Home / Blogs

How Big Is the Domain Business?

John Levine

When you're standing close to ICANN, the domain business may seem pretty big, but when you stand farther away, not so much.

Verisign's revenues are about $1 billion/year. The .COM and .NET top-level domains together have about 150M names. The next biggest gTLDS are .ORG with 25M and .INFO with 12M. The biggest new TLDs are TOP with 2.9M and .XYZ with 1.8M, with both bloated by firesale prices. The rest are smaller, mostly much smaller.

The biggest registrar is Godaddy whose revenue is $2.5G/yr which includes a lot of non-registrar revenue. So I'm guessing that the total registry and registrar business is in the range of $3 to $5G (billion), growing slowly.

Meanwhile, Amazon's cloud revenue this year is about $20G, Google Cloud and Microsoft Azure both about $10G, each of which are likely to double by 2020. Facebook's revenue is around $40G, Google's is $100G, jd.com (the Chinese e-commerce company) $55G. Amazon's overall revenue is $180G, but most of that is stuff, not bits. The total Internet business is on the order of $1 trillion, with the domain business being less than 1% of that. So let us not get swelled heads.

This also reminds us that even though $100M to get control of .WEB seemed like a lot of money in the domain world, that would have been 0.15% of the cash Google currently has on hand.

By John Levine, Author, Consultant & Speaker
Follow CircleID on
Related topics: Domain Names
SHARE THIS POST

If you are pressed for time ...

... this is for you. More and more professionals are choosing to publish critical posts on CircleID from all corners of the Internet industry. If you find it hard to keep up daily, consider subscribing to our weekly digest. We will provide you a convenient summary report once a week sent directly to your inbox. It's a quick and easy read.

I make a point of reading CircleID. There is no getting around the utility of knowing what thoughtful people are thinking and saying about our industry.

Vinton Cerf, Co-designer of the TCP/IP Protocols & the Architecture of the Internet

Share your comments

Incorrect Stats George Kirikos  –  Aug 15, 2018 5:21 AM PST

.org doesn't have 25 million domains registered as is claimed, and .info doesn't have 12 million domains. Sites like DomainNameStats.com have current figures (roughly 10.3 million and 5.1 million respectively).

Also, one has to be careful about double-counting of wholesale revenues. i.e. domain name revenues of a retail registrar like GoDaddy include their input costs that go to the registry operators (VeriSign, for .com), plus their margin/profits. So, one has to make sure to only count the revenue once (otherwise, registry operator revenue can get counted multiple times).

However, once you include ccTLD revenues and the secondary market for domains, the above estimate of the annual turnover might be in the correct ballpark.

You're right, there was a bug in John Levine  –  Aug 15, 2018 8:32 AM PST

You're right, there was a bug in my name counting scripts.  ORG has about 10.3 million, INFO has 5.05 million. But the result is still the same, Verisign's COM and NET are bigger than everything else, even with fairly optimistic estimates for ccTLDs, and the whole business is a rounding error on the whole Internet industry.

If the speculative market for domains is a substantial fraction of the domain industry, which it might well be, it reminds us how broken the whole thing is.

>it reminds us how broken the whole Charles Christopher  –  Aug 15, 2018 1:04 PM PST

>it reminds us how broken the whole thing is.

Please define "broken" as you have used it in this context.

The cost of electricity and natural gas to a large successful manufacturing facility would hopefully be "rounding error" to its total costs, especially considering employees and raw materials.

Try running it without that rounding error energy source ....

Take away Google, does the internet still work? Yes.

Take away domain names, does the internet still work? No.

When we take away domain names everything is still there, its no different than taking away all the street signs in a city all the streets homes and businesses are unaffected. .... But the internet will not work anymore, and local commerce will fail when addresses are no longer physically discoverable, those cheap street signs are actually kinda important after all even if their relative price/cost is sometimes hard to understand.

