Estibot: [estibot.com] Hot or Not?
Have you played with it?
Did it work for you?
Have automated domain appraisals gotten any better since the early days of GreatDomains 7-12+ page automated reports?
Since this only makes sense if we deal in specifics "naming names" will be okay for this thread - only. Pick a domain. Get it appraised. Report back why the automated value makes sense or makes no sense at all. Argue your case.
While your at it, does any other automated appraisal bot do a better job? Which one? Prove it - only if we all can have a go at it.
What ARE automated appraisals good for? Entertainment purposes only?
Automated appraisals: Hot or Not?
[edited by: Webwork at 5:56 pm (utc) on Sep. 15, 2008]
For the domain names I tried, its valuations ranged from less than one-tenth of what I paid for the domain to eleven times what I know was paid for the domain (no site or traffic yet).
Its traffic estimates are wildly inaccurate. The "valuation considering traffic" for one domain was less than the site earns in an average week.
Here's one tidbit that makes little sense:
MAX PPC Income/day $392.6
Overture/mo616908
Wordtracker/day725
[EstiBot.com...]
Domain Appraisal for domain name only : USD 6,600
It estimates Max PPC income per day at nearly four hundred dollars, but values the domain name at only $6600?
That's interesting math, to say the least ...
Hot or not? NOT
From EstiBot:
Detected keywords: "r v rental" -- Note please: If this is not correct, try again and define keywords by separating them with a space in the domain name.Domain: rvrental.com
Keywords (User-Defined) r v rental
Frequency (Google) 742000
in Anchor Text 5990000
in Title 50100
in URL 15800
Backlinks 61
PageRank 4
Alexa Rank 4352529 **
Traffic (Visits / Day) some**
PPC Ads # 36
Max PPC Bid $1.74
MAX PPC Income/day $52.2
Overture/mo 565
Wordtracker/day 13EstiBot.com Domain Appraisal : USD 13,000
Click Here to find out how to make money with RVRental.com
Hmmm . . How to make money with RVRental.com?
Step 1: Thy own counsel or that of thy well informed domainer friends thou shalt keep.
Step 2: Ignore the advice of thy well informed domainer friends and do your own deep research.
Step 3: NEVER FORGET: The inherent value of a type-in domain is THE VALUE OF THE CONVERTED LEADS IT GENERATES.
Here's one tidbit that makes little sense:
MAX PPC Income/day $392.6

Overture/mo616908

Wordtracker/day725
[EstiBot.com...] 
Domain Appraisal for domain name only : USD 6,600
It estimates Max PPC income per day at nearly four hundred dollars, but values the domain name at only $6600?
Its traffic estimates are wildly inaccurate. The "valuation considering traffic" for one domain was less than the site earns in an average week.
Estibot cannot possibly know how much your site earns in an average week, so it won't take that into account. You can calculate your site's worth yourself according to the revenue - Estibot was not designed to valuate sites, but domain names. The Traffic estimate and valuation is just a bonus feature. It can only make a rough estimate of traffic, and while it's often in the right ballpark, it's not always accurate (neither is Compete.com or Alexa)
Classic example: RVRental.com, reported sold via DNJournal for $325,000.00.EstiBot.com Domain Appraisal : USD 13,000
If you had used it correctly (typed in RvRental.com or rv rental.com as suggested on the front page (rv is a single word for these purposes), then you would have seen the valuation $190,000 which is in the right ballpark (especially if you read the interpretation guide). FIt also says on the results page, "Detected keywords: "rv rental" -- Note please: If this is not correct, try again and define keywords by separating them with a space in the domain name. So you should try again if the keywords are not parsed right.
I realize that it should parse the keywords better. I am in the process of improving the keyword parser, so users won't have to define the keywords because they usually don't seem to bother with that anyway.
Criticism is welcome, and I'm the first to admit the appraisal algorithm is far from perfect, but a lot of what I see is unfair criticism based on incorrect use and not bothering to read any of the explanations or the user guide, or the interpretation guide...webmasters are an impatient bunch :)
However, many professional domainers and domain companies use the appraisal tool extensively for business purposes - they use it as it is intended, a tool to help them make their own valuation, not as a literal valuation. My site doesn't make any big claims about the accuracy - please read the user's guide and the disclaimers that are all over the place.
