go to sh*t, which is happening, they are all going to sh*t.

I see odd spikes all the time with .XYZ. Unsurprising that Namecheap stopped support. Now a Chinese registrar has dropped .TOP.

New gTLDs are majority owned by Chinese. 70% are parked and not in active use. The rest that are owned by US/CA/EU most are not going to be used. You can zone stuff all you want, that won't encourage active usage.

[jmccormac]

You have to understand the dynamics of the domain business. Otherwise you will end up categorising all new gTLDs as a single set. What you refer to as "odd spikes" are the effects of various registrar promotion deals and the drops of these discounted domain names. The renewal rates on discounted domain names are much lower than those for full price registrations. The xyz business model is based on doing enough deals to keep the new registrations ahead of the deletions.

One of the main motivations for registrars to do these promotions is to upsell the customer and get them to buy hosting or another product at the same time. Without the promotional deals, there's little demand for some of these discounted gTLDs and as a result, registrars stop promoting them.

I don't consider the ntldstats claims on parking levels to be accurate. The percentages vary from gTLD to gTLD and the 70% claim is wildly optimistic for some. For some gTLDs the web usage and development levels are below 10%. A few are even below 5%. Usage on ccTLDs and .com tends to be higher due to their age and markets. The non-core gTLDs (net/org/biz/info) have lower usage. For most economically developed countries, the com/ccTLD axis occupies over 80% of the domain footprint.

The "most of the new gTLDs are owned by the Chinese", thing is misleading. The reality lies in the distribution of the Chinese registrations rather than the volume. Some of the new gTLDs specifically targeting the Chinese market are almost completely Chinese dominated. This is not a surprise. They are Chinese gTLDs.

Some of the new gTLD registries like those that run .xyz and .top have made the Chinese market the core of their business along with heavy discounts and promotions. There are some Chinese targeted gTLDs which are completely dominated by Chinese registrations. But there are many other new gTLDs that have minimal Chinese registration volume. At the moment, the number of new gTLDs is over 1100. Many of them are brands. Others are geo gTLDs and a few are generic. There are zombie TLDs and bubble TLDs there too. Some are in the second year of operation. Others have just launched. The problem is that people think of the new gTLDs as a single TLD rather than a set of TLDs.

The .com TLD zone count (127,034,762) is now close to that for 01 September 2016. A lot of the domains that have dropped in the last week or so are Chinese registrations. And this is only the start.

Regards...jmcc

[accurate]

Saying there are some good new domains is like saying there are goods parts of Aleppo. Most of it has been destroyed by spamming and bogus registrations.

Many web hosts I've talked with have removed new tlds as options. They said there was no demand. There is no upsell happening with new gTLDs.

Even you admit John that the market right now is just country extensions and .com.

[jmccormac]

Read what I wrote. I don't share your view that the market is just .com and .ccTLDs. It is not a single market but rather a small global market with a large number of country level markets. I stated that the ccTLD/com axis for most developed countries represents approximately 80% or more of the domain footprint in those countries. That 20% or so is occupied by other ccTLDs (the adjacent markets effect), legacy non-com gTLDs and new gTLDs.

The classic domainer response to the drops in the new gTLDs is a typical "sky is falling" one. When loads of domains drop in a new gTLD, domainers focus on the drops rather than on the ones that are renewed. For the registry, registrars and hosters, it is a numbers game. The registries need to build awareness and some of them have decided to stuff their zones with cheap or free registrations. Quality-wise, it is very bad. However, the registries and the registrars know something that most people do not: some of these domain names will be renewed. Thus they are doing deals based on a small percentage of the domain names renewing. It can lock a TLD into a boom and bust cycle where the registry has to keep doing discounting deals to keep the new registrations ahead of the deletions. But in terms of volume and awareness, it can work. But there has to be some normal registration volume.

Getting back to the market issue, it is not a single market. It is a set of markets and some of them are mature and some of them are in their early phases. The same applies to TLDs. The generic ones might appear to be global but the registrant base is not.

The .com zone count is now 126,963,922 as of 27/November/2016. That's lower than the 01/September/2016 zone count.

Regards...jmcc

[accurate]

You clearly spend a lot of time on this John and I respect HosterStats and the info you provide.

There are always going to be bright spots, which these new domain companies focus on. In reality most of the city is crumbling. They are pointing out where the development is while ignoring the zone stuffing.

Honestly the experience myself and most people have with new TLDs at this point is getting lots of spam.

[jmccormac]

Honestly the experience myself and most people have with new TLDs at this point is getting lots of spam.