The speculative angle is sometimes overhyped, John,It John McCormac  –  Aug 16, 2018 6:47 AM PST

The speculative angle is sometimes overhyped, John,
It is a small part of the overall market. Most domain names are registered but not in use for active websites. With the legacy gTLDs, some of the active web would be be around 25% of the registered domain names. The rate varies and some of the new gTLDs are not doing well on web usage but discounting has lead to very low renewal rates in some of the new gTLDs. Discounting, over the last fifteen years or so, has changed the renewal rates and though the blended renewal rates for some of the legacy gTLDs are strong, (>70%) some of the one year renewal rates are around 15% lower. Domain names are often treated by registrars and hosters as a means to upsell the customer to more lucrative hosting packages and other services. One factor that might be missing from your analysis is the monetisation of undeveloped and unused domain names. Some registrars will automatically park these undeveloped domain names on their own PPC parking operation. It is possible that they generate some revenue or the registrars would not be doing it.

To much speculation can kill a TLD. The best example of this was the .EU fiasco where the European Commission, its advisors and the registry hadn't a clue about they were doing. Massive speculation from non-EU actors basically killed the TLD in the English language market (Ireland and the UK). It never recovered and it is now considered a brand protection registration option rather than a primary brand identity for registrants. Things changed after that with premium domain names being held back by the registries (.MOBI, .ASIA etc) and auctioned off for large sums. The same thing has happened in some of the new gTLDs but with some of these gTLDs, the retention of large numbers of premium domain names by the registries killed off any interest in the gTLDs. Speculation does give a land rush boost to a TLD and increases awareness but taking that out of the equation leads to lower registration numbers. However, without development and usage, a TLD will become a zombie TLD where domain names will be flipped for the price of a hamburger.

The .WEB issue is interesting because it is, perhaps, one of the few strings that could be a strong (possibly 10M or more registrations in the first few years) competitor in the existing COM/NET/ORG TLD market. Both .NET and .ORG are struggling in many country level markets and there has been a long-term decline with .NET that has been masked by the Chinese speculative bubbles of the last few years.

The .TOP and .XYZ gTLDs have, along with some other new gTLDs, different dynamics to the global, legacy gTLDs. They are dominated by registrations from the Chinese market. They are not global but rather Chinese gTLDs in all but name. Some of the non-core legacy gTLDs have also used discounting in the Chinese market to drive registration volume with the inevitable bust following a year or so later from these booms in new registrations.

"The speculative angle is sometimes overhyped,"And yet:"To Charles Christopher  –  Aug 17, 2018 10:51 AM PST

"The speculative angle is sometimes overhyped,"

And yet:

"To much speculation can kill a TLD. "

Which is it? From EURID itself:

https://eurid.eu/media/filer_public/d3/85/d38538c1-dac5-4e28-a779-cc31b8259697/boek_dot_eu_v05.pdf

"The relevant Ec legislation does not allow speculative and abusive registration."

The document acknowledges the bureaucracy of "verification" costs and procedures were primary causes of .EU registration issues, not speculation. It seems registrants like "just works" behavior, where they just pay a registration fee and just get their domain name and they just get to keep it. It is just to simple.

In the crash of 2007, speculators were blamed, they always are. And yet when one looked under the hood, ignoring inability to pay debt, and the MERS system (which sidestepped 400 years of property rights accountability) were ignored.

Its always the speculators fault, even when its not.

So long as arbitrage exists, there will be speculation. In fact, the commodities demonstrate the benefit of speculation, and the domain industry is the same:

https://www.investopedia.com/financial-edge/0412/the-role-of-speculators-in-the-commodity-market.aspx

"For starters, speculators serve an important role in [domain name] markets. There are many legitimate purposes for trading in [domain names] or by actually holding [domain names]. Companies do it to hedge against rapid rises and falls in price, and farmers may do it to try and offset swings in the weather or price swings that result in rapid shifts in demand. These market participants need someone to take the offsetting positions to their positions, which is exactly what speculators do. In this respect, they provide the markets with an invaluable service and help encourage liquidity, which serves to encourage the ability to easily make trades with reasonable commissions."

In other words, as I said earlier, they make domains available to business by holding them out of use ["offsetting positions"] and perceiving no emotional or other value beyond the investment value (typically derived from arbitrage). I make no claim the analog is particularly precise, but this is a just a break from my work and it is a reasonable analogy to see the point.