The appraisal thingy is just one part of the site, it's mostly a professional level keyword / domain research service
Thanks
Detected keywords: "r v s"
Domain Appraisal for domain name only : USD 5,800
USD 11,000
Domain: rvs.com
Keywords (User-Defined) rv s
[EstiBot.com...] Domain Appraisal : USD 7,800
Domain: rv.com
[EstiBot.com...] Domain Appraisal : USD 190,000
Is RV.com worth less than RVRental.com? Don’t think so.
Would “the bot” have pegged RVRental.com before the sale? Possibly, but let’s compare the bots “success” in predicting “a recent sale” with a similar un-sold domain.
LI.com is a domain that was recently sold for $500K. It is reported in NameBio.com
Domain: li.com
Keywords (User-Defined) l i
[EstiBot.com...] Domain Appraisal : USD 530,000
So, you might think the bot “predicted” the sales price but let’s look at another 2 letter domain – LA, as in LosAngeles – population ~7 million:
Detected keywords: "l a"
Domain: la.com
Domain Appraisal for domain name only : USD 19,000
Detected keywords: "l a"
Domain: la.com
[EstiBot.com...] Domain Appraisal : USD 6,900
So LI.com was estiboted at $530,000, shortly after the sale was announced, BUT LA.com - which most folks know stands for Los Angeles comes in at $19K and $7K. That leads me to believe that the bot does a better job assigning values after a sale. However, to be sure I checked a few more “famous” 2 letter domains.
Detected keywords: "n y"
Domain: ny.com
Domain Appraisal for domain name only : USD 120,000
Proably $120,000. is off by about an additional 0+.
What about 3 letter domains, parcularly those with large geo-targeting relevance:
Detected keywords: "n y c"
Domain: nyc.com
Domain Appraisal for domain name only : USD 8,000
Estibot has trouble with 3 letter domains that stand for one of the largest cities in the world.
Let's try 4 letters:
Detected keywords: "f s b o"
Domain: fsbo.com
Domain Appraisal for domain name only :USD 1,200
Detected keywords: "f s b o"
Domain: fsbo.com
[EstiBot.com...] Domain Appraisal : USD 160
Seriously undervalued for the famous “for sale by owner” or fizzbo acronym. Again, no big surprise that the bot cannot ascertain it’s value. To the bot FSBO looks a lot like another 4 letter domain.
The bot also presents some interesting pricing patterns. Can you spot the pattern?:
Detected keywords: "hotels"
Domain: hotels.com
Domain Appraisal for domain name only : USD 1,200,000
Detected keywords: "travel"
Domain: travel.com
[EstiBot.com...] Domain Appraisal : USD 1,200,000
Detected keywords: "car"
Domain: car.com
[EstiBot.com...] Domain Appraisal : USD 1,200,000
It appears the bot may have a “default value” so be aware of that. I’m sure no one playing in this range would rely on the bot, but the values are illustrative.
Detected keywords: "real estate"
[EstiBot.com...] Domain Appraisal : USD 660,000
Hmmm . . YES, I'll take Hotels.com, Travel.com, Car.com and RealEstate.com for $3.86 million. In a heartbeat.
Now, here’s another quirk:
Detected keywords: "loan"
Domain: loan.com
Domain Appraisal for domain name only : USD 150,000
Detected keywords: "loans"
Domain: loans.com
[EstiBot.com...] Domain Appraisal : USD 3,300,000
I'm guessing Estibot was fed the data that Loans.com sold for $3 million but that the bot has trouble parsing the "s" and interpreting the value of Loan.com in the light of the Loans.com sale. Loan.com is not a $150K domain.
What about valuing .info domains?
Detected keywords: "garden"
Domain: garden.info
[EstiBot.com...] Domain Appraisal : USD 70,000
Garden.info sold at a Moniker domain auction a few months ago for $5,000. However, NameBio didn't record the sale, as I checked and it’s not in their database. $5,000 vs. $70,000.
