poll What makes a domain valuable (or not) in 2020 ?

[Gube]

With the idea of building a new, open, appraisal formula in mind, we thought the domaining community could help.
Cast your vote on what influences the most a name's value to you.

An appraiser should help investors to invest at the right price, but can also be used when selling a domain to an end-user ("Proving" the price).
Making the number random wouldn't work..
But here are some elements we could use in a future (free) Appraiser :

• The domain length
  "the shorter the better" or another rule

• TLD used (Does a .com have the same value as a .paris, Increased complexity here, as some GTLDs might be a better fit than a general TLD)

• The domain age
  I think it's still believed age can be a deciding factor when buying or selling a domain. (More names available and registered in the past, meaning the average hand reg quality was better)

• Word balance ratio
  I did use that factor on my personnal scoring system, it worked that way : A domain with a small word size balance had a better score than others (Example : The dog was better than wonderful dog)

• Vowels/Consonants ratio
• TLDs taken count
• Potential end users
  In my formula, I was checking for the amount of similar domains registered (Count of all domains countaining the SLD of the name searched, it's not precise on short SLD, example : the.paris, it was matching all domains containing "the", but .paris was not taken into account - Should it be improved and used ?)

• Readability
  This factor might not be clear, but basically, we would check if a given name can be spelled and read easily

• Dictionary factor (If the name is part of a dictionary, and is commonly used for example, could it influences its value?)

• Similar names sold (Similarity % multiplicated by the sale price?)

• More factors? Feel free to discuss about them in the thread

You can vote for up to 3 different factors, but if you have more in mind, or would like to defend your point of view, please do.

Gube

 

[koolishman]

Well, a price range is not subjective.
If you search for something totally random, it won't tell you it's worth 5M.

The thing is that, the more info you have, the less subjective it is

Good luck!

 

[BrandAptly.com]

I believe a platform that factors user ratings of domain names (based on a number of essential criteria) into it's algorithm may be a more efficient appraisal vehicle.

 

[Ostrados]

#1 factor is potential end users, or in general can the domain be used by a brand or business. There can be many high value 3L .com domains with no end users, and many 3 words domains with plenty of end users.

Another important factor not mentioned here (rarely mentioned anywhere) is name scarcity, scarcity leads to higher value because of low supply. And this is not only limited to liquid domains (2L, 3L, 4L) but also to one word domains and even 2 word domains. For example category killer one word domains have high value because they are rare, similarly good 2 word combos in some niches & categories are limited and taken.

Brandable domains are the hardest to evaluate by a computer, because there can be made up names, unexpected combos, creative blends.. and many other possibilities that need human judgment to be evaluated correctly.

 

[koolishman]

Brandable domains are the hardest to evaluate by a computer, because there can be made up names, unexpected combos, creative blends.. and many other possibilities that need human judgment to be evaluated correctly

This is 100% correct.

 

[YourBestTrader]

This is one of my favorite topics to read about, so thank you for starting this thread.

Over the last decade, I've determined that there are between 100 and 200 data points that might influence the value of a domain... a truly fascinating study.

The way you want to go about it for accuracy would be to have it act as a Wizard, with the first page asking which type of domain it is: Dictionary, Exact Match, Geo, Liquid, Numeric, etc.

The next page would ask relevant questions for that domain type if needed, and pull the data for all of the other factors either from its own database or querying other databases using API calls.

You'd have to run a LOT of testing using both programmers and mathematicians to figure out an accurate "weighting" of the data points... this is where the highest level of difficulty lies.

Of course, Pareto's law will end up holding true and only 20% of the data points will end up accounting for 80% of domain value... It will be interesting to see what those data points are!

A study showed Mike Mann's valuations could be predicted to 66% variance with just 4 data points.

This would line up with me imagining that of the 200 data points, you'd use only 100 for a given domain type... then Pareto's law would say only 20 of those points should account for 80% of the price.. then if you use Pareto again on the 20 data points... that leaves you with 4 data points... much like what Mike Mann is using.