Be careful what you wish for when you believe speculators should not be part of the domain name industry ...

That was mainly an "aren't we great" book John McCormac  –  Aug 17, 2018 12:07 PM PST

That was mainly an "aren't we great" book from Eurid. It papered over all the bad things that happened such as launching without a functioning transfer system and the registration system falling over after a few hours on the land rush day.

The speculative angle gets overhyped because people don't understand the extent or function of speculation in new TLDs. Ordinary people think that just because someone got there first and registered a domain name and are not using it that the registrant is speculating or worse.

The EC, its advisors and Eurid were completely unable to anticipate what was going to happen because they hadn't much of a clue about the domain name business or what was happening at the time. The didn't understand that that the .EU was really a generic TLD in all but name with approximately 27 different languages and a multitude of markets. The Eurid people came mainly from a ccTLD background rather than a gTLD background. They assumed that the market would play by their rules. The Eurid people were totally unprepared for the firestorm into which they walked.

Creating strong legislation and then not enforcing it just doesn't work. There was massive speculation and, worse, cybersquatting in the .EU land rush. Front companies were being used to register tens of thousands of domain names in .EU ccTLD. Eurid was shamed into taking action against the Ovidio group and then bungled that too. It suspended about 73K of its domains and then had to reinstate them after some court in Belgium found that they didn't follow the correct procedures.  The effect of all this over speculation and cybersquatting effectively killed the .EU ccTLD in the Irish and UK markets. It didn't help it in other European markets either. It also had the unintended effect of kickstarting growth in many of the real ccTLDs in the European Union because registrants considered their local ccTLD to be more trustworthy than a pseudo ccTLD.

It took nearly five years for most of the speculative and cybersquatted registrations to wash out of the zone. A properly regulated and properly run .EU could have been an major competitor to the .COM in the EU markets. Now it is struggling to break 5% of the country level domain name footprints in the EU member states. The blame for the .EU fiasco lies with the European Commission and its advisors. If these people had a clue about what was happening with the legacy gTLDs (this was when Domain Tasting was kicking off and the blogs were filled with posts about how .EU was going to be the next big thing), then .EU ccTLD would not be struggling in most EU markets.

I consider speculation to be a necessary part of the domain name markets. It drives growth in the first six months of the land rush period in most new TLDs. That's approximately the length of a land rush period in a new TLD. After that the growth patterns tend to settle down. Without the land rush period driving awareness and registrations, growth in a new TLD is massively lower. This can be seen in some of the new gTLDs where the registries held back numbers of generic keyword domain names in the new gTLDs so that there was effectively no land rush attraction for anyone. A few of these new gTLDs flatlined. Speculation drives awareness and, if handled properly, drives registrations and helps development.

In the .EU ccTLD, the percentage of domain names on sale is quite low. There is development and usage but a .EU domain name is more likely to be a brand protection registration rather than a primary brand website. It was lucky to survive and Eurid seemed to have learned from its mistakes.

As regards the statistics on speculation and land rushes, I don't need to quote some investment encyclopedia. I can quote the actual numbers and graph the effect. Because I track domain names at both domain name and nameserver/hoster level, it is possible to measure the activity of some of the domain name sales and auction hosters/registrars in various TLD markets. The echo of the land rush (the deletion spike around the anniversary of the landrush start in 2006) in .EU could still be detected right ten years after the TLD launched.

In a strong TLD with good prospects, speculation is only a small part of the registrations base and it generally targets generic keywords. The bulk of registrations will be from people who intend to use the TLD whether for e-mail or web deveopment and brand protection. Brand protection isn't simply trademark owners protecting their mark. Small businesses will protect their brand if they think that the TLD has a future in their market. If those small businesses aren't aware of the TLD, then they won't register domain names and won't develop websites on them. Growth and awareness in a TLD arises from these small business websites rather than big brand businesses but that's a completely different topic.

>I consider speculation to be a necessary Charles Christopher  –  Aug 17, 2018 2:42 PM PST

>I consider speculation to be a necessary part of the domain name markets.