I also checked a few high endser value domains, one’s where the value of a converted lead can be quite high, and therefore the underlying domain can generate huge enduser value with just a few clicks – driving up the sales figure considerably:
Detected keywords: "personal injury lawyer"
Domain: personalinjurylawyer.com
[EstiBot.com...] Domain Appraisal : USD 88,000.
Anyone who knows anything about personal injury law/lawyers knows that I’d pay $88,000 for THAT domain.
Domain: injurylawyer.com
Keywords (User-Defined) injury lawyer
[EstiBot.com...] Domain Appraisal : USD 37,000
From the above I infer that the bot has value estimating the likely market value of domains with high “enduser conversion values”. A single personal injury case can easily generate $100,000 in fees, making the value of the domain far greater than $88K.
What about localization?
Detected keywords: "new jersey injury lawyer"
Domain: newjerseyinjurylawyer.com
[EstiBot.com...] Domain Appraisal : USD 70
Detected keywords: "nj injury lawyer"
Domain: njinjurylawyer.com
Keywords (User-Defined) nj injury lawyer
[EstiBot.com...] Domain Appraisal : Reg Fee
Detected keywords: "nj lawyer"
Domain: njlawyer.com
[EstiBot.com...] Domain Appraisal : USD 10,000
Detected keywords: "new jersey lawyer"
Domain: newjerseylawyer.com
[EstiBot.com...] Domain Appraisal : USD 17,000
Detected keywords: "california lawyer”
Domain: californialawyer.com
[EstiBot.com...] Domain Appraisal : USD 46,000
Detected keywords: "la plastic surgeon"
Domain: laplasticsurgeon.com
Keywords (User-Defined) la plastic surgeon
[EstiBot.com...] Domain Appraisal : USD 70
Detected keywords: "los angeles plastic surgeon"
Domain: losangelesplasticsurgeon.com
Keywords (User-Defined) los angeles plastic surgeon
[EstiBot.com...] Domain Appraisal : USD 23,000
I’ll say, without hesitation, that the bot has trouble handling high value localized professional service domains. NJLawyer.com for $10K would have probably 200 law firms stumbling over each other to get to the domain first.
So, what's the take-away? Probably that a lot more data is needed to make value prediction reliable. Unfortunately, most domain sales go unreported. I've never reported a domain sale. Estibot was off by 50-60Xs the sale price of one domain I sold privately. It's not the bot's fault, though, it's the lack of data.
Bottom line: Most automated domain appraisal tools should be considered "for entertainment purposes only" as of this date. Each domain needs it's own work up, preferably by someone who is not only familiar with the domain vertical but also has a depth of knowledge about the enduser market.
Remember that the true enduser value of a domain name is reflected in the converted lead value to an enduser, someoine who is able to directly convert the lead to it's highest and best use/value. Someone who practices "injury lawyer" and is an "injury lawyer" or "personal injury lawyer" would likely pay a significant premium for exclusive control of a domain such as PersonalInjuryLawyer.com - appraised at $88,000.
As of 2008 a substantial number of domain sales are still domainer-to-domainer or "wholesale value" sales. Domainers are loath to pay more than 10% of a domains likely enduser value to acquire it. To the degree that Estibot relies on reported sales it is likely undervalue domains significantly.
[edited by: Webwork at 12:04 am (utc) on Sep. 22, 2008]
Detected keywords: "r v s" 
Domain Appraisal for domain name only : USD 5,800 
USD 11,000
Domain: rvs.com 
Keywords (User-Defined) rv s 
http://estibot.com Domain Appraisal : USD 7,800
May I stress first that I recommend reading the "guide" and the "about" pages - a lot of these questions are answered there already.
I think you may be putting spaces between the letters, because it's putting a lot of spaces there? Also, sometimes estibot parses the keywords incorrectly - that's why there is the user definition option (it's also because domainers want that freedom), which you should absolutely use. If estibot gives you "mo vie guide" as the keyword, it is obviously wrong and then you should try inputing "MovieGuide.com", capitalizing the keywords, or alternatively "movie guide.com" separating them with a space character.
Domain:
rvs.com
Keywords (User-Defined)
rvs
[EstiBot.com...] 