Personally, for exact match domains, I'm using my own version of Andrew Rosener's valuation formula...

"Valuation" = Searches x CPC x Unique Competition Modifier x 12 month x 0.33 CTR x TLD Modifier x Niche Modifier

It gives me a good feel for a domain's potential value especially when comparing two similar ones and having to choose between them.

Someday when I have the budget I'll build the Ultimate Valuation System... but for now... I gather my data in secret and super enjoy threads like this

 

[BBQ Names]

Everything is a factor. Domain valuation is highly subjective and requires a lot of nuance.

Obviously some things matter more than others, like extension.
A top branding word or keyword in .COM is obviously going to be worth more than some lesser extension.

Also, the potential pool of end users. The more potential buyers, the more potential value.

Brad

Boom. Of course it us subjective. That is the artistic JOY of domaining! Sure, a range i suppose could be estimated, but, really, pricing (and then selling) is so much a salesman’s art. There are objective aspects of it, but “beauty is in the eye of the beholder”

 

[dande]

Nothing beats potential end users to me. Everything else comes in second place.

 

[Martins B.]

TLD+Dictionary facotr+end users is probably the golden triangle of profit.

Brandables are a whole different can of worms though.

 

[ThatNameGuy]

With the idea of building a new, open, appraisal formula in mind, we thought the domaining community could help.
Cast your vote on what influences the most a name's value to you.

An appraiser should help investors to invest at the right price, but can also be used when selling a domain to an end-user ("Proving" the price).
Making the number random wouldn't work..
But here are some elements we could use in a future (free) Appraiser :

• The domain length
  "the shorter the better" or another rule

• TLD used (Does a .com have the same value as a .paris, Increased complexity here, as some GTLDs might be a better fit than a general TLD)

• The domain age
  I think it's still believed age can be a deciding factor when buying or selling a domain. (More names available and registered in the past, meaning the average hand reg quality was better)

• Word balance ratio
  I did use that factor on my personnal scoring system, it worked that way : A domain with a small word size balance had a better score than others (Example : The dog was better than wonderful dog)

• Vowels/Consonants ratio
• TLDs taken count
• Potential end users
  In my formula, I was checking for the amount of similar domains registered (Count of all domains countaining the SLD of the name searched, it's not precise on short SLD, example : the.paris, it was matching all domains containing "the", but .paris was not taken into account - Should it be improved and used ?)

• Readability
  This factor might not be clear, but basically, we would check if a given name can be spelled and read easily

• Dictionary factor (If the name is part of a dictionary, and is commonly used for example, could it influences its value?)

• Similar names sold (Similarity % multiplicated by the sale price?)

• More factors? Feel free to discuss about them in the thread

You can vote for up to 3 different factors, but if you have more in mind, or would like to defend your point of view, please do.

Gube

First let me say, one of the reasons I will continue to register some of my domains at GD is due to their valuation/appraisal program. I can provide hundreds of examples where GD values domains at $xxxx.xx and Epik (I believe mirrors Estibots) at $0.00.

I have experience in several industries where "scoring" is a big deal, and no doubt some sort of "scoring" plays a roll in GD's valuations.

Generally speaking with "the idea of building a new, open, appraisal formula in mind", Epik should look closely at GD's model, and hire me with my experience in "scoring" to improve on what I believe to be the best valuation tool in the domain industry☺

Ironically I registered a two word domain at Epik a week ago that pertains to "scoring", and here are the valuations for:

NameXXXXXXX.com

Epik $80 - at least it's not $0.00
Free Valuation: $275
Go Daddy: $1,362
NameWorth: $5,540

Finally, like in real estate, comparable's are key to respected appraisals, and the same thing goes for names/domains imho.

 

[Namersen]

If you're training an automated appraisal tool, does it make sense to let the user give it hints, for fine tuning the result? I am thinking of suggesting the root words for creative blends, or a dictionary entry url in case the tool tried to find keywords where there aren't, inside a more obscure dictionary word.