I am not sure it is necessary, but agree without it there would likely be far more problems for others to complain about.

That was the primary point I was trying to make. Thank you.

And please note, I have been at this for 20 years and my own "stash" of zone files, whois scans, and DNS scans, is many terabytes, compressed, in size. I'm not missing very much data regarding this issue either.

It is a complex issue. John McCormac  –  Aug 17, 2018 3:23 PM PST

Some TLDs can get away without having to use speculation to build awareness. These are generally niche TLDs or highly regulated ccTLDs. The problem that the highly regulated ccTLDs encounter is that there are always alternatives and that's why the main part of most country level markets is around 80% .COM/.ccTLD. As a market matures, the .COM goes legacy and the bulk of new registrations switches to the ccTLD.

The odds are very much stacked against any new TLD and it generally needs all the help it can get. Even with a mature TLD like the .EU has to work hard to build usage and awareness. The Content / No Content / Redirects percentages were 17.18% - 53.17% - 29.65% (including 2.75% HTTPS redirects) on a May 2018 web usage survey. In an ideal model, all registered domain names in a TLD would be used for e-mail or for a website. It just doesn't work like that in reality. The rise of shared hosting with automated setup of domain names has improved the overall number of working domain names in TLDs but these domain names are not necessarily active in terms of e-mail or website usage. And the reliance on discounting in some of the top new gTLDs has made zone file counts a somewhat less reliable metric than it was previously.

I was going to comment on that Charles Christopher  –  Aug 17, 2018 4:05 PM PST

I was going to comment on that but did not.

There has always been a certain "cockoff" regarding reg counts in the industry.

I have gone through the COM/NET zone file by eyeball twice. How many are so silly as to do such a thing? But I encourage anybody that actually wants to discuss this stuff to do so. The COM/NET zone file is mostly computer barf, it makes an impact once you see it for yourself. It begs LOTS of questions .....

I also have watched land rushes very closely and watched large blocks of domains delete at unexpected times. The clockwork of it suggested to me they they inflating their own reg numbers to make it appear there was more interest than there really was. As a registrar, you only know what is in your account based on the provided reports. You can't know the registry created "shadow" registrations in your account unless you run the zone file to look at the whois, and why would any registrar bother?

Speaking loosely:

Reg counts really don't matter, other than to investor relations. So long as I can register a domain, it keeps working, and I don't get extorted on future renewal fees, I as an end user don't care what the reg count is. Yesteryear's one to one relationship between TLD and registry biased things more towards individual regs counts mattering. Today the nTLDs are part of a pool in a given registry. The "failing" ones being picked up by those backends. Once you've been doing this the cost of adding another TLD is about $0.00.

Many years ago via a friend, long before today's nTLDs, I obtained a quote from one of the large backend providers "What if I owned my own TLD and wanted you to backend it?" the number was stunningly low, and I promised not to share it.

One example I like to keep in mind is .COOP. Its reg fee is reasonable and its reg count is very low. Yet I never hear doom and gloom comments regarding it, only others TLDs. Back to reg counts don't matter, unless P&L;is the only concern.

>As a market matures, the .COM goes legacy and the
>bulk of new registrations switches to the ccTLD.

As I recently pointed out to a peer:

"As time goes on the bias of us old farts means nothing. The young squirts are growing up without our bias and simply don't care about the TLD so long as it make sense and is memorable" ... Classical Plato's Cave.

Recently Versign has been sending out a lot of emails trying to get attention for itself. At some point 6 or 7 chars in an obscure TLD is a lot more customer friendly than 24 chars in .com.

>but these domain names are not necessarily active in terms of e-mail

Agreed. The industry has done a great job at taming website authorship, but not so good regarding email. I have some pretty cynical views as to why that may be so, but I will not distract the conversation further.

>And the reliance on discounting in some of the top new gTLDs has made
>zone file counts a somewhat less reliable metric than it was previously.

Like I say, I don't think much of zone counts. But the changes in zone counts I agree says a lot.

