I know estibot is just a tool but I have a question about the value from estibot. Is there improper weighting on .com from every other domain. I know .com is king, but I noticed a $40,000 .com is only valued at $470 in a .net. I just think the balance is just a little off.
Estibot
- Thread starter Debt.tv
- Start date
The views expressed on this page by users and staff are their own, not those of NamePros.
- Impact
- 136
Thanks guys. Trying to make an accurate automated domain appraisal system is like learning to play the octraventral heebiephone - a pleasantly futile task (wrong number of mouths).
Having said that, I do believe I have already programmed the first such system that is beyond being just a novelty, and I will pursue my goal of making a system that can be a consistently useful tool in helping users valuate their domains.
Having said that, I do believe I have already programmed the first such system that is beyond being just a novelty, and I will pursue my goal of making a system that can be a consistently useful tool in helping users valuate their domains.
- Impact
- 198
If you don't mind me asking - Was estibot "tweaked" any before the T.R.A.F.F.I.C. lists were run ? Very impressive to see the relationship between Actual sold prices and the amount you came up with. Were those run individually or via the Bulk tool ?
(Yes I could answer the last one myself - But see no reason to waste your BW :p)
(Yes I could answer the last one myself - But see no reason to waste your BW :p)
- Impact
- 136
Mark, no, they were not "tweaked", those were just the results of a pure automatic appraisal based on the domain name itself. The names were run by the Bulk Tool http://estibot.com/batch.phpMark said:
Note that while the Bulk Tool is great for generic keyword domains, it does not do a very good job with anything else, and usually does not even give an appraisal to subpar names.
- Impact
- 198
Thanks 
Was just curious. :tu:
People will always argue whether there is anyway to actually appraise domains - and which methods work best if they do put faith in them. I think you are definitely on the right track myself ...
Considering how many factors it really takes - and the fact this tool is real time ..... I can't imagine anything or anyone competing with you any time soon.
Domainers are some of the worst "Visitors" to websites when it comes to monetizing "IMO" - So I wish you Great luck on balancing the costs vs. income on it :tu:
Was just curious. :tu:People will always argue whether there is anyway to actually appraise domains - and which methods work best if they do put faith in them. I think you are definitely on the right track myself ...
Considering how many factors it really takes - and the fact this tool is real time ..... I can't imagine anything or anyone competing with you any time soon.
Domainers are some of the worst "Visitors" to websites when it comes to monetizing "IMO"
- So I wish you Great luck on balancing the costs vs. income on it :tu:
- Impact
- 136
mlee said:
It's true. Here is a quote from http://estibot.com/guide.html
EstiBot is optimized for the low-to-mid end domains, because you can't optimize for both ends, and the great majority of domains are not high-end domains and the great majority of domainers mostly want to research low-to-mid end domains.
With a name like invest.com, you don't need an appraisal of any kind
estibot.com said:It's true. Here is a quote from http://estibot.com/guide.html
EstiBot is optimized for the low-to-mid end domains, because you can't optimize for both ends, and the great majority of domains are not high-end domains and the great majority of domainers mostly want to research low-to-mid end domains.
With a name like invest.com, you don't need an appraisal of any kind
Well. If that is the case, it would be very helpful you have this statement posted on your website. I happen to have few high end names. I got few offers which claims that value should be near Estibot. I refused the offers but did not really know why. Now I know why I refused the offer.
in my opinion, I think estibot's valuation of .com and .net are usually very close to home but CCTLDS are not part estibot's strongpoints.
that is bcos it is almost impossible to ahve an instatnt appraisal tool that can factor all necessary factors including CCTLDs to come up with a perfect estimate.
That being said estibot is doing a wonderful job of giving us an idea of good domains but the final decision will always be up to us, the domainers and the buyers.
VAlue is relative what is valuable to one is useless to another
that is bcos it is almost impossible to ahve an instatnt appraisal tool that can factor all necessary factors including CCTLDs to come up with a perfect estimate.
That being said estibot is doing a wonderful job of giving us an idea of good domains but the final decision will always be up to us, the domainers and the buyers.