 

[Gube]

If you're training an automated appraisal tool, does it make sense to let the user give it hints, for fine tuning the result? I am thinking of suggesting the root words for creative blends, or a dictionary entry url in case the tool tried to find keywords where there aren't, inside a more obscure dictionary word.

If it's automated, as mentionned earlier in the thread, it won't do the job accurately.
Here the idea would be to complete manually what the tool can get directly and correct parts he might have got wrong, for example, if you know the end-user, it could be more accurate by adding the data you have.
Or if the system got the words separation(s) wrong, it can be adjusted

I will start working on a draft project soon, and share the method + formulas used for all automated parts (made in-house) in case it might help others.
It won't be done in one-time and will likely need improvements, formula changes, but it's likely worth it

 

[h2o]

I would also say "Brandability" is a big factor as well that is implied in those multiple choice options but not explicitly stated. I have seen super memorable, catchy domains that could easily be used for multi-million dollar national ad campaigns but the data from Estibot or GD Appraisals, for example will only show them as being only worth a few hundred dollars at most.

You need to have good intuition and judgement as well as looking at "all of the above".

 

[kemjika11]

ALl i know and live by is this :

Short names in good/decent extensions = best insurance collateral the common domainer can get.

 

[cooljub]

It's an impossible task Gube, you may as well give up now.

 

[Gube]

It's an impossible task Gube, you may as well give up now.

It might sound difficult, but not impossible

 

[ivey]

Brandability #1

 

[Ostrados]

It might sound difficult, but not impossible

Only possible using deep learning AI.
Forget about finding a formula or equation it will never work.

 

[Ostrados]

BTW GD claim that their appraisal tool is using AI, their appraisal tool did improve a little but no big improvement that you would expect from a learning AI, and it is still far from being a reliable appraisal tool.

 

[biggie]

What makes a domain valuable (or not) in 2020 ?

for some domains, it's the same things that made them valuable in 2019.
and for others,
it could be any news surrounding an event, product, service, etc that occurred in 2020 or anticipated to launch in near future.

as for the appraisal bot:

unless you can program common sense into that time sucker
then....do we really need another?

imo...

 

[Name Trader]

@biggie - isn't common sense programmable?

 

[biggie]

@biggie - isn't common sense programmable?

obviously, it's not

having been around some years before the bot, and seeing all the competitors and rivals which have came after,
one can see the effect they have had on domainers, who started after the bot, over the years.

on NP in particular, at least 2-3 times a year, there is a "how accurate is bot" post.

and.., when you view the lists of sales threads with bot evaluations, there is plenty of evidence that majority of those registrations, were not made using a "common sense" approach.

in the argument of bot appraisals, there are two sides
the builder of the tool
and the user

and to me, common sense says
the more dependent the user becomes on a tool for pricing,
the more power the builder, has over prices.

if you can find some fallacies there, i'll listen.

imo….

 

[NickB]

"An insightful opinion piece on the future of common sense in AI."

https://www.kdnuggets.com/2016/08/common-sense-artificial-intelligence-2026.html

give it a bit of of time and the machine will develop / have common sense - for now no, so trust those appraisal tools at your peril

 

[ThatNameGuy]

"An insightful opinion piece on the future of common sense in AI."

give it a bit of of time and the machine will develop / have common sense - for now no, so trust those appraisal tools at your peril

Seriously...i just got an email from a friend, Subject; Toilet Paper Tips

"Toilet Paper will go twice as far if you use both sides"

Is this the sort of common sense AI can add to the appraisal process?

 

[Croissant]

Thank for the poll Gube

(of course all options/parameters on the poll are important)

my answer is:
I would like to repeat, Dictionary =)

 

[Samer]

Thank for the poll Gube

(of course all options/parameters on the poll are important)

my answer is:
I would like to repeat, Dictionary =)

too many dictionary words to count;

many unknown ones; i dont like that answer;

How will you measure “dictionary”“?