Domain Appraisal for domain name only : USD 110,000
Domain: rv.com 
http://estibot.com Domain Appraisal : USD 190,000
Is RV.com worth less than RVRental.com? Don’t think so.
Domain: rv.com
Keywords (Autodetected) rv
[EstiBot.com...] Domain Appraisal : USD 3,800,000
Would “the bot” have pegged RVRental.com before the sale? Possibly, but let’s compare the bots “success” in predicting “a recent sale” with a similar un-sold domain.
LI.com is a domain that was recently sold for $500K. It is reported in NameBio.com
Domain: li.com 
Keywords (User-Defined) l i 
http://estibot.com Domain Appraisal : USD 530,000
When Namebio data exists of a previous sale of the exact domain, estibot adjusts its appraisal to better reflect the known value, and estimates for increase in market value over time - that's no secret, just common sense
So, you might think the bot “predicted” the sales price but let’s look at another 2 letter domain – LA, as in LosAngeles – population ~7 million:
Detected keywords: "l a" 
Domain: la.com 
Domain Appraisal for domain name only : USD 19,000
Detected keywords: "l a" 
Domain: la.com 
http://estibot.com Domain Appraisal : USD 6,900
So LI.com was estiboted at $530,000, shortly after the sale was announced, BUT LA.com - which most folks know stands for Los Angeles comes in at $19K and $7K. That leads me to believe that the bot does a better job assigning values after a sale. However, to be sure I checked a few more “famous” 2 letter domains.
Again, without the spaces, la.com
Domain:la.com
Keywords (Autodetected)la
[EstiBot.com...]
Domain Appraisal for domain name only : USD 600,000
...et cetera. Whenever there are spaces between the letters in the "detected keywords" field, please try again. Find the correct input that will give you the most reasonable appraisal, and then, as recommended in the user's guide:
Really, take a look at [estibot.com...] and all will be clear, it even has a rough interpretation guide, which says among other things:
----quote----
EstiBot is optimized for low-to-mid end domains, because that's the range of interest for the vast majority of domainers. If you get an EstiBot valuation of >$100,000 or so, and you are unsure of the value of your domain, it may be worth getting a professional appraisal. The domain can be worth hundreds of thousands if not millions to the right buyer.
---end of quote---
So you see, any appraisal above $100,000 means you have a highly expensive domain in your hands and should not need estibot to evaluate the value, and the algo isn't optimized to tell the difference between a $100k domain and a $10 million domain, although it often does tell the difference quite well.
Domain: fsbo.com 
http://estibot.com Domain Appraisal : USD 160
Seriously undervalued for the famous “for sale by owner” or fizzbo acronym. Again, no big surprise that the bot cannot ascertain it’s value. To the bot FSBO looks a lot like another 4 letter domain.
Again,
Domain: fsbo.com
Keywords (Autodetected) fsbo (without the spaces)
[EstiBot.com...] Domain Appraisal : USD 180,000
The bot also presents some interesting pricing patterns. Can you spot the pattern?:
Detected keywords: "hotels" 
Domain: hotels.com 
Domain Appraisal for domain name only : USD 1,200,000Detected keywords: "travel" 
Domain: travel.com 
http://estibot.com Domain Appraisal : USD 1,200,000
Detected keywords: "car" 
Domain: car.com 
http://estibot.com Domain Appraisal : USD 1,200,000
It appears the bot may have a “default value” so be aware of that. I’m sure no one playing in this range would rely on the bot, but the values are illustrative.
Actually it does not have a default value, but it does have an error correction value. It checks the appraised value against the keyword data, and if it looks like it is probable that the appraisal is too low, it will nudge the domain name to the next category and it looks like for some domains it produces the $1.2m "minimum value". You may notice that the above domains have a different appraisal "considering traffic"
Detected keywords: "real estate" 
http://estibot.com Domain Appraisal : USD 660,000
Hmmm . . YES, I'll take Hotels.com, Travel.com, Car.com and RealEstate.com for $3.86 million. In a heartbeat.
Again I refer to the guide, above $100k means "very valuable" and does not attempt to make an exact valuation. As you said, owners of these domains would not rely on the bot, and if you read the guide, neither should you, but the bot can give you a lot of data and a ballpark appraisal to help YOU valuate your domain name. That's what it's for. The dollar value should not be taken literally.