VAlue is relative what is valuable to one is useless to another
estibot.com said:
Does it take the age of the domain into consideration? I've noticed pretty ordinary domain names fetch $xxx just because they've been registered for a long time - its a consideration.
A field to enter in average monthly income of the domain. You can use it to show the estimated value with that factored in.
- Impact
- 136
mlee said:
Well, I've tried to pack EstiBot full of disclaimers and information...there are always those who don't even read the most obvious ones, like the keyword help on the front page.
I doubt that anyone seriously considering buying Invest.com is going to base their offer on an automatic appraisal - they are just as likely just "taking the piss" - but you can direct them to http://estibot.com/guide.html - I've added yet more disclaimers there now, including the statement you suggested (thanks for the suggestion, appreciated).
gwkg said:
This feature is coming next. I'm just right now analyzing the effect that domain age has on the sales prices.
Thanks. I'll look into it.
Last edited:
- Impact
- 1,230
Bottomline, only you know the inherent value of your domain. You might have some local / social insight into the particular niche.
Plus the variables include stuff like traffic, revenue and ppc value which estibot does not take into account. It uses alexa traffic figures and which are US centric to say the very least and ditto for ppc value.
But estibot throws up some figures which are not available anywhere else together, like the ones mentioned in my previous post. So its a very useful tool, and in the right hands it can do enough to allow you to do the rest easier.
Plus the variables include stuff like traffic, revenue and ppc value which estibot does not take into account. It uses alexa traffic figures and which are US centric to say the very least and ditto for ppc value.
But estibot throws up some figures which are not available anywhere else together, like the ones mentioned in my previous post. So its a very useful tool, and in the right hands it can do enough to allow you to do the rest easier.
Last edited:
I have been trying to understand what I get when Estibot gives me a nice number and I think I have found a plum.
Have you considered that sales records are incomplete data? Not the unrecorded, private sales, although those exist. This is bigger - how about all the domains that were not sold at all? The vast majority of them. Reserve not met at auction. Sitting on Sedo year after year. And yet the pool of inquiries to Estibot actually leans towards those unsold and unsalable domains because those are the domains that drop and are being offered in the aftermarket.
I use Estibot a lot, as a first filter. Just trying to understand what that valuaton means.
Have you considered that sales records are incomplete data? Not the unrecorded, private sales, although those exist. This is bigger - how about all the domains that were not sold at all? The vast majority of them. Reserve not met at auction. Sitting on Sedo year after year. And yet the pool of inquiries to Estibot actually leans towards those unsold and unsalable domains because those are the domains that drop and are being offered in the aftermarket.
I use Estibot a lot, as a first filter. Just trying to understand what that valuaton means.
- Impact
- 1,230
accentnepal said:
I fail to see how or why unsold domains should be included in the valuation process to start with.
Someone's asking price is really no indication of the actual worth of a domain.
accentnepal said:
When they are sold in the aftermarket...if they are, the value is reflected then...imo.
accentnepal said:
Zigackly.
Well my question was really to Esa, but I will try to clarify.
When you enter a domain name into Estibot various factors are considered in producing an estimated value of that name. A prime factor, as Esa just mentioned, is previous domain sales. But previous domain sales represent only a part of the domains offered for sale - those that actually did sell. So previous sales data is biased towards the better (most salable) names in any sample. Estibot tries to identify why those names are better, but I suspect it is comparing domains that sold for small prices vs those that sold for large prices, because I cannot see how the universe of unsold domains can be included. (I agree that the price unsold domains are offered for is irrelevant, that is not what I am talking about)
On the other side, the names searched in Estibot by domainers come from a different pool - the domains that have not yet sold. The domains previously sold to end users over the last ten+ years are not usually searched by domainers, neither are the domains that long term domain consolidators have locked away. So the average value of the domains we check does not represent the average value of that type of domain, it is the average value of what is left available in the market.
So we are looking for the value of the pool: (domains available for sale)
and Estibot is basing it's values on the pool: (domains that have sold recently).
Not the same thing and there is a bias of quality toward the second pool.
Getting much too complicated for what was just intended as a bit of trivia.