The difference is that I run full new gTLD web usage surveys and I see all the spam, all the zone stuffing, the all holding pages and all the abandoned websites. There are millions of them. But there are also sites being developed. I'm working on the November 2016 survey data at the moment.

Usage on newly launched TLDs is always going to be far lower than established TLDs like .com or the ccTLDs. It can take five years or more for a TLD to get established. There is no doubt that some of the new gTLDs will fail. Some already have as they have failed to get enough registrations. Some registry operators had never run a TLD before and seemed to be very optimistic on their projections. Others spread the risk over a range of gTLDs so that they effectively ended up running one large gTLD with a lot of niches.

What really screwed up this new gTLD round was the simple fact that the market changed when ICANN wasn't looking and ICANN proceeded anyway. When Domain Tasting was killed by ICANN's introduction of the "restocking fee", it effectively destroyed the artificial scarcity that was the basis for the new gTLDs. People simply could not get the domains that they wanted because most of them were being snarfed by Domain Tasting operations and put on PPC. (Domain Tasting actually kickstarted demand for ccTLDs in the home markets and eventually led to .com being eclipsed in some of those markets.) That changed with ICANN's implementation of a fee for Add Grace Period deletions over a percentage of the registrar's monthly registrations. What had effectively been a cost-free business model suddenly involved became expensive. And Domain Tasting at the scale prior to the ICANN action stopped. The new gTLD launches proceeded even though the market landscape had changed dramatically. The launch schedule for the new gTLDs was a complete clusterf*ck of dimwitted planning.

The domain name business actually has seasons. You can see the ebb and flow of demand if you study the statistics. By forcing the launch of hundreds of new gTLDs in a a short period of time and without adjusting for the seasonality of demand, ICANN screwed up an already delicate situation. Not only were new gTLDs competing against .com, the other legacy TLDs and the ccTLDs, they were competing against each other. Throw in this "premium domains" rubbish and self-dealing and it really gets messy.

One of the things I've been working on for a while is mapping market trends in the gTLDs. This demonstrates in simple terms (essentially a spreadsheet and a graph) where a gTLD flips from being a Rest of World dominated TLD to being a CN/HK/JP (the notorious Chip market) dominated gTLD. When that happens, things shift to a completely different gear in terms of registration prices, renewal rates and the proliferation of spam websites. It is a completely different dynamic to a well distributed (geographically) gTLD. The gTLD effectively becomes a Chinese TLD in all but name. I was going to put it on the website but it seems increasingly pointless. Google doesn't like numbers on websites and its kludgefest of an algorithm thinks that the data for hundreds of millions of domain names is "thin content". On the upside, Google is as clueless about the new gTLDs as a dog that sees a car it has been chasing stopped at the traffic lights.

Regards...jmcc

[accurate]

You can definitely go register good dropped domains that dropcatchers aren't getting. You just have to monitor vigilantly.

So you are admitting that a new domain extension that is too heavily CN/HK/JP is bad?

I'd be interested to see the stats John. :)

[jmccormac]

You can definitely go register good dropped domains that dropcatchers aren't getting. You just have to monitor vigilantly.

There are more deleted domains in the legacy gTLDs than there are active domain names. But the database on these goes back a long way. The problem with a lot of these drops is that they are related to businesses or aspirations that either failed or never happened. Some might be useful but the insanity of the dotCom bubble drops takes some beating.

So you are admitting that a new domain extension that is too heavily CN/HK/JP is bad?

I think in terms of markets, usage and economics when considering a TLD rather than simply TLDs being simply good or bad.

A distributed TLD tends to do better in terms of renewals and development as there are different dynamics at work and the unstable elements can be somewhat balanced by the stable elements (Newbies versus veterans. Speculation versus development.). A CN/HK/JP dominated TLD rapidly takes on some unstable characteristics (very wide ranges of renewal rates across hosters), discounting, the rapid growth of clone (pointed to primary websites in other TLDs without a 301/302 redirect)/affiliate lander content as opposed to development. It is almost as if a domain in this kind of a market is a commodity rather than an asset. There is an awful lot of speculative registrations in some of the Chip gTLDs. While some of them will have a rarity value (3Ls, short numerics and keywords), many of them will not. The speculated domains generally end up on auction sites or for sale.

I'd be interested to see the stats John. :)

I've a lot of survey work to get through first.

Regards...jmcc

[accurate]

Noticing even some of the more popular new TLDs are really shedding domains. :)

[accurate]

Interesting data from CIRA's yearly factbook.

[cira.ca...]
[cira.ca...]