What would you do if you had a list of best 50.000 unregistered domains?

[twiki]

Hypothetical question:

What would you do if you had a list of best ~50.000 unregistered domains?

But not enough cash today to register them. (that'd be like half mil)

Names that WILL sell in large % the next 5 years on 4-5 digit figures, based on growing scarcity, industry analysis, price history analysis and everything together analyzed and ranked by a computer big brain.

Sort of like an AI system of largest domain platforms would generate, if you want.

 

[manpreet]

Hypothetical question:

What would you do if you had a list of best ~50.000 unregistered domains?

But not enough cash today to register them. (that'd be like half mil)

Names that WILL sell in large % the next 5 years on 4-5 digit figures, based on growing scarcity, industry analysis, price history analysis and everything together analyzed and ranked by a computer big brain.

Sort of like an AI system of largest domain platforms would generate, if you want.

Nice thought, but 50,000 is a large number to even shortlist! Investing in Quality rather than Quantity should be the motto.

 

[BrandEntrance.com]

My medium is reality... whatever that may be.. or as close to that as it gets.

 

[jmcc]

Out of a billion domains that are not there anymore, or ever registered, there has to be something left. Statistically at least. But that's a mammoth task for a very powerful system. The 50K number is also a hypothesis; maybe the right number is 50 or 500. I can't tell and know no one who can tell right now.

The Domain Tasting registrars were, at the time, keeping 1 out of every 330 domain names tasted. The aim of some of the early tasters was to build a set of domain names that would earn from PPC while waiting to be sold. The later tasters were just registering for PPC value. What was valuable then may not be valuable now. And what was junk then might be treasure now.

The AI will follow current trends and emerging markets. That's where the money is, and greatest likelihood to get something nice. I haven't mentioned this because I thought it's kind of obvious. So rather than scouring and regging, you have this theoretical system that can - woosh - reg out the niche.

That's similar to what happened to the new gTLD registries that decided to keep the premiums for themselves. They killed the market for their gTLDs. By effectively registering every potentially valuable domain name in a niche, it kills market interest in the niche. No interest = no market = no sales.There is always a human element to registrations and robot registrations tend to stand out. One of the tests that is used to detect this kind of registration pattern in a TLD is graphing the domain name length of registrations with the .COM being a reference graph. The gTLDs with robot registrations (mainly webspam and e-mail spam registrations due to discounting offers) then to have large spikes in 4, 5 and 6 length domain names.

The other thing that is sometimes done is a DNA test. ( http://www.hosterstats.com/com-tld.php ) That's where a TLD is compared against all others to see which terms are available in other TLDs. Most domainers do this by checking WHOIS records. It makes sense to do it this way with limited data and resources. The other way is to compare zone files. It takes a while but there is a large set of data involved. With the right database schema and algorithms, it can be done on a desktop PC. The reason why many of those potentially valuable domain names that are long gone in .COM are unregistered in other TLDs is because those other TLDs have different dynamics (ccTLDs) or the gTLD is relatively small/high-priced.

Regards...jmcc

 

[iGreen]

Filter out your list for the top 1% and try registering just 500. I think that too is a big number for unregistered. But can take chance if you are confident and experienced enough in the domain industry.

 

[twiki]

The Domain Tasting registrars were, at the time, keeping 1 out of every 330 domain names tasted. The aim of some of the early tasters was to build a set of domain names that would earn from PPC while waiting to be sold. The later tasters were just registering for PPC value. What was valuable then may not be valuable now. And what was junk then might be treasure now.

That's similar to what happened to the new gTLD registries that decided to keep the premiums for themselves. They killed the market for their gTLDs. By effectively registering every potentially valuable domain name in a niche, it kills market interest in the niche. No interest = no market = no sales.There is always a human element to registrations and robot registrations tend to stand out. One of the tests that is used to detect this kind of registration pattern in a TLD is graphing the domain name length of registrations with the .COM being a reference graph. The gTLDs with robot registrations (mainly webspam and e-mail spam registrations due to discounting offers) then to have large spikes in 4, 5 and 6 length domain names.

The other thing that is sometimes done is a DNA test. ( http://www.hosterstats.com/com-tld.php ) That's where a TLD is compared against all others to see which terms are available in other TLDs. Most domainers do this by checking WHOIS records. It makes sense to do it this way with limited data and resources. The other way is to compare zone files. It takes a while but there is a large set of data involved. With the right database schema and algorithms, it can be done on a desktop PC. The reason why many of those potentially valuable domain names that are long gone in .COM are unregistered in other TLDs is because those other TLDs have different dynamics (ccTLDs) or the gTLD is relatively small/high-priced.

