Hacker News new | past | comments | ask | show | jobs | submit login
How many .com domain names are unused? (singaporedatacompany.com)
275 points by jekor 47 days ago | hide | past | web | favorite | 207 comments

I'm not really well informed about the politics of DNS at the top, but I've always had the impression ICANN has been grossly negligent / corrupted over the last twenty years.

At the end of the day there is profound capital interest in making name records unfair and noncompetitive in order to rent seek fortunes off the Internet and I see no indication ICANN is anything but compromised by those interests in the way they hand out TLDs and enable registrars to operate as for profit corporations. Nothing about a database translating human readable strings to IP addresses should be a damn profit center.

I'm surprised the EFF, FSF, etc haven't ever shown interest in replacing ICANN's registrar model with something that enables registration without squatting, doesn't let anyone rent seek a fortune off of naming things, or create profit incentives that leads to the mess where some things are named strangely because someone is trying to extort money for it.

There are some edge technologies like namecoin and IPNS that seem to be trying to approach this problem in novel ways but its a shame we don't have a fair consensus model on naming things that actually works.

This is mathematically simple to fix.

Charge US$0.50 per month ($6 per year) for every domain reservation regardless of who is reserving. The fee must be charged per month and not per year, which intentionally results in greater maintenance overhead.

I don't care if you are a registrar, government, charity, school, or whatever. US$0.50 each and every month for each and every name. Registrars can slap whatever markup they want on top of that. The free market will lower the price where competition matters and raise the price where demand is huge for a specified name.

If you are holding 100,000 domain names and doing nothing but serving ads on them you better hope you are generating greater than $800,000 (conservative estimate) in advertising revenue off those domains. That number is the cost of reserving the name plus your business expenses to maintain those names and the associated advertising content. I find this scenario to not likely to be profitable.

Unfortunately this solution won't work in reality, because everybody will bitch about their business interests and self-justifying political reasons.

All you've done is increase the registration fee by $6 per year as the registrar will happily handle the monthly aspect. The money will just increase the profit of the rent seekers who control DNS.

I don't know, it might work. For me as an individual, my costs go up $6 per year. For a domain squatter, their costs go up hundreds of thousands.

It makes sense to me, like a property tax. You are basically renting a valuable natural resource from everybody else, so you pay a little bit to repay them and prove you're getting some use out of it.

not necessarily, if it reduces the number of squatted domains.

The benefit of increasing the registration fees is actually an concept that the Swedish registry is currently following. They found that by increasing prices the new registrations did not suffer, and the renewal actually increased, with the theory being that because people pay more they value the domain name ownership more.

A cheap registration price make sense when you look to expand new registrations to a fresh market, but what most parties in this industry want now is a higher rate of renewals. For that a higher price can actually be beneficial for both registries and registers.

most registrars already charge a yearly fee per domain. Does ICANN not already charge a fee per domain?

ICANN charges a ridiculously low maintenance fee to the registrars at something like $0.18 per domain per year.


Why wouldn't the monthly renewal be automated away?

There is liability in frequency of transactions that requires human monitoring. The more frequent a transaction is the more liability there is. If the charges were automated would you risk the presence, trademark, and health of your business that your domain will always be there due to that automation?

> There is liability in frequency of transactions that requires human monitoring. The more frequent a transaction is the more liability there is. If the charges were automated would you risk the presence, trademark, and health of your business that your domain will always be there due to that automation?

I don't understand..large companies currently outsource domain management to others, effectively relying on automation for their business. Same goes for any business that pays any bill with automation. Do you also cut a check monthly for phone and internet service?

Bills and accounting at any large company are managed by a team of accountants with the support of automation.

The difference is that utilities do not expire. If you don’t pay your bills you may lose service temporarily, but they won’t delete your account and reprocess your facility. A domain name is essentially leased property. If you let it expire not only will you lose it but then somebody else can take it and you won’t get it back. The real world equivalent is failing to pay mortgage or property taxes on your house, except consumer protection laws require a large number of failures (sometimes years without payment) before they can take ownership.

Advertisers often pay 3€ - 5€ to google for one click. If your 100,000 domains are not complete garbage, you should easily get a million out of it.

That is not really accurate. I used to work with this madness. Google ad displays/impressions are the result of transactional data micro-auctions. If there is high demand for a particular and well known data market, like Las Vegas hotel booking, the ad rate will be more.

The auction rate is determined by the winning bid rate of the competing content providers. In the Las Vegas hotel booking example the hotel properties might pay up to $16 for their ad display along side Google search results and up to $24 for a click through. Most ads, though, are rated at pennies. Anything that can achieve bid rates over $0.50 is a gold mine.

So, yes, some providers will pay far more than $7.00 for a click, but that is extraordinarily rare. This is even more rare when source and context are taken into account. Impressions that look organic, like on top of search results, are worth more than those that look like spam. I highly doubt ads that look like spam on a parked domain are doing more than pennies, likely fractions of a penny. The commonality of low rates and weak impressions on online advertising is also why Facebook shares aren't worth more than they are given the traffic and dedicated eyeballs that they have.

In some industries, you may get those levels of CPC on a high value web property, but a parked/error/empty domain will not be valued by a reputed ad network in the same way as an Alexa Top 10000 website would, the CPC/CPM will not even be close.

> 3€ - 5€

Isn't that for 1000 clicks?

Apparently not


> Below are the 97 countries represented in the map, in order of highest average cost per click to lowest average cost per click, compared to the US average CPC, which is between $1 and $2 on the Search network.

Now I'm unsure if each individual click is actually billed for that much. Or if that's a sum / average of all the advertising costs until someone finally clicks.

It is what Google gets when somebody/something clicks on an ad in SERPs.

