What is the cost to operate a new extension?

[Mr Dobalina]

I understand there is the application fee for every ngtld application. Beyond that, what are the main variable costs?

For example, Donuts has all the support and backend in place so if they bring on another ngtld, what costs do they incur? It is plausible that a less popular extension can essentially have a free ride because the more popular extensions can make up any budget deficit.

The reason I ask is to look at the actual viability of these new extensions. Even the unpopular ones.

 

[Kate]

Have a look here: https://newgtlds.icann.org/en/applicants/agb

In addition to the application fee, there are various costs, some fixed, some variable depending on volume:

Pay recurring fees to ICANN. In addition to supporting
expenditures made to accomplish the objectives set out in
ICANN’s mission statement, these funds enable the support
required for new gTLDs, including: contractual
compliance, registry liaison, increased registrar
accreditations, and other registry support activities. The
fees include both a fixed component (USD 25,000 annually)
and, where the TLD exceeds a transaction volume, a
variable fee based on transaction volume. See Article 6 of
the registry agreement.

And then your operating costs + various compliance costs + the cost of building or outsourcing the registry backend. You have to shop around if you want to outsource the technical part. It's hard to give an accurate figure, but running a TLD is expensive.

Indeed there are economies of scale can be achieved when running several extensions on one single platform but the application fee + annual fee remain per TLD and they are not cheap.

 

[betthelot]

The lowest you could run would be around 150,000usd if you had economies of scale such as donuts and spread most of admin costs tech support against the other strings.

 

[Mr Dobalina]

Thank you both for your thoughtful answers.

So it sounds like Donuts' business model is to hit major keywords since I imagine there is a lot more than $150,000/extension to be made over the course of many years. The ngtlds were never meant for domain investors in the first place. If a stronger secondary market forms then more power to Donuts but that's not what these were created for.

Edit: .Club however is pushing for a secondary so they can sell their names at a premium for the first year then a steady renewal revenue stream to keep everything running. They have a lot more fixed costs as a percent of sales.

Edit 2: .xyz is pushing as many names as possible to try and catch end users then whatever sticks they will pull in that revenue consistently. They are also pushing for a strong secondary market so they can sell their premium priced names.

 

[FNet]

The vast majority of buyers are the domainers. This is a fact. But many registries got very greedy. So they are using the so called 'premium' price model for many domain names. They will learn the hard way.
After buying a string from ICANN, and if you are already operating many strings, the operations costs are very low except the quarter fee you have to pay to ICANN. So you can run the registry with a fixed price model with a logical renewal fee (less than ten bucks). BUT, you have to make over money somehow to repay the money you owe to the creditors (banks, stockholders or whoever provided you the loan or capital). With today's data the money invested in strings can not be repayed. So this craziness about buying new TLDs who have zero end user demand will stop soon. The forces of the market are always stronger from marketing tricks, consumer stupidity or even regulations. I like some new gTLDs, but let s face it. .horse? .guitar? .cricket? .whoiswho? .theworstextentionsthatcanbeexisted?
It will take some years but market forces will restore normal order.

 

[dordomai]

I think application fees around 200k + 25k/year + staff + backend + capital(to prove that you have sufficient funds)+ legal fees(lawyer etc.)+marketing fees(not required but you might not be successful without)

my guess is you need a minimum of 500k at least and possibly more.

once you have running system in place and run alot of extensions it could be quite cheap 25k/year after investments + cost of running the company.

i think what is costly is the marketing, you need to promote the strings and that can cost you a lot of money.

my feeling is that many of the small registries with a single TLD could be mildly profitable if they keep costs low. Nothing exciting but they can keep their head above the water and see a positive ROI. The question is what happens if more strings get released and demand does not go up? Then it will get harder and harder to make money if everyone wants a slice of the same pie.

so far it appears that larger operators like Rightside and MMX are not profitable. I don't know why this is. perhaps they spend so much on marketing or we underestimate the cost of running a registry.

my feeling is that most of the money does not come from premium sales but from simple renewal fees. It's too hard to sell overpriced premiums even when the registry does it. End-user demand for premiums is low.

That is why they try to give so many domains away. they hope that at least some of them will renew at the standard rate.

 

[FNet]

What marketing? You are joking, right? I hardly saw any registry spending money to seriously promote any string.
You cannot be profitable because the buying of a string is high. I mean, 200k is the minimum, the strings are auctioned and cost much more. The string must be really crap to have only one bidder. Finally, as time passes the demand per string lowers because in reality the strings compete each other. The more strings, the less demand per string. The only winner is ICANN.

 

[Mr Dobalina]

The startup costs for Donuts is probably in the realm of $100-125mln. They own 197 extentions with over 2 million registrations. Once the dust settles they may have a great cash flow. Many of their extentions have over 5k registrations. If they can keep it that way, they'll break even in a few years. Assuming they are ready to drop revenue to the bottom line of course. They might keep spending on new extensions to get as big as possible and that is why it looks like they are losing money. When they stop buying extensions, they can profit and payout investors/creditors. I like the model from a business perspective actually. It just doesn't work for domainers very well.

 

[dordomai]

What marketing? You are joking, right? I hardly saw any registry spending money to seriously promote any string.
You cannot be profitable because the buying of a string is high. I mean, 200k is the minimum, the strings are auctioned and cost much more. The string must be really crap to have only one bidder. Finally, as time passes the demand per string lowers because in reality the strings compete each other. The more strings, the less demand per string. The only winner is ICANN.

I am not sure but I think giving away domains could lose them money. That would be marketing as well.

Also registries need to pay to have their extension featured in a prominent position with a registrar, there are many registrars they could have agreements with. The smaller registries don't do much promotion if any probably.