Regards...jmcc

Appreciate the insight! Some notes, there are tools out there that do the test you mentioned and show names status across various gTLDs. I believe though that each gTLD is a different animal, and if you'd ask me personally, I only deal in .coms.

Too bad that other TLDs are discounting so much - this creates a wave of spam registrations and despite bringing cash in, just as in your referenced book, it degrades the TLDs quality in the eyes of the people actually looking to build normal sites and trust.

 

[twiki]

Filter out your list for the top 1% and try registering just 500. I think that too is a big number for unregistered. But can take chance if you are confident and experienced enough in the domain industry.

Correct. But then the list might only be 500 names, who knows. Still the same approach would be recommended.

Indeed, selection quality depends on experience, and that comes with time. When staring domaining, you're kind of walking blind and I guess most of us shed a lot of portfolio at the end of first year and even later. But after some time, experience kicks in and you can tell straight off which domain is worth something and stands out. Anyway, what you suggested is exactly what I would do in such a situation.

 

[Name Trader]

People still use Estibot?

Yep. You are going to need something to weed 50,000 domains to a manageable size. Ok. It's not perfect. You might throw some out with the bathwater (inevitable anyhow). You also need to use your own intuition about each valuation, high or low. What would you suggest to replace it?

 

[.X.]

I wouldn't sleep for 5 days straight first off, I would take the best 2000 domains in the list of 50,000 then i would sleep for 2 days straight.

 

[Name Trader]

I wouldn't sleep for 5 days straight first off, I would take the best 2000 domains in the list of 50,000 then i would sleep for 2 days straight.

How would you wittle it down from 50,000 to 2,000?

 

[.X.]

How would you wittle it down from 50,000 to 2,000?

whew, that is the tough part, but i would go with my instincts + research on my choices, i would target 1 word dictionary .com , I know i would make mistakes in my choices but that is just domaining IMO

 

[jmcc]

I wouldn't sleep for 5 days straight first off, I would take the best 2000 domains in the list of 50,000 then i would sleep for 2 days straight.

Before or after registering them? Reading through 50K regs can be like reading a telephone directory to find the address associated with a single phone number. It might be better to automate the approach so that you don't end up with a headache worse than a bad hangover. The danger with a list of what may be the best 50K and a strictly manual approach is that you will come up with a good excuse to register them all. They would need to be prioritised.

Regards...jmcc

 

[frank-germany]

Hypothetical question:

What would you do if you had a list of best ~50.000 unregistered domains?

But not enough cash today to register them. (that'd be like half mil)

Names that WILL sell in large % the next 5 years on 4-5 digit figures, based on growing scarcity, industry analysis, price history analysis and everything together analyzed and ranked by a computer big brain.

Sort of like an AI system of largest domain platforms would generate, if you want.

I would consult a psychiatrist
and re-read this forum

 

[twiki]

I would consult a psychiatrist
and re-read this forum

It's still a hypothetical, and largely fictional thread. When you're dreaming, everything is possible, isn't it...?

 

[tomcarl]

Yep. You are going to need something to weed 50,000 domains to a manageable size. Ok. It's not perfect. You might throw some out with the bathwater (inevitable anyhow). You also need to use your own intuition about each valuation, high or low. What would you suggest to replace it?

Agreed, but I feel that GoDaddy's domain appraisal tool has been a lot more "accurate" as far as guessing what to price for, which isn't saying much but I'd put more weight in it than estibot.from personal experience

 

[twiki]

Agreed, but I feel that GoDaddy's domain appraisal tool has been a lot more "accurate", which isn't saying much but I'd put more weight in it than estibot.

It's often better to price your domains by GD appraisal, so you have room to negotiate, if you're unsure what price to use. But otherwise the value itself is just as rubbish as Estibot, if not worse.

I trust neither.

 

[twiki]

Update:

In an attempt to at least test the theory, during the last 24 hours I've enhanced my tools and used the thing to full scan all the niches I deal in.

What I was mostly curious about is how deep is the .com niche scanned, and if there are any exceptions, anything left of value.