No for example in the drink space in which I work the CPC for popular cocktail terms can easily be in the 1-2$ range

eg Negroni Recipe is $2.4 in the US

Couldn't agree more. Not sure why the task of operating DNS' isn't the domain of governments the world over. It's kind of like operating a post office. Post office's are the ones who have the routing information and capability to route physical messages and packages. They are still operated on a for profit model, but they are also generally very well regulated. Same thing should apply with registrars

When you scan for uncorrupt institutions, how the hell do you end up on "governments the world over"?

Because some of us don't believe governments are inherently evil?

Almost no one believes governments are inherently evil, but that doesn’t mean many, or most, aren’t corrupt. How many anarchists are there in your neighborhood?

I've never understood why governments (or labor unions) default to "corrupt" and private firms default to "virtuous".

It seems to be that if you lean Republican, governments are corrupt and private firms virtuous because freedom and the magic of the Invisible Hand of free markets make private firms virtuous (individual choice) and governments corrupt (coercive force).

If you lean Democrat, then governments are virtuous and private firms corrupt because the free market doesn't solve all the problems, and for those you need regulation (government: coercive force). In fact, the regulated free market leads huge wealth disparities, which leads to social problems, and since private firms that are seeking to maximize their wealth disparity, obviously they must be corrupt.

Naturally, neither of these pictures is right, or rather, they are both partly right.

I don’t think that’s accurate. Companies and individuals benefit from outsourcing government work, that creates a huge incentive to paint government as incompetent.

The US military which hands crazy money to private firms is viewed very highly by the Republican’s and less so by Democrats.

Power corrupts when governments extend excessive control over markets.

Power corrupts when private firms extend excessive influence over government.

The most consistent principle is to oppose concentration of power in all its forms.

I mean, lots of people think both are corrupt.

because private firms make money by making things people want and every consumer is by choice.

government makes money by threatening you with violence (pay tax or go to jail) and its subjects are so rarely by choice, and if by choice, it's between one government and another.

The goal of the government is not to make money.

That's why people suggest to let the government run the nuclear power plants. Unlike a for-profit corporation, governmental organizations have little incentive to save money for maintenance by cutting corners.

Private firms aren't inherently "virtuous," but their motivations are at least clear. A business seeking to maximize financial returns for its owners isn't "corrupt," it's "working as expected."

The corruption of governments and labor unions often results in unexpected financial and non-financial returns for various people in ways that are more difficult to detect and measure.

Mainly because most governments around the globe have the monopoly of public force, creating the perfect ground for arbitrary decisions. Its not black and white, by those defaults seems accurate.

Are there private firms that don't make arbitrary decisions?

it's more like both are made up of individuals with high self-interest. The difference is you can choose not to interact with a private firm.

I thought about this in the wake of the BP Deepwater oil spill. How exactly can I choose to avoid interacting with any company that produces a fungible commodity (e.g BP, Archer Daniels Midland, Monsanto)?

I can stop visiting BP branded gas stations but that has nothing to do with where the oil/gasoline actually comes from.

I agree, that's an issue. Or if you wanted to stop funding exploitative mining but still wanted to own electronics, you might have a hard time. But I prefer the "you can often avoid interacting with them" option of private companies, to the "you cannot avoid them and are entirely at the mercy of their whims" option of government.

Not to say government isn't necessary - it is. But I don't understand why people treat it as virtuous, especially when those same people will express distrust in politicians. I suppose it's because government is the only way to legitimately force one's will on others.

I prefer the ability to get rid of a government by voting elsewhere.



Also what's the proof that private organizations are a better solution to the problem?

Just look at the US prison system.

Not to undermine your point, but you have used the worst possible example. In the top 1000 of worst prisons around the world, most if not all of them are not private. Just take a look a Brazil prison system as en example.

Does Brazil have private prisons that are much better than their public ones?

If not, I don't think you have contradicted his point at all.

There are not too many places that we can look at public vs private institutions for direct comparisons (in the contexts that they operate). Prisons are one, I can think of FedEx/UPS/DHL vs the Post Office as another, and it's difficult for me to come up with much after that.

One glaring problem with private prisons is that it creates an incentive to put more people in jail and keep them there longer.

With the government, you get a vote. With other institutions, you have to hope their stakeholders' interests align with yours.

I can have a vote, but government will still force me to pay tributes at gunpoint. On the other hand, im not forced to buy most of private goods. At the end, the evil organizations are those that became monopolies, no matter if it is public or private.

Sure, I'm not forced to buy food, or medicine, or heat, or for somewhere to live, I could just decide to starve, or freeze, or die of plague.

> I can have a vote, but government will still force me to pay tributes at gunpoint

No, at most you'll get a letter in a mail and a lien if you aren't actively trying to evade taxes. No reason to engage in hyperbole.

And then if you don’t leave your house you get put in the last socially acceptable debtors prison. And if you try leave that they will kill you. Ultimately the threat of murder and kidnapping are there

Or you could just be a decent person and pay your taxes.

We're talking about "governments the world over", and they are typically not fairly run democracies.

I believe that is how the country code TLDs do work, but most nations (who are given ownership of them) they just hire a corporation to sell domains and make rent seeking gross revenue off a few characters of a URL.

But those are eclipsed by the generic TLDs where its just a happy accident registrars and rent seekers are making fortunes off of words with com or net at the end of them.

The responses below are kind of funny. The government already runs key infrastructure. I'm not sure if the argument is government vs. Corp. Frankly governments are more likely to consider public interest than corps are. That's not to say that they always do, but corps almost never do. This is why government sells things like spectrum or land for railways if they don't own the rail themselves. They also own things like roads. It's important to realise that there is no incorruptible institution in the world. Maybe domain name management is actually a great use case for blockchain or some other truly democratic distributed accounting/ledger system.

>It's kind of like operating a post office.

The reason that there is a namespace problem in the first place is that most people don't want the kind of domain name that post office numbering would give them.