I do not think that we can release an infinite number of strings. Eventually they will have to stop. The best strings were already released, more and more crap would need to be added.

I don't think strings like .poo or .crap would turn out to be profitable the number of words that can be profitable must be limited. The good ones are all taken.

 

[FNet]

I think that you don't t put ICANN s fees in the equation. It is 197 times 25k. Also In these numbers many sold just for pennies (promotions). And if they buy new strings they will cannibalize their sales/renewals of old strings.
Things are not so simple as they look. You can not create demand if it is not really exists. I mean, to create more than the consumer needs.

 

[FNet]

Doromai, you know that a domain name for the registry is just some data in a database. The only cost for a domain, if you already own the extention, is the fee that ICANN asks per domain. Which is very low. No the heavy cost, if you have the infrastructure, is to buy the extention.

 

[dordomai]

I think that you don't t put ICANN s fees in the equation. It is 197 times 25k. Also In these numbers many sold just for pennies (promotions). And if they buy new strings they will cannibalize their sales/renewals of old strings.
Things are not so simple as they look. You can not create demand if it is not really exists. I mean, to create more than the consumer needs.

i think a lot of the demand is not real demand but either TM holders protecting their mark or speculators hoping to resell the domain.

 

[FNet]

TM is real demand for the registries. You see, the company will keep renewing them. If they are just redirecting is irrelevant. Speculation is another matter ...

 

[Mr Dobalina]

If you divide 2 million renewals into 197 extensions you're looking at 10,000 per. Renewals are typically between $10-40. They are looking at $100-400k/year for each extension. The 25k a year fee is the only big variable cost that anyone has mentioned. That is about $5mln per year. With 2 million registrations with $10 to 40 renewals, they'll pull in 20 to 80 million. That should cover the backend fixed costs and the 5 million fee per year.

 

[NigelD]

Have donuts ever given any reason why every extension is run by a separate LLC?

 

[Mr Dobalina]

Have donuts ever given any reason why every extension is run by a separate LLC?

Interesting. If they ever shut one down it would protect the rest of them. Like if they "bankrupt" one the creditors may not be able to go after other assets. Idk enough about business law so it is just speculation.

 

[FNet]

If you divide 2 million renewals into 197 extensions you're looking at 10,000 per. Renewals are typically between $10-40. They are looking at $100-400k/year for each extension. The 25k a year fee is the only big variable cost that anyone has mentioned. That is about $5mln per year. With 2 million registrations with $10 to 40 renewals, they'll pull in 20 to 80 million. That should cover the backend fixed costs and the 5 million fee per year.

2 million registrations contain many domains with promotion prices. Off course they have sold some premium domains, but usually this is an one time buy. Keep in mind that the trend to mediocre extensions will disappear. You can not repay the aquasition of the string with just ten thousand domain names per string. You can break even with operational costs maybe. You have other costs except the ICANN s yearly fees. Not much, but they add up. The marginal profits, if any, can not cover the aquatisions.

 

[NigelD]

Yes - Mr Dobalina - that's the first impression it gave to me - or perhaps they're looking to sell off individually sometime in the future

 

[FNet]

NigeID, this very simple I believe. As a last resort, they can bankrupt the company only for the specific extentions with big losses.

 

[FNet]

You don t need separate companies to sell assets. On the other hand, maybe it was easier to them to gather capital or take loan etc with this scheme.

 

[Mr Dobalina]

Domainers aren't their target audience. That's why they don't let their premium names go for $10 just to be sat on for 20 years. They want real end users representing their best names. I think the model works great. They wouldn't be doing it if their projections showed otherwise. They're not in business to make us money.

 

[Kate]

Domainers aren't their target audience.

So they say. Of course, they would rather have real end users than domainers, but registries are nonetheless very dependent on domainers and speculators of all kinds. They don't choose their customers, although they can adjust pricing so as to appeal or deter certain types of customers. For instance spammers will always be attracted by low fee (<= $1) extensions.
The parking rate is very high in many TLDs, in some TLDs it's well above 50%. Clearly, they are not all being bought with a development plan in mind.

That's why they don't let their premium names go for $10 just to be sat on for 20 years. They want real end users representing their best names.

It's not the main reason. The reason is, they want a return on their investment.
Ironically, hoarding names like they do is the best way to ensure that these names will never be developed and will collect dust for 20 years

if they really wanted end users, they could award domain names for free or for a nominal fee on a RFP basis: the end user must present a development project that will enhance the TLD and in exchange they get a 'great' domain. It's been done with .mobi but in the past but I don't think a lot of names were awarded through RFP. Anyway, it didn't make a lasting impact.
Reality is, end users are not lining up for those names.

As for marketing: the registries are not doing much because it is expensive and probably will not translate to a significant increase in sales. Not cost-effective.
Instead they rely on the registrars to promote TLDs in front of their customers. Since there are so many TLDs competing for a share of small pie, I suppose there has to be negotiations and registrars have to be incentivized to prioritize certain extensions over others...
Registrars also have to make a profit and there has to be some volume.

Namecheap dumped .xyz if I remember well. I can imagine that the reason was the low margins on penny domains vs the clerical/legal work dealing with abuse (spam, malware).

 

[Mr Dobalina]

if they really wanted end users, they could award domain names for free or for a nominal fee on a RFP basis: the end user must present a development project that will enhance the TLD and in exchange they get a 'great' domain.

This would be the cheapest form of advertising. I really doubt Google is paying whatever ungodly high renewal rate abc.xyz would have. That had a big impact on the extension. I was under the assumption that the premium priced names are only premium for domainers. If you are an enduser with a solid marketing budget and development plan, you could contact the registry and get your hands on one of those for cheap. Idk for sure but it is just good business.