The results are the expected ones:

- The .com TLD is VERY DEEP scanned. Almost everything left of value has been taken already. In competitive niches 100% is taken. That's where money is for most domainers.
- After a lot of crunching, I've ended up with an initial list of roughly 200k names, all of them 2-words (what I deal in)
- Out of them, about 30k were free, most of which obviously rubbish. The software was good at it; but from the human view most don't make any sense whatsoever. AI could be far better at disseminating what's what.
- Human factor is still vital to determine if anything is worth taking, in the end;
- I have a pre-final list of 5K and currently digging more. But most of it is to be scrapped.
- Got a few I selected, quite nice but need to be further validated. Guess there might be 50 domains overall worth taking, or less, but there's a lot of manual work in order to decide which is worth anything.

- Unsurprisingly, in newer niches and trends there is slightly more worth assessing.

I believe a true AI system might be better at finding brandables though. And yes, a partner like Google could be the key to finding untapped segments.

But otherwise, at least to me, the question can be put to rest.

There is no chance you will find a large number of untapped domains to reg, at least in the main gTLDs (dunno about newer TLDs). You might find a few to your liking, but nothing close to the hypothesis being tested.

Still, it was quite fun to explore.

 

[twiki]

Side note, one tip for hand regs. This is what works for me:

When you do hand regs, chances are great that you're registering crap. Even if you're already done all the due diligence you normally do. (note, when you do dropcatch, chances are great you won't get the domain at all... the world ain't perfect, it seems )

What I do with regs, is to let the domains I have found just sit for a while. Then I come back several times in other parts of the day and review the list again, even if it's a short list. Each time I see a different angle, yet another detail, and cross out a few more domains from the list. So you end up the best of it. Or with nothing to reg, but that's a good thing if you got there.

I think this works for two reasons.

First, you cut off your own momentary impulse to reg your brand new find right away. With regs, this is again a good thing.

Second, in different moments of the day and often coming from different activities, your brain (or mine at least) is in a different state, and sees other aspects that were missed.

Edit: For example, for me there is a huge difference before and after coffee.

 

[tonyk2000]

Then I come back several times in other parts of the day and review the list again, even if it's a short list. Each time I see a different angle, yet another detail, and cross out a few more domains from the list.

Words of wisdom Somewhat similar applies to the initial review of a 30K list, you'll need breaks, or the quality of the review will decrease after each 1K reviewed.

Google could be the key to finding untapped segments

Unofficial databases of google searches, including country-based (google .ca vs google .co.uk), with some metrics like term popularuty, do exist. They are not free. Sometimes, it is funny to read what the people are searching for. But such databases are barely usable for domaining purposes. I have some, so it is personal experience. Too many long phrases, too many mistypes like different "versions" of Mississippi, etc. They may be well used for SEO though.

 

[twiki]

Words of wisdom Somewhat similar applies to the initial review of a 30K list, you'll need breaks, or the quality of the review will decrease after each 1K reviewed.

I have observed the same thing. Although I am quite alert during the day, after each 1K I feel less alert. And sometimes I need to stop after say 3K cause I'm already sleepy. If I'd continue then at some point I would go straight to the bed. It definitely drains brain power.

 

[twiki]

Oh and about the end result of my list:

People already snatched a couple domains today from right under my nose. So yeah, the list is good. (though the thing pisses me off)

 

[4better]

It will all depend on my budget and financial capability. Will only try to register the most I can to the most I can afford.

 

[twiki]

Yet another piece of advice for reg people that do this professionally:

Do not use registrar search for your reg lists! Use a Whois provider instead. Especially when you've found a very good name.

I just got bitten today, yet again.

Due to my whois calls being out of acceptable volume, have been throttled - so I've decided to use a registrar search for the last part in the list so I can get it done.

Well, few minutes after that, one of my domains was registered at the very same registrar. It was kind of the best in the list, or at least one of those domains that many people go for. This ain't no coincidence. Timing was clear. And yes, it happened in the past as well.

Learned my lesson hard, once again. Better wait a little for more whois quota, than hand over my data to registrars. (Edit: the kind that do this, cause not every registrar does).

 

[MapleDots]

I would sell the list to Mike Mann for a 10% commission

 

[twiki]

I would sell the list to Mike Mann for a 10% commission

Ahhh.... you must be the one involved I knew it!

P.S. You're selling my domains too cheap

 

[tonyk2000]

I would sell the list to Mike Mann for a 10% commission

Meet the 'Mann' who registered 14,962 domains in 24 hours:

https://www.cnet.com/news/meet-the-mann-who-registered-14962-domains-in-24-hours/