When the internet was newer, my school email address was @schooldistrict.k12.state.ca.us, by the time I graduated highschool it was @schoolname.org.

I am against privatisation of key infrastructure and I would certainly count the internet as key infrastructure. However I am not sure the government would be entirely benevolent. So I agree that some good regulations are probably a good start.

I agree that ICANN is either corrupt or useless, but you still have the problem of scarcity to address. Domain names are important assets for businesses, and ‘good’ ones are going to be in high demand. Squatters add essentially no value to the system, but there’s no fair way to determine who gets a domain name that multiple parties want, aside from letting the market take care of it. This is a flaw in the system itself, but I can’t see many realistic ways of addressing it.

Domain names could be auctioned, then it would not be allowed to resell them. For example you make a bid on how much you want to pay per year, then you pay that amount until you no longer need the domain name, without the ability to resell it.

Then how do you sell your business? What if inflation or deflation happens?

A business is usually a separate entity, and it's that entity that have made a contract with the domain issuer. It's however possible that a domain squatter would setup bare corporate with only the domain as asset, which can also be done when selling goods to avoid VAT. It is however too much paperwork to be worth it. And if it's abused there can be a specific rule to prevent it. The price can be adjusted for inflation and deflation ... Came to think about my old employer feeling generous for increasing salaries, while he had actually lowered them as inflation had went up more then the pay raise.

Domains already require your real life info. Just give something like 5 free domain names per person and tie it to their ID.

There exists more than enough useful domains for everyone its just most go unused

That’s a completely impractical suggestion, that ignores how domain names are actually used. Domain names are used primarily by organisations, not individuals. But even for individuals this wouldn’t work. Vanity domains are still going to be scarce, and the names in demand for organisations are still going to be scarce. And why should there be any limit on the number of domains you can buy? What if I have 6 projects I want to host? What if I own a company and want to buy my name in multiple TLDs to prevent phishing? What if I want to have a different domain for my dev/test/sandbox/prod environments?

“For every complex problem, there is an answer that is clear, simple, and wrong.”

Businesses can still buy all the domain names they want just like they can now. The owner of a small business just needs 1 domain for their cafe and maybe one more for personal email.

You know these speculators are businesses right? What problem do you think your suggestion is solving?

I propose a .pubkey TLD that doesn't need any registration and can be administered by the owner of an associated public key.

The mapping to the nameservers could be maintained on a blockchain (to fund the storage costs).

One of the worst squatters right now is www.hugedomains.com

They buy up expired domains and immediately put them up for sale for $XXXX

At that price, they can sit on the domains for decades and still make a profit when it's sold.

Eg: https://www.hugedomains.com/domain_search.cfm?domain_name=cs...

They really are the worst. I got in touch with them to buy one, as unfortunately they had one that I needed for business (speculatively registered on their part). When I asked how much they wanted for it, they then doubled the price to $10000 (9 characters, not commonly used words).

In the end I went with an alternative TLD. it isn't as good, and I'd have been willing to pay a decent amount for the one I wanted (even their original price, reluctantly), but I refused on principle when they got silly.

They are just parasites.

This is something I’m terrified of.

I feel like I can never let a domain that I may want to use again in the future ever laps or someone will buy it and extort some exorbitant price from me to get it back.

That would be correct. Why are you terrified? Set up autorenewal in two places.

how can you do that ... via two different registrars?

There are grassroots efforts to make DNS more fair springing up. Handshake.org is one example (with backing from A16Z, Sequoia, and SV Angel) — they're aiming to decentralize the root zone and allow anyone to own their own TLD.

Disclaimer: I'm building a registrar for Handshake called namebase.io.

I've always wondered if they could somehow price domains logorithmically.

First domain is cheap and then cost goes up x% per additional domain. Stop people holding swaths of domains at each gTLD level.

Would people end up just having a bunch of shelf corps...

> Would people end up just having a bunch of shelf corps...

Yes. They already do something similar when registering in a ccTLD which requires a “local presence” in that country – they simply pay somebody in that country to be a proxy.

It costs $360 to incorporate at minimum, at least here in Ontario, so that would still be substantially more expensive than it is currently.

The CIRA (Canadian Internet Registration Authority) at least, probably others, used to give away domains for free but stopped the practise due to squatting.

But one company can deal with multiple domains.

There are plenty of countries where it's cheaper though.

I wish there was a system where each individual is entitled to a few domain names for no cost on their local cctld. That way real users get the domains they want and hoarders have to pay.

Wait, what??? Think really long and hard about how that would play out. This is a scarce resource, that like many digital items feels cheaper than it really is. I personally think that domain registration should cost more. That would a) stop casual speculation and b ) make those speculators feel like holding on to a domain less appealing. I own way more domain names than I have time to work on and until there is any reason to stop renewing them probably won’t.

Perhaps you could have a ratcheting cost, where the first 100 domains are $10 each, the next 100 are $20 each, and so on up to some upper bound, with provisions for shell companies that solely register domains.

That is part of how FreeNom [0] (ex dot.tk) advertise themselves. They have had serious issues in the past, and most of the domains of the countries involved are automatically treated like spam by a lot of internet services.

> With the free domain model, partner countries can offer their local internet population the tools to get online quickly and at no cost. This proven and innovative model dramatically encourages the creation of local content and stimulates the registrar.

This started out with Tokelau, a New Zealand territory, IIRC.

I love the idea, but it seems the commercial world hates it.

[0] https://www.freenom.com

Several ISPs do that actually. Not many but I've seen this a few times.

One need only look at the $185,000 "application fees" paid to icann by companies like donuts, and play follow the money, for all these new insipid GTLDs.