~~~~~~~~~~~~~~~~~~~
While I am here - what is with the differences I am seeing sometimes between the Google results in Estibot and those on Google itself?
domain: announced.***
Frequency (Google) 802000000
Results 1 - 10 of about 157,000,000 for ANNOUNCED
If there is a bug in there then that could really mess up results - or are you getting data from a different source?
When you enter a domain name into Estibot various factors are considered in producing an estimated value of that name. A prime factor, as Esa just mentioned, is previous domain sales. But previous domain sales represent only a part of the domains offered for sale - those that actually did sell. So previous sales data is biased towards the better (most salable) names in any sample. Estibot tries to identify why those names are better, but I suspect it is comparing domains that sold for small prices vs those that sold for large prices, because I cannot see how the universe of unsold domains can be included. (I agree that the price unsold domains are offered for is irrelevant, that is not what I am talking about)
On the other side, the names searched in Estibot by domainers come from a different pool - the domains that have not yet sold. The domains previously sold to end users over the last ten+ years are not usually searched by domainers, neither are the domains that long term domain consolidators have locked away. So the average value of the domains we check does not represent the average value of that type of domain, it is the average value of what is left available in the market.
So we are looking for the value of the pool: (domains available for sale)
and Estibot is basing it's values on the pool: (domains that have sold recently).
Not the same thing and there is a bias of quality toward the second pool.
Getting much too complicated for what was just intended as a bit of trivia.
~~~~~~~~~~~~~~~~~~~
While I am here - what is with the differences I am seeing sometimes between the Google results in Estibot and those on Google itself?
domain: announced.***
Frequency (Google) 802000000
Results 1 - 10 of about 157,000,000 for ANNOUNCED
If there is a bug in there then that could really mess up results - or are you getting data from a different source?
Last edited:
I have to agree with what accentnepal said about the unsold names. Lets say there are 100 names all registered in the same category, basically the same name. If they are all for sale for $1000 and 1 ends up selling, even if the other 99 never sell, they continue to get renewed or dropped, then past sales data would value these 99 names that seem to be worth nothing, because they are unsalable, at $1000. Any system based on past sales data would give a value to these name that appear to actually be worth reg fee. That is why the data from unsold domains would be valuable, if it could possibly be implemented in an automated system like that. "This many LLLL.ws have sold over this time period, but this many have not". That line would convince many people to stay away from .ws, even though some have sold for a lot.
This is where you were going with this, right accentnepal?
Also, this is not a knock on estibot, which is a very good automated appraial tool.
This is where you were going with this, right accentnepal?
Also, this is not a knock on estibot, which is a very good automated appraial tool.
- Impact
- 136
Accentnepal I think you hit the nail on the head. That's why your head is hurting
The pool of unsold names is important. So is the pool of private/NDA sales. They are the dark energy of the domain valuation universe.
Then, of course, there is the pool of unregisterd domain names, which is practically infinite.
The vast majority of registered domain names are never sold. Therefore, their value is reg fee by definition.
EstiBot tries to look at the potential value. The way I built my program is that it puts itself in the position of a motivated buyer, either end user or a reseller with great expectations of future value.
It takes the domain name, and compares its characteristics to similar names that have actually been sold. It transforms the domain name into numerical representations of the properties of those keywords (frequency, search popularity, PPC advertising bids, and many more), and of the value of other domain name characteristics, such as tld, hyphens, numbers, length, Then, based on comparison of a large database of previous sales, it performs a statistical calculation, and it gives you a price that it would be willing to pay you for that domain as a motivated buyer.
A price that, given all the data, is most likely (statistics, probability stuff) to reflect the qualities of the domain name in today's market.
This method results in a bias, just like Accentnepal said, towards the values of the pool of previously sold domain names. It ignores the unsold premium names and the unsold and unregistered subpar names.
In my experience, this generally leads to overestimation of the value of low-end domain names, and underestimation of the value of high-end domain names. Generally means just that - not always, but more often than not.
I've built dozens of error reduction routines. Once the value has been calculated, it will be analyzed against various known error sources, such as high Overture when in fact the Overture figure is unreliable. Many more similar routines check for other anomalies in the domain name numerical data. These anomalies cause EstiBot to give penalties, which result in a decrease in the final valuation.