For example the promotion that .INFO did which started these promotions to begin with, if memory serves. Enom took advantage of the Afilias free domain offer and registered the adjacent .INFO for the .COMs (and others?) in their customers accounts. The spammers nabbed domains to causing Google to "ban" .INFOs for some time. It was a mess but something instructing happened. From memory Enom registered 2.5 millions .INFOs. The next year when renewals came 1.25 million here renewed. At the time it can be assume that all those renewals were for domains in speculator's accounts. That was a big statement for .INFO, as big of a mess as they caused it showed interest once people had the domains and could see traffic and get some PPC income.

I never thought that 50% of those free regs would stick. Its the equivalent of PendingDelete's, just because it was registered others think is has value. Thus the regs counts are naturally increase just because newbies come in with this attitude and do some holding up of the reg counts especially in .COM.

But back to the original post. Domain Names are the foundation of everything else. The fact that it's market share by value is something that we should be happy with and not critical of.

The reference I like to use is a smart motivated single mother. For $15 a year she has the opportunity to build a home based business, give her kids time many can't, and compete with any company in the world. That is because the cost of that domain name is so low and her website cost go up very little as her revenue/traffic goes up. And THAT is something we need to be proud of and not criticize or look down on.

Its that low cost that provides the opportunity many in history have dreamed of but few have ever really had.

# 13 Reply (max. reply level reached)  |  Link  |  Report Problems
Having domain names as identifiers is a John Levine  –  Aug 15, 2018 1:43 PM PST

Having domain names as identifiers is a dandy idea. Trading domain names as speculative trinkets adds negative value for the users.

>Trading domain names as speculative trinkets adds Charles Christopher  –  Aug 15, 2018 2:15 PM PST

>Trading domain names as speculative trinkets adds negative value for the users.

Only the ones after the one that buys it and creates a website with it.

The part that people miss is removing "speculation" will never increase availability, it will in fact DECREASE availability.

Its the speculator holding the domain for years that allows the "user" of great interest to actually obtain the domain name. Had there been no speculation its more likely someone would have build a site on the domain, which apparently has value, thus actually DRIVING UP the cost relative to the speculator who perceives no value in the domain beyond an investment to sell. Speculators, by definition, place no emotional or other value on a domain name beyond it being an investment to sell.

The value of a domain name to an "end user" in fact, rarely include the domains intrinsic value. They actually see the domain as an expense and only an expense.

A business is not likely to pay more for a domain that the value of the company it plans to build on it. And thus if there is already a website, or business, obtaining the domain will have a price that pays for the work/business which is of no value to the buyer with other intentions.

Like it or no, speculators are the UNPAID sales force of registries. Beyond the domain sale they also, in various ways, drive interest in the various TLDs. Registries now recognize this as former "speculators" have moved up to the registry level .....

Oh, I don't think there is any John Levine  –  Aug 15, 2018 2:25 PM PST

Oh, I don't think there is any realistic way to make speculation go away. But that doesn't mean it's of any social benefit.

Like it or no, speculators are the UNPAID sales force of registries.

Why should anyone care? I have nothing against registries but I see no reason to encourage people to buy names that they aren't going to use as identifiers.

>I see no reason to encourage people Charles Christopher  –  Aug 15, 2018 2:38 PM PST

>I see no reason to encourage people to buy names that they
>aren't going to use as identifiers.

I don't either. And I don't recall ever seeing that happen.

I've only ever seen people using their freewill to decide to do this, many having great success.

For others its a hobby, and better (actually learning economics) than pretending to be vegetable in front of a TV.

>But that doesn't mean it's of any social benefit.

I suspect those who have purchased domains from "speculators" might have a different view. Clearly they benefited otherwise they never would have paid or even inquired as to price.

I always wonder about those who tell me what my "social values" should or should not be .... I left home at age 19 so I could decide those for myself. :)

To post comments, please login or create an account.

Related

Topics

DNS Security

Sponsored byAfilias

New TLDs

Sponsored byAfilias

Domain Names

Sponsored byVerisign

IP Addressing

Sponsored byAvenue4 LLC

Cybersecurity

Sponsored byVerisign