I'm still waiting for ICANN to approve some scammer that realizes he can buy the rights to the .conn TLD and phish the entire world with apple.conn / google.conn etc... talk about juicy ROI

Hmm, .corn would be even better. With modified kerning, links will look like .com.

Yes that would be even more devious!

Can a domain name registrar register a domain name for themselves for free or for a minimal fee? What about Verisign -- if they registered ".com" domain names for themselves, who would they even pay since there is no one above them? Seriously, does anyone know the answer to this? I've never found an authoritative answer.

I've wondered if domain name registrars are speculating in the domain name market themselves. Perhaps even insidiously grabbing domain names that you've shown an interest in. Maybe millions of domain names are reserved by the registrars themselves!

Wasn't godaddy or someone like them doing exactly that, a number of years ago? I seem to recall a situation where you'd go to their site and search for a domain, and if you didn't buy it, then the next day it would mysteriously show as no longer available, but purchasable.

It's called domain name front running. GoDaddy have repeatedly been accused of it, as have many other registrars...


I think this specific process is referred to as domain tasting. Front running is using insider information, such as whois logs, while tasting is actively registering domains and measuring how much traffic the domain gets.


Yea they'd abuse the 3 day grace period that was put in place to squat the domains that were searched. That let them temporarily lock up the domain and also announce to the world that there was interest in the domain, leading to more squatting.

This GoDaddy domain front-running thread is from just 2 months ago:


A long long time ago, I do remember that Verisign did something with the .com namespace so that unregistered domain queries got redirected to a default webpage they hosted (no doubt with plans to monetize). My memory is hazy but I've included some links below.

There was a massive backlash for obvious reasons and they reversed the change quite quickly.

Whenever I use Chrome and it redirects a legitimate website URL that I've hosted locally to their search engine as a search query... I feel a sense of disappointment with the IT community, in that we let Google hijack everything in a similar way without much criticism at all.

[1] https://arstechnica.com/uncategorized/2003/09/2824-2/

[2] https://arstechnica.com/uncategorized/2008/05/verisign-gets-...

> Whenever I use Chrome and it redirects a legitimate website URL that I've hosted locally to their search engine as a search query...

German internet provider T-Online pretty much does the same by default. When you enter a domain without a DNS record, they will redirect you to one of their pages.

I think it's something like "That domain does not exist" at the top and then their "news" site which is plastered with ads and more.

Shady tactics, but you can disable that in their settings when they are your ISP.

Many years ago, in the heyday of domain speculation (after the dot-com crash), some of the largest domain speculators became registrars to gain advantages such as the one you mention.

This conflict of interest was not limited to registrars. Registries got into the game as well, reserving names for the sole purpose of exploiting speculators or speculating themselves.

What the OP could have done instead of taking a random sample is to examine certain nameservers in the com zone. That is, look at certain registrars. This reveals the true extent of parking.

Not sure about the situation today, but in the past many domain registrants were unaware that their registrar was deriving revenue from their "unused" domains The registrants were paying the fees to keep the registrations active and the regsistrar was collecting the ad revenue by setting up parking on the domains.

There is a fee to ICANN in the case of .com's, although its tiny compared to the fee you're charged - $0.18 currently.

hugedomains, as mentioned above for being one of the worst, do this. They own their own registrar/are part of the same company / some convoluted ownership structure (namecheap LLC I think)?

Do you have a source for that? Because it's a pretty big claim. I have used namecheap for years, and they have been absolutely wonderful, but if they are also hugedomains I will start migrating.

I just googled for the thread I remember reading, and it is namebright that is apparently linked to them: https://www.scottphillips.com/post/2010/buying_a_domain_from...

This is the thread where someone mentions they /had/ their domain with namecheap, which is why I confused them: https://www.namepros.com/threads/held-hostage-by-huge-domain...

Personally I've never used either - I'm in the UK so it is different companies (oneandone are who I've used, who have been fine)

I appreciate you looking and following up. I have had nothing but wonderful experiences with namecheap so I was extremely concerned.

Just because a domain has ads on it or no webserver, does not mean it is unused.

I have quite a few domains I use for email and such that I park with ads. They make me $20-$30 a month and I still get full use of the domain for what I need it for.

Edit: People are asking what I use. DomainAdvertising.com. When I originally signed up with them you had to have thousands of domains (which I have sold off), I have no idea what the requirements are now.

> Just because a domain has ads on it or no webserver, does not mean it is unused.

Yep. I have two very old domains, obtained back when domains were a) free, b) referred to as "delegated" from hostmaster, and c) placed in service by pulling the Very Special Form(tm) from ftp.internic.net and filling it out and e-mailing it to hostmaster and hoping you didn't delete a stray space in there along the way thus causing the hostmaster-bot to yell at you versus delegating your requested name in two or three weeks.

I've had them for so long that they crop up in weird places and one of them has been my e-mail domain for over two decades. Neither has a web site or other "normal" services on it. Once or twice a year, I'll get an offer to buy the three-character one (I've replied to every single one with "sure, you can send the cashier's check to [address]" and none ever do). Also once or twice per year, I'll get either a nice e-mail asking for the other domain--it is a common word that some historical groups like to use--or a flame mail complaining that I'm "camping" on the domain and being mad that I won't hand it over.

placed in service by pulling the Very Special Form(tm) from ftp.internic.net

Man, those were the days, weren't they? I remember being really proud of having a two digit alias with Hostmaster...

> I have quite a few domains I use for email and such that I park with ads. They make me $20-$30 a month and I still get full use of the domain for what I need it for.

Can you go into more detail about how you did that? I'm surprised that a random domain would get enough traffic, unless it was a common word or typo away from a major domain.

He did look for domains with MX records in his study

> Mail (2.6% or ~3.5 million)

> Any domain not in any other category, but with MX DNS records (for email), I categorized as Mail. I did not attempt to see if the mail server was working or if delivery was possible. It's possible that many of these domains are not actually used for email, but I've given them the benefit of the doubt.