Now, here’s another quirk:
Detected keywords: "loan" 
Domain: loan.com 
Domain Appraisal for domain name only : USD 150,000
Detected keywords: "loans" 
Domain: loans.com 
http://estibot.com Domain Appraisal : USD 3,300,000
I'm guessing estibot was fed the data that Loans.com sold for $3 million but that the bot has trouble parsing the "s" and interpreting the value of Loan.com in the light of the Loans.com sale. Loan.com is not a $150K domain.
Agreed, that's a quirk, but within Estibot's own rules it's still correct: Loan.com is definitely above $100k domain.
I’ll say, without hesitation, that the bot has trouble handling high value localized professional service domains. NJLawyer.com for $10K would have probably 200 law firms stumbling over each other to get to the domain first.
Agreed. Geo service domains is one of the major weak points, and I am working to fix that.
The rest of your post does point out errors in estibot's valuations. It has plenty of room for improvement, I consider it a good start for a system. I have a plan in place to take it to the next level pretty soon.
However, again I refer to the guide and the "about estibot" pages, which say among other things:
--quote--
More than anything, EstiBot is here to help you valuate your domain name. It will provide you with keyword data and domain registration information, and a ballpark valuation to give you a starting point for your own, critical appraisal of the domain name.
--end of quote---
It's a tool.
So, what's the take-away? Probably that a lot more data is needed to make value prediction reliable.
Agreed 100%
Remember that the true enduser value of a domain name is reflected in the converted lead value to an enduser, someoine who is able to directly convert the lead to it's highest and best use/value. Someone who practices "injury lawyer" and is an "injury lawyer" or "personal injury lawyer" would likely pay a significant premium for exclusive control of a domain such as PersonalInjuryLawyer.com - appraised at $88,000.
Agreed 100%
As of 2008 a substantial number of domain sales are still domainer-to-domainer or "wholesale value" sales. Domainers are loath to pay more than 10% of a domains likely enduser value to acquire it. To the degree that estibot relies on reported sales it is likely undervalue domains significantly.
Agreed and I am presently acquiring a lot more domain sales data and fresh keyword data in my endeavour to make a meaningful valuator, which, based on the current one, I think is quite possible to do.
You've got your work cut out for you since, as I said, most private sales to endusers are never reported and most professional domainers that I know are loath to disclose such info for various good reasons. Also, most folks that I know with quality portfolios aren't running their domains through Afternic, Sedo, etc. because they don't need the auction house's assistance to bring in offers.
I don't now how you convey this either, to users, but you really should place a caveat that auction sales - at least through 2008 - tend to be skewed towards wholesale domainer pricing,. THAT is a really important fact in getting one's head around domain value data. I think much of the bot's data is calculated to mislead IF someone uses "the tool" to estimate an enduser sale value. Potential buyers may get their hopes up.
I think the bot is nice try and that you are doing the best you can with what's available. However, any warnings - about not relying on the values or taking them for face value - likely should be much larger and front and center. Might lend, ultimately, to the bot's long term credibility - as it presumably will improve over the years as more data is accumulated.
You, with your value, you also might want to display some data on the number of comparable domain sales that "the system" relied upon to estimate value. You also might spit out a more relevant set of related sales, presenting not just a set of "same first word domains" (I just queried India Exports.com - which spit out "India domains") - but also spit out an equal number of "second word" sales. So, reported sales of domains with "India" in the domain and reported sales of domains with "Export(s)" in the domain.
Lastly, instead of 10+ affiliate ads per page you might lighten up a bit on ads. My $.02 is that by taking the lead in this space you might gain an advantage, but since most of the data you rely on is in the public domain, another competitor may come along and "best you" by simply offering a better user experience. Make the bot less "I wanna sell stuff" somewhat 1999-ish desperate spammy looking experience. That might just lend to the bots street cred.
Oh yeah, and encourage users to NOT post - in the various domain forums - that "Estibot estimated my domain to be worth" when they're looking for offers for their domains.