This is why I always say: Great, your domain name was valued at $850. But remember, this is the estimated market price for your domain assuming it is actually sold. which is like saying, "if you sell this domain it may occur at a price in this ballpark, but if you don't sell this domain, the value is reg fee".
The most important thing is to look at the valuation number as a rough pointer, a quick-look summary of the desirability of the domain name according to EstiBot.
What it cannot do is give you a probability of selling your domain within the next 2 years. I would love to add that to the analysis, but I haven't figured out how to do someting like that yet.
The pool of unsold names is important. So is the pool of private/NDA sales. They are the dark energy of the domain valuation universe.
Then, of course, there is the pool of unregisterd domain names, which is practically infinite.
The vast majority of registered domain names are never sold. Therefore, their value is reg fee by definition.
EstiBot tries to look at the potential value. The way I built my program is that it puts itself in the position of a motivated buyer, either end user or a reseller with great expectations of future value.
It takes the domain name, and compares its characteristics to similar names that have actually been sold. It transforms the domain name into numerical representations of the properties of those keywords (frequency, search popularity, PPC advertising bids, and many more), and of the value of other domain name characteristics, such as tld, hyphens, numbers, length, Then, based on comparison of a large database of previous sales, it performs a statistical calculation, and it gives you a price that it would be willing to pay you for that domain as a motivated buyer.
A price that, given all the data, is most likely (statistics, probability stuff) to reflect the qualities of the domain name in today's market.
This method results in a bias, just like Accentnepal said, towards the values of the pool of previously sold domain names. It ignores the unsold premium names and the unsold and unregistered subpar names.
In my experience, this generally leads to overestimation of the value of low-end domain names, and underestimation of the value of high-end domain names. Generally means just that - not always, but more often than not.
I've built dozens of error reduction routines. Once the value has been calculated, it will be analyzed against various known error sources, such as high Overture when in fact the Overture figure is unreliable. Many more similar routines check for other anomalies in the domain name numerical data. These anomalies cause EstiBot to give penalties, which result in a decrease in the final valuation.
This is why I always say: Great, your domain name was valued at $850. But remember, this is the estimated market price for your domain assuming it is actually sold. which is like saying, "if you sell this domain it may occur at a price in this ballpark, but if you don't sell this domain, the value is reg fee".
The most important thing is to look at the valuation number as a rough pointer, a quick-look summary of the desirability of the domain name according to EstiBot.
What it cannot do is give you a probability of selling your domain within the next 2 years. I would love to add that to the analysis, but I haven't figured out how to do someting like that yet.
- Impact
- 38
I like it for my site PeterFoti.com is said to be worth 3100 with traffic
PeterFoti.com is said to be worth 3100 with traffic It seems as though Estibot tends to UNDERvalue domains.
It valued one of mine, BeforeUsing.com at $5,900. That must be a great sign if it's undervaluing them! :hehe:
Or course I've also been told my domain is worthless. So.... I have not a clue.
It valued one of mine, BeforeUsing.com at $5,900. That must be a great sign if it's undervaluing them! :hehe:
Or course I've also been told my domain is worthless. So.... I have not a clue.
Last edited:
Similar threads
Polls of interest ➔
-
What are your strengths as an investor?- 188 voters
- 27 replies
-
What's your domain portfolio size?
- 474 voters
- 139 replies
-
How do you think a potential buyer first checks if your domain is for sale?- 533 voters
- 52 replies
Hot this week ➔
-
Stop chasing buyers — make them come to you.
- 37 replies
- 3 points
-
Is Sedo Quietly Becoming the Hottest Marketplace Again?- 34 replies
- 1 point
-
Official Thread for Reasons .SI sucks as an Extension- 23 replies
- 3 points
-
.Vegas After 10 Weeks: What the Data Shows About This Geo-TLD- 17 replies
- 4 points
Hot this month ➔
-
The $99 Experiment Is Teaching Me More About Domain Liquidity Than Appraisal Tools Ever Did- 178 comments
- 67 points
-
Feedback on DM offers- 119 replies
- 17 posters
-
Domains wanted: 5k-10k(per item)
- 53 replies
- 5 points