Also interested in what you use to park your extra domains. I have a dozen or so domains of former/future ideas and would love to at least have them maintain themselves. :)

Something like Netlify would do the trick. Zero cost or maintenance once you set it up with AdSense code, and better than using some parking service that would probably want to keep some of the revenue.


Google has ended the "AdSesne for Domains" direct program. You have to use a parking company.

Just putting AdSense code up a bare domain is a good way to get banned.

Interesting to note, domainadvertising.com doesn't seem to support https, which makes it even just that much sketchier to me.

I also have a domain that doesn't have a site or email attached to it that I use as a code namespace prefix. I believe this is a common practice? (Although many have a business or personal webpage on the same domain)

Can I ask what you're using to park them? I have a few unused domains myself.

Unless I misunderstand the term, isn't your domain already "parked", i.e. bought but not connected to a server?

In the image in the linked post, the "No Web Server" would be what you describe. "Parked" is where you point the domain at a generic page so that visitors see a page, but you're not actively hosting any kind of web server there.

It's just a short name for that category which he specifically includes parked domains and even recognizes "Some of these domains likely have some non-web use".

Call it "Not overtly used" if you want.

You mean AdSense ads or from a different provider?

I'm surprised that such a large percentage of domains are registered with GoDaddy. What leads people to pick them? They're such a shitty registrar, every time I have to deal with them it's awful.

It's trustworthy and their domain management tools, particularly for dealing with large numbers of domains, is on point. I've transferred to other registrars before, and am moving some of my less critical domains to Cloudflare for the cost savings currently[1]. But my main domains will remain with GoDaddy for the foreseeable future.

Oh, and they have 24/7 phone support, that at least, last time I had to call, was US-based. I don't truly trust anyone who doesn't have good phone support. Ticket systems are great and all for low priority requests, but it's too easy for a place with only a ticket system to handle you in a low priority way.

[1]The transfer process is straightforward and the price can't be beat, but their domain management features are still incredibly immature. You can't have name servers set that aren't Cloudflare's, every time you go to domain registration, it wants you to transfer ALL of your domains. And bulk management of domains on Cloudflare is nonexistent: You can't even find the expiration/renewal dates of your domains on Cloudflare without visiting the page for each one individually. It's actually easier to find the expiration dates for all your domains that are not on Cloudflare (on the Domain Registration page). I do like it but they have work to do.

Just curious, how many domains do you have and what kinds of bulk management tools do you need?

I've been very happy with Cloudflare's registrar myself, even though they aren't always the cheapest[1]. Cloudflare just silently auto-renews all of my domains so I never have to think about when they're going to expire. And I don't mind the lack of nameservers since I can't think of a reason to use non-cloudflare nameservers for any of my domains.

[1] Namecheap offers .io for $32 (not a promo, standard price) whereas Cloudflare charges $45.

About 20-30 at any given time. I expect to be able to see the list of domains I own, when they expire, and if auto-renew is on or off from one convenient panel.

Also "just silently auto-renews" is not an okay behavior for me either. Even when I have auto-renew on, GoDaddy sends me email notices around renewal time for any of my domains. I like to manually renew them a bit early to ensure there's no problems with payment processing. Obviously I have like a year until any of my Cloudflare-held domains renew, but if I don't get email notices in advance, I will likely look at transferring them away. Proper notice of account/billing activity is a must for trustworthy service.

For one, my host auto-configures a lot of features on the DNS for my domains. Cloudflare imports it, obviously, when I transfer over, but if my host changes something, Cloudflare isn't going to get it, and I don't even have the option to change my name server provider even temporarily for troubleshooting purposes. That's a big flaw, and Cloudflare will need to fix it to keep me.

I don't have any .io domains, but my understanding is Cloudflare keeps literally nothing for any of their domains, so presumably Namecheap was able to cut a better deal with the registry.

I'm curious how you determined GoDaddy to be trustworthy. A quick search for "godaddy horror story" returns a long history of warnings about shady practices...

I've used them for over a decade without issue, for one. But as I said, 24/7 phone support, the fact that domains are their primary business, not an addon, and that their management tools are absolutely top shelf of any domain registrar I've ever dealt with.

I wouldn't use them personally but I understand the appeal. A lot of other web hosts and registrars have serious flaws. Godaddy picks up the phone when you call them, 24/7. You can also contact them via live chat or email. They have actual 2 factor support and require a 2 day waiting period and manual verification of ID to remove 2fa. It is a real company that has been around for a while and will not disappear. They won't hold your domain hostage if you cancel web hosting or want to transfer it. They offer a ton of products, and they advertise a lot. They use modern control panels or similar interfaces(really the same as many other web hosting companies, but not bad).

I worked at GoDaddy for a while. They have been around a really long time, so I'm not surprised there are so many registered there. They also advertised a lot more than any other registrar. Honestly, having worked there, they are not a bad registrar, they're just more focused on small businesses than on techies.

Another comment in this thread claims GoDaddy has a history of speculatively registering domains people search for but don't purchase in the hopes of extorting them later. Perhaps this practice explains the discrepancy.

I wouldn't be surprised if it's more the case GoDaddy snatches them all up and people don't bother to leave.

The biggest problem with GoDaddy's interface is how they try and hide access to changing nameservers or DNS. I don't want to "Use this domain"!

There are better registrars and some offer reseller accounts which is necessary when you manage a lot of domains and not all your own.

Back in the day, they were much less expensive than most of the other registrars. Now it's mainly inertia in the same way people still have AOL or Yahoo accounts.

They're usually the only registrar people are aware of.

I guess those lewd superbowl commercials paid off then.