It just kinda doesn't work . . since it's sortof a tough crowd on the buying side. :P
[edited by: Webwork at 12:01 am (utc) on Sep. 24, 2008]
FYI, I tried the various 2,3 & 4 letter domains both with and without spaces between letters, so the new and improved value estimates you are reporting, for example - LA.com - is just that: new and improved.
No. I would not touch the algorithm for the purposes of pleasing a forum poster. IN fact I have not changed it in many months. It's pretty complicated and I hate to tamper with it - that's why I'm working on a whole new version. I've run these names before, and so have thousands of users, I don't know why you got the funny results. Maybe just try again.
for example - LA.com
You've got your work cut out for you since, as I said, most private sales to endusers are never reported and most professional domainers that I know are loath to disclose such info for various good reasons. Also, most folks that I know with quality portfolios aren't running their domains through Afternic, Sedo, etc. because they don't need the auction house's assistance to bring in offers.I don't now how you convey this either, to users, but you really should place a caveat that auction sales - at least through 2008 - tend to be skewed towards wholesale domainer pricing,. THAT is a really important fact in getting one's head around domain value data. I think much of the bot's data is calculated to mislead IF someone uses "the tool" to estimate an enduser sale value. Potential buyers may get their hopes up.
I'm aware of these things and am accumulating private sales data to address just this.
I think the bot is nice try and that you are doing the best you can with what's available.
However, any warnings - about not relying on the values or taking them for face value - likely should be much larger and front and center. Might lend, ultimately, to the bot's long term credibility - as it presumably will improve over the years as more data is accumulated.
You, with your value, you also might want to display some data on the number of comparable domain sales that "the system" relied upon to estimate value. You also might spit out a more relevant set of related sales, presenting not just a set of "same first word domains" (I just queried India Exports.com - which spit out "India domains") - but also spit out an equal number of "second word" sales. So, reported sales of domains with "India" in the domain and reported sales of domains with "Export(s)" in the domain.
It does display about 100 related sales for registered members, for both keywords.
Lastly, instead of 10+ affiliate ads per page you might lighten up a bit on ads. My $.02 is that by taking the lead in this space you might gain an advantage, but since most of the data you rely on is in the public domain, another competitor may come along and "best you" by simply offering a better user experience. Make the bot less "I wanna sell stuff" somewhat 1999-ish desperate spammy looking experience. That might just lend to the bots street cred.
As much as I'd like to reduce the ads, affiliate ads do work and they pay for the 2 dedicated servers and paid API requests it takes to run the thing. (Paid) subscribers do not see the ads nor the captcha.
I do agree that the site needs a redesign. It's a one man show and I'm focusing on developing the tools so don't have time to pay enough attention to design.
Re user experience, the site gets over 800,000 hits monthly, and has 2,500 registered members, and as I handle all support I can say that the vast majority of paying members have been quite content with what's on offer so I must be doing something right :)
Oh yeah, and encourage users to NOT post - in the various domain forums - that "Estibot estimated my domain to be worth" when they're looking for offers for their domains.
I don't think domainers need babysitting. I can't tell people what to post in forums and what not to post in forums..
Anyway, as it says on the site, the dollar value is just a rough pointer at best. It's all mainly about the keyword and domain finder tools.
Cheers
Domain Estibot Sold
groping.com $130,000 $5,000
seafoodrestaurant.com $47,000 $3,500
selfmotivation.com $45,000 $2,600
executivejetcharter.com $44,000 $2,900
absent.com $36,000 $1600
feels.com $25,000 $500
admitted.com $24,000 $2,250
commonstock.com $24,000 $2,500
men.org $21,000 $12,500
children.mobi $16,000 $800
collegenewspaper.com $11,000 $2,750
rattlers.com $7,100 $4,500
dirtyjokes.net $4,400 $650
educates.com $4,300 $1,000
beef.net $4,200 $2,750
Just an object lesson.
Domainers don't need babysitting. However noobs may benefit from a little bit of coaching as they begin their journey.
Hopefully noobs aren't basing their decisions to buy unregistered domains upon a conclusion - an impulsive decision - that there are quick riches to be had after the system reports that a domain they just checked is worth 10x, 20x or more than the registration fee - and it's available right now, so click here.