Compounding superbowl and name recognition: Their referral incentives are also quite transparent and competitive compared to other quite large registrars that have clauses like, and this is paraphrasing, "payouts of $75 dollars must be accrued within 3 months or will be forfeited" or "must refer $300 dollars of signups per month".

I imagine this causes referring 3rd parties that read small print to align their product with GoDaddy than a competitor.

On https://dofo.com we have all this data and more, you can check it freely.

By using dozens of filters, you can refine the results and see whatever you want.

For example, 66,237 .com domain names contain "coin" and are listed for sale on marketplaces: https://dofo.com/search/?contains=coin&on_sale=y&extension=c...

You can also check the popularity of any keyword on the trends page: https://dofo.com/trends/keyword/coin It says, how popular is the keyword, how many registered domain names exist with that keyword, what percentage of these domain names are used actively as a website and more.

As for the registrars; yes Godaddy is the first one with about ~70 million domain names under management and Namecheap and Tucows are following it: https://dofo.com/registrars

Disclaimer: I'm the co-founder of dofo.com

You may want to fix the headline "we work with the bests"

Thanks, you're right :) We will fix it today.

We've got to the stage now where if you've got a handful of .coms sharing a server, they'd have to be pretty substantial sites before your hosting costs would exceed your domain renewal fees. That's one hell of a market failure.

Huh? $10/year doesn't buy all that much hosting.

Considering the site:server ratio for 99% of the web it's far above $10/y/server

I'm shocked by the lack of porn on the Internet. Gambling appears to be a much more popular activity.

Though I have absolutely no evidence to support this, I suspect that gambling doesn't suffer the same, uh.. stamina issues, that porn probably does. I assume that the sustained stream of small amounts of currency would easily add up to more revenue for gambling than porn, resulting a larger market.

> 0.59%

I researched it and most sites seem to agree on the fact that at least 4% of websites are sex-related, seems a lot more realistic to me.

I think the study is classifying normal sex-related websites as "Content" and porn filled parkings domains as "Porn".

I'd be curious how many domain names are used, but not for the web or email. No idea how to measure this, and I'm sure it's tiny.

For instance, I have a .io domain I use as my own little personal dynamic dns service since dyndns turned evil. The TLD doesn't do anything interesting, and any web pages served by the subdomains are done on nonstandard ports, so if this study scanned .io domains, they'd classify my domain as not in use, when it very much is.

I recently discovered Freenom[0], which offers free .ga domains - I think you need to manually reconfirm it once a year, but other than that, there doesn't appear to be any downsides. Perfect for personal stuff!

[0] https://www.freenom.com/en/index.html?lang=en

I have to assume this kind of use represents such a small amount of domains it is not worth discussing.

> .. dyndns turned evil.

Care to elaborate?

In the grand scheme of things, it's not the evilest thing to have ever happened. At some point the Windows auto-updater decided to push a new feature that changes the network DNS settings to make them point to some Oracle DNS server. If you undo the changes without disabling the "feature" in the app, it'll reset them for you. It was done with the claim of speeding up DNS lookup or somesuch.

It's a checkbox, and no doubt it would remember the setting if I changed it, but the idea an app would change my DNS settings (which, in my case broke an internal app that was admittedly fragile to begin with) without my permission was enough for me to cancel my DynDNS account there and then while I cobbled together a replacement.

I recently quit DynDNS after many years of use because of sheer frustration. I dropped my paid subscription because their Mac OSX updater hasn't been working correctly for years. It would forget its settings and/or fail to update. I filed a bug report with DynDNS and spoke to their tech support, and it never got solved. (I suppose I could have found a third party updater, but I expected their own updater to work.)

Hurricane Electric does free DNS hosting and supports dynamic DNS

Since they stopped offering a free plan, I imagine. I just paid them the money, it's worth it.

There are better free alternatives like freedns.afraid.com out there if you dont really need to rely on your router that may only support Dyn.com

Excellent research! Finding available and meaningful domain names is a big problem. One solution is to chose a non-.com TLD. There are many but it's still an unpopular choice. I believe this is changing and other TLD will become more popular as .com names become more limited. With all the TLDs out there it becomes cost prohibitive to buy them all.

List of TLDs: http://data.iana.org/TLD/tlds-alpha-by-domain.txt

Aside from anything else, there is a lot of work out into this - it's a pretty robust piece of research, and is (I assume) a fairly good calling card as there are many business questions that can only be answered with this sort of skill set / stubbornness

I was hoping to see there a block with unregistered domain names.

One could do it by, for example, training a Markov chain/neural network to spew out (sample) domains, and then see how many are currently registered, to compute the general XXXX.com space utilization fraction.

This approach would be powered by a more sensible, IMHO, probability distribution than, say, taking all domain names with up to 30 {a-z0-9} characters with uniform probabilities.

One intriguing proposal to limit domain name squatting is "Harberger Taxation"[1]

Under this scheme, every owner of a domain would declare a value for the domains they own, and pay a percentage of that value as a tax (replacing the current annual registration fees).

Owners are incentivized to post an accurate value for their domain because they must sell their domain for the price they list if anyone offers to buy at that price.

While it makes sense in theory, I think there are some practical problems that would have to be addressed before such a system could be put into use. For example, what stops a company with a large bankroll from squashing competitors by purchasing their domain?

[1] - https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2818494

That seems totally bonkers for the case of an entity like Google or Apple. Would they really list their domains worth $1,000,000,000,000 dollars? and pay some percent of that a year?

Once you do this, squatters will just find an easy way to make a domain seem active: email traffic, basic websites, etc.

Also sometimes there are legit uses, like squatting names in order to protect your brand. For example I could see google squatting google-search.com

This scheme has the advantage of it not mattering if something is active. Instead you pay fees based on what it's worth to you.

If someone is squatting on a high-value domain, they will have to pay high fees or sell the domain to someone who can make better use of it.