As someone concerned about newcomers not getting burned I'm glad to see that noobs are warned against such impulsive actions - but isn't the impulse actually encouraged - excited - by the appraisals in excess of registration fees?
Of course there's the flip side of a "registration value only" fee. "NJ Injury Lawyer.com" EstiBot.com Domain Appraisal : Reg Fee"
So there's a bit of balancing.
Well, hopefully there won't bells and lights that go off, in newer competing versions of such systems, when a noob "picks a winner". :P JK (Though I can see that coming, maybe not on your site.)
Currently, such bells ring and lights flash only in the minds of noobs.
[edited by: Webwork at 3:47 pm (utc) on Sep. 24, 2008]
Just an object lesson.
I'm aware of the results and the TheDomains post
I see two points in this lesson:
1) Hindsight is a wonderful thing.
2) While some of the appraisals are a somewhat high, most are perfectly reasonable at least from end-user pespective (which is what you originally argued the system should focus on) - and most domainers agree that the domains sold at disappointingly low prices
To the degree that Estibot relies on reported sales it is likely undervalue domains significantly.
I'm not entirely clear which way the argument goes here. I don't know why you are so fixated on taking the dollar appraisals literally (which I prominently advice against) and the argumentation is starting to have an ad hoc feel to it. I don't know if you read my responses, but if you did, you definitely disregarded any of my attempts at constructive discourse, and the tone is getting a tad self-important ("just an object lesson" / "so there's a bit balancing") and so I'll refrain from continuing this thread on my part as I don't think it's a good use of anyone's time anymore.
taking the dollar appraisals literally (which I prominently advice against)
Well, at least that still leaves us with the entertainment value ...
If you want to try it, just create a free account and then go to the contact page and drop me an email to claim your free month at Pro level. If you are not interested, I won't take offense..
I don't know why you are so fixated on taking the dollar appraisals literally
I'm not even close to taking the values literally, though it's values that you offer as the lead-in service. And, as I said, I suspect that it's values the system generates that help to sell other services, like on the spot domain registrations. So, in that context I think it's actually quite important to drive home the point, as you do, that the numbers aren't anything to get excited about. That, however, does not appear to stop people from listing domains for sale and reporting "their estibot value", leaving me to suspect that some users ignore the warning signs - and it's no wonder - as I discussed above.
Hindsight is a wonderful thing
In more ways that I care to analyze or express, that just about sums it up. The bot seems to have a knack for esti-mating values after the fact. Indeed, after the sale of RVRentals.com I could have sworn you attempted to suggest that "the bot got it right". Did it - beforehand? We don't know. However, due to the number crunching before the recent auction we can see that the bot's skill is in hindsight.
most are perfectly reasonable at least from end-user pespective
I'm beginning to understand that whatever the bot outputs there will be a rationale explanation. However, since a considerable amount of input to the system must be domainer-to-domainer, since that's what is most often publicly reported, the auction predictions should have reflected "that market". It didn't. No big deal. No one should rely on the bot.
Maybe you need to tune the bot up a bit and offer 2 values: domainer-to-domain values (wholesale, liquidity market at auction) and enduser sales.
I'm not sure but I'll guess that you haven't been segregating your data, the data the bot is using to appraise domains, between data/sales in the wholesale market and sales to endusers. If not, that's likely a system problem that needs to be corrected.
I'd also look to graph/input the variation in PPC revenue from parking and wholesale market prices. It's pretty clear to me that if you are going to use wholesale market data then you have to "normalize" for swings in PPC. That a domain sells today for $50K, when a "related domain" sold for $100K 12-24 months ago, may reflect a swing in market enthusiasm based upon excess PPC revenue (then) versus PPC revenue uncertainty in the here-and-now, and nothing much else.
However, on the enduser side of that traffic the value of a converted lead may remain substantially the same, which would argue for adjusting values or the weighting to account for "other factors".
So much to actually think about when pricing domains. One can only hope that no one is using the bot to make domain purchase or sale decisions, as you so correctly counsel.
[edited by: Webwork at 7:41 pm (utc) on Sep. 24, 2008]