This prevents someone else from squatting on google-search.com and allows it to go to the highest value use -- protecting Google's brand.

So, I'm an average person hosting my email on my domain. I use it for both personal and professional reasons. I'm by no means wealthy, so even though my email address has huge value to me (from clients, to friends, to banks and digital services) I can't pay a huge price for it.

With your suggestion, someone with enough money would be able to force me to sell it. Identity thieves would have a field day with this!

> While it makes sense in theory, I think there are some practical problems that would have to be addressed before such a system could be put into use. For example, what stops a company with a large bankroll from squashing competitors by purchasing their domain?

They don't have to sell if the offer sucks.

The idea is that if I'm a startup that owns google-disruptor.com, I have to declare a value for that domain name (let's say $100,000), and pay a percentage of that value as a tax. I'm then obligated to sell the domain to anyone who is willing to pay $100,000. This prevents me from undervaluing the domain.

Since Google has a much larger bankroll than me, they can purchase the domain for $100,000, and revalue the domain at $500,000 -- out of my price range.

Since Google has a much larger bankroll than you, if they decided that they wanted prickledpear.com, wouldn't you have to sell it to them, or any other party with deep enough pockets?

Yes, that is the problem I mentioned in the original comment that I'm not sure how to solve :)

So as a competitor to Google with $1000 in my pocket, Google could domain-snipe me out of existence, legally, for $1001 dollars?

To clarify, under the scheme being proposed, they actually do have to sell. Otherwise you would just declare your domain has zero value, you'd pay no tax, and then if anyone offered to buy it, you'd say the offer sucks and refuse to sell. So declaring a value is a commitment to sell and pay the tax.

> To clarify, under the scheme being proposed, they actually do have to sell.

That just sounds broken to me. If I buy a domain, I should own it. Someone who wants it should offer what they are willing to pay for it, if I don't like their offer, they can find another domain.

It's definitely a strange idea especially the first time you hear it. I don't know if I like it for domain names.

I own nibble.com and rolled my own very simple static "parked" page using Bootstrap and an Escrow.com button.

In my opinion the parking services like Sedo etc. aren't worth it just for capturing sales leads...the commission is too high.

Just in the last few days, I've been looking for something simple like this to put on userhelp.com, thanks!

"4. Not all records in the zone file are valid domains. Some do not have a WHOIS record and may act as honeypots to catch people distributing and using zone files without permission."

It has been a while since I looked at the com zone file access ageement but as I recall the most substantial restriction was on redistribution.

A wee bit of web searching shows there are still plenty of non-official sources selling zone files.

Perhaps it's just me but sometimes I think the number is increasing.

Are the honeypots designed to catch these sellers and stop them from selling?

Does Verisign enforce the ZFA agreement?

Maybe all Verisign can do is revoke the access of the licensee who breached their agreement?

Domain registration is one of the most perfect use scenarios for a blockchain type system. It would be entirely decentralized meaning no party has any unfair advantage which leads to these sort of systems where players obtain massive amounts of domains for non-market prices and then sit on them indefinitely relying on the occasional way over market price sale to justify the business model. And while views may vary here, I think it would also generally be a positive for governments to no longer be able to strip individuals or companies of their domains, such as for instance SciHub.

One practical problem in the early stages is that various domains could be snagged and used to extort the companies/individuals who 'rightfully' own them for unreasonably large sums of money if they wanted access to their domains. This issue could be resolved by premining and mirroring the current domain system before public launch with domains transferred to their rightful owners upon proof of ownership akin to the systems used by Let's Encrypt.

Probably the most challenging problem would be in determining how many domain registration coins would be generated per unit time. Too few and prices inflate too high and destroy the system. Too many and prices deflate and the coins become worthless - destroying motivation to keep mining, in turn again destroying the system. Lots of possible solutions though. Fun problem!

"There are currently 137 million .com domain names registered."

But how many is regularly visited today by most of the population? A good 10k max?

Alexa can give you an idea of the most popular sites. https://www.alexa.com/topsites

Well, there are 500 on the top 500. I don't think Alexa goes low enough to answer the GP.

10k sounds low by at least a couple orders of magnitude.

www.steam.com. They've been holding out for almost a decade trying to get money from Valve.

Its a shame Armenia controls .am because ste.am isn't in use right now.

Domain names are incredibly underpriced (I'm referring to the "good" ones). With the evolution of the internet, where startups are now willing to spend 10k-100k on a domain, it makes no sense that its original cost was 10$, and that insentivises speculators.

If prime domains would cost 100$ or even 1000$ per year it wouldn't affect a running business much, but would almost completely eliminate the viablilty of domain squatting. And if you're an individual you'll still have many other options.

This could be applied to the entire COM TLD (it does stand for commercial after all) or only for parts of it under some criteria (like length etc).

Just because a domain may appear to be unused, it doesn't necessarily mean it is unused.

For instance, I have a domain I use as a file hosting which I can link to when I want to share something (I used to use Imgur for this purpose before, but frankly I don't trust any image host at this point, so I set up nginx to host a single directory). As you need to know the file names to download them, the domain would appear to be unused.

Although now that I look at I suppose I do explain its purpose, so here is that: https://i.xfix.pw/.

And what percentage of domains do you think represent use cases like this? I don't think it significantly changes the statistical analysis. Sure, there are some tiny fraction of a percent of weird uses from people like us, but it's hardly worth even mentioning.

But xfix.pw also has a web page.

Anybody who is interested by the economics of domain names should read Henry George who laid all this out over one hundred years ago in Progress and Poverty. Domains are just the new frontier since we discovered and claimed all physical land. The 1990's were a land grab and now we are entering the period of calcification.

I wonder how many of the 25% newly registered domains were registered for the first time ever?

What is the statistical power and what makes the sample "random"?

It seems like it has been this way for a long time. I remember back in the dialup days in college sitting around trying out random urls and even then many of them were ad sites that were for sale.

That reminded me, I have domains to renew. I better take care of that.

"Domain Usage, from a sample of 2,188 domains" isnt that to low sized sample to analyse data of more than 20 million domains?

It should be fine - strong law of big numbers. They even computed some confidence intervals (99% CI ±3%), but it's not precisely explained how, just references a paper.

Quote from the research:

> the crawling server was only configured for IPv4

He may have misqualified 11% or more of those domains.

I seriously doubt 11% of domains are IPv6 only, but I'd love to see someone do some research on it. I would assume its only a fraction of a percent that ONLY responds via IPv6.

All the ones that I want!

Domain renewals should be priced higher. $500/yr, $1000/yr, $5000/yr would prevent squatters.

I'm a hobbyist with a few domain names (under 10 total); of those, the majority are in "use" (that is, there is something of mine they resolve to - not to a blank page or a "for sale" page or anything like that). But a couple or so have no content behind them (and honestly, some day I'll probably let them go back to the wild).

Should I have to pay the amounts you list for something that is a hobby? What about others with hobby domains and sites? Should they have to pay such extreme amounts just to keep their hobby site going?

There are people out there who run sites "just because" - giving away information and content for free, like the web used to be - like it was meant to be, not this almost seeming abomination we have today, built on ephemerous "likes" and other useless metrics.

Your solution would do nothing but turn the web even further to corporate control, and leave hobby sites to only the well heeled, or to those who can scrape by (perhaps NPR fund-drive style). The majority would simply fold up shop, never to be seen again.

I get it - I don't like true squatters any more than the next person; that is, those who sit on a domain hoping to make a buck off of some words. I have no problem with a business establishing a valuable domain based on their name and reputation. I don't like those who sit on domain names that sound or seem like a similar company's name, and try to bank of that (or worse).

But I think your solution of "raising rates" - at least unilaterally - isn't the answer, unless we want a worse web experience than we already have today.

Wouldn't this completely prevent small local organizations/business owners from running their own websites and push them in the Facebook/Google/etc ecosystem? or do I not understand what a "renewal" is?

No you understood right. He's suggesting that a domain should be 5000 a year so that you or I could never possibly own any .com domain for personal use.

He didn't limit his suggestion to .com

It would also prevent most private individuals from owning their own domain - hardly a solution without collateral damage.

And they would keep the web from being usable by anyone that is not in it for business, but wants to put up a site.

$20 a year for a domain plus $50 a year is an easy price for getting a server to experiment with. If things cost 10 times that, it would price out a lot of activity at the margins of the web.

Wouldn't that be harmful to small businesses though?

That's a horrible solution. How about we add more TLD's so it costs effectively $5000/yr to squat the same domain on all the alternative TLD's

Do that and I'll drop my personal.com so will lose control of my domain for email and handover more lock in to Google/email provider of choice.

I absolutely disagree, but I think it is reasonable for domain authorities of various gTLDs to incorporate company names into domain name rights a bit more - if your company is FooBar and foobar.com is taken you may get some legal advantage in any efforts to secure it, this could just shift the domain name squatting issue to a company name squatting issue though. And, I'm also making the assumption that we're fine with pushing personal pages and other domain uses off on a different gTLD.

> And, I'm also making the assumption that we're fine with pushing personal pages and other domain uses off on a different gTLD.

This is a sure sign of getting old, in the vein of "old man yells at cloud," but you can have my personal domain that's been registered in a common gTLD long enough for it to buy its own beer when you pry it out of my registrar's hands.

Yeah, it sucks to have come along at the end of the landrush and all of the good spots are taken. That's, sadly, life when it comes to scarce things like that. I sure wish I'd been alive in the 40s or old enough in the 80s to buy waterfront property in Ballard when it was cheap but that time has passed, too.

This is one benefit of having a unique name. My good friend, "Robert Miller" is out of luck. Me, well, yeah. Only one person with my name in the United States, possibly the world.


> I think it is reasonable for domain authorities of various gTLDs to incorporate company names into domain name rights a bit more - if your company is FooBar and foobar.com is taken you may get some legal advantage in any efforts to secure

What about Foobar Plubbers? Do they get Foobar.com? And Foobar Restaurant in Dallas? Split it to FoobarPlubmers.com and FoobaRestaurantDallas.com? And FoobarPlubmersOhioLimited, vs FoobarPlumbersTennesseePartners? Mr Foobar for a personal domain, or Ms Foobar for a portfolio of work?

Perhaps a better solution would be for each subsequent domain cost $1 extra to renew. I.e. your 11th domain would cost you $10 extra per year, your 101st domain a hundred extra a year.

I don't have a big problem with people holding onto a half dozen domains for years at a time and not making much use of them, but companies owning thousands of domains as speculation? F that noise.

Holding companies would break an approach like this very quickly - it's likely that the cost burden of this change would hit smaller companies much quicker than large ones.

Correct me if I am wrong, but are you saying domain sqautters could spawn 1000s of holding to maneuver around my proposal? I guess there could be countries/states where you can register a company for a very low cost. What is the cheapest place one can start a company?

$50 is the price to incorporate in Arkansas, Colorado, Hawaii, Iowa, Oklahoma and Mississippi[1] apparently

[1] https://www.entrepreneur.com/article/241528

Interesting, but there are typically recurring costs involved in keeping a company open. Mind you, I guess accounting/filing costs also can get pretty cheap at scale, especially if each company is identical in nature.

This seems pretty hard to enforce. What if instead the fee per domain just increased by a dollar per year? The first few dozen years shouldn't be too onerous for those who aren't squatting on a mountain of domains and if your domain is 50+ years old maybe you've found a way to monetize it.

This seems reasonable as well.

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact