Estibot 2.0 official launch

[DomainingCom]

After a long beta period we are proud to present you the new version of Estibot.
A lot of improvements and new features have been added.

Thanks Esa and Luc for this great service.

 

[Carlton]

Estibot and .biz

To further the conversation on Estibot, I thought this public statement from Sedo is certainly worthy of consideration. It echoes my earlier point that Estibot needs to be modified so as to factor one character domain sales into an extension's value/weighting:

SEDO:

In terms of the best-performing extensions for average sales price, Sedo’s study saw the sharpest rise in .biz domains. This growth can be attributed in part to the growth in sales of one character .biz domains, which first became available through the .biz registrar in September 2009.

 

[Carlton]

For further discussion and analysis of the integrity factor ...

Just sold a .biz city + real estate domain to an end user. Estibot's stats are ...

Appraisal Value = $10, OVT = 13,000 searches per month, CPC = $4.60. The sale was astronomically higher than estibot's appraisal.

This illustrates the concern I have about anyone in the general market or domainer market using estibot for appraisal value.

That being said, it seems that estibot could have additional algorithms added that would clean up such terribly inaccurate results. My domain name was comprised of:

A huge metro city name + real estate + the .biz web extension. Real estate is obviously a perfect fit for the .biz (business) address. That estibot comes up with a $10 value means that it is specifically deficient in its mathematical analysis in at least 1 of 3 elements. They are:

1. Assigning proper value to a large U.S. city
2. Assigning proper value to the popular key phrase "real estate"
3. Assigning proper value to .BIZ (business)

 

[noy.com]

I just checked and seems like the problem with LLL.coms still is there, though, again those should be the easiest to valuate. So, again, I suspect some hidden interest in undervaluing those.

And, second point, I checked a random LLL.com, ilw.com, for value and related sales and it is interesting what related sales estibot uses to come up with value:

ilw.info, etailworld.com, mailwasher.com, civilwatch.com, briorailway.com.

Now, question is what kind of stupid logic is in using etailworld sale for valuation of ilw.com? or civilwatch or briorailway? Just because there is the same letter combination inside those words?

And these guys have an ambition to influence the market...

 

[accent]

Noy- I have noticed the appraisals of four letter domains also draw heavily on words in which those four letters occur. This is of course nonsense and another reason that a human is a required part of any Estibot appraisal. And I use Estibot daily.

There does not appear to be a way to over-ride the parser, which often causes bad valuations of LLLL.coms. Preferably there could be an option on the results screen that would add a line for the un-parsed version, to keep all the information in one place.

There is an option for "fast appraisal" below the input area, at this time it does not seem to be any different. I am hoping this is where the previous version of Estibot will be available. I think it gave a valid second opinion using partly different data. If there was a major difference in the appraisals it was a signal for further research. Actually I feel the earlier version was a little more accurate. Could be a comfort thing, but there is real value in having both valuations - I would prefer them on the same page, but I guess that would be confusing to others.

I do think we will be seeing end users presenting Estibot appraisals and expecting a discount from there. Giving "wholesale and retail" prices would only cause them to grab on to the wholesale version and want to work down. Most end users know very little about domain valuations so they cling to the little information they can find. Their chief fear is of paying too much (and the boss finding out). So, since Estibot is going to set the MAXIMUM value for a domain it should reflect only end user pricing. Domainers use it to establish relative values, the actual numbers are not important.

Perhaps there needs to be a limit on usage by paying customers, but 250 per day for domainer level is too low. Once you hit that max you can't even appraise single names. Perhaps there could be the ability to save unused appraisals from day to day? I don't do a whole lot most days but sometimes I get on a roll....

 

[estibot.com]

I just checked and seems like the problem with LLL.coms still is there, though, again those should be the easiest to valuate. So, again, I suspect some hidden interest in undervaluing those.

Suspect all you want. Like I said there is a bug and we're working on it.

And, second point, I checked a random LLL.com, ilw.com, for value and related sales and it is interesting what related sales estibot uses to come up with value:

ilw.info, etailworld.com, mailwasher.com, civilwatch.com, briorailway.com.

FYI those sales are not used to estimate ilw.com value. They are only displayed as an additional feature for the user's convenience. The 'related sales' list has nothing to do with valuation. They are simply fetched based on word / string occurrence. The valuation engine uses an entirely different method for matching comparable sales. For instance, in the case of LLL.coms, the valuation engine uses other LLL.com sales as one (important, but not the only) factor.

Now, question is what kind of stupid logic is in using etailworld sale for valuation of ilw.com? or civilwatch or briorailway? Just because there is the same letter combination inside those words?

Yes, that is a stupid idea indeed

 

[noy.com]

yeah, yeah... it is a bug... It is probably such a sofisticated algorithm behind the thing.. right... It takes few months to dig and still not find the cause or not to be able to fix it.

Regarding ilw vs etaILWorld, then don't call it related sales. There is NOTHING related between the two or it is as much related as a cow is related to an asteroid.

In fact, if you are as advanced in your "black box" as you claim, you could have your script to provide couple really related.

 

[Carlton]

More "appaisal" values for Namepros analysis

Thought I'd investigate estibot a little deeper to see what kind of valuation accuracy it might produce with geo domains in their native country code (cctld) vs. .com.

What's wrong with the following picture?

Canada.ca $9700
Canada.com $1,170,000

Montreal.ca $6200
Montreal.com $638,000

Paris.fr $1600 (All I can say is wow!)
Paris.com $1,280,000

London.co.uk $39,000
London.com $3,830,000

Birmingham.co.uk $21,000
Birmingham.com $298,000

Mexico.mx $40
Mexico.com $1,380,000

Cancun.mx $20
Cancun.com $1,050,000

Berlin.de $21,000
Berlin.com $524,000

I think these clearly illustrate the danger inherent in automated appaisals. The necessary "judgement" & logic required to appraise domain names is just not there.

Estibot's algorithms must receive much more intensive work. To the extent that estibot (and "Valuate.com") are advertised as appraisal tools, their danger to mispresent investor assets must be exposed in a big way.

 

[noy.com]

other examples of estibot's "sofisticatedness"...

wz.com is valued at $9300. I assume even in reseller market that name would easily sell for 25-30K. And, given that there are only 676 ll.coms, those should be easy to valuate even manually.

I assume, that is some bug in the super-advanced algorithm as well and would take 12-14 months to fix.

Or more from this market-forming machine:

infoad.com is valued at $50. And it is not surprising, as Estibot "thinks" that it is combination of "in+foad".

Another interesting one, Estibot values e-mail.com at mere $103K, the reason, I assume, being that it has programmed in heavy penalty for dash in it. And this is with 834000+ exact searches at $1.72 cpc.

I think the whole sofisticated algorithm story is bogus. It is quite simple ratios from some benchmark numbers.

 

[DutchBoyd]

I <3 Estibot... the best automated appraisals hands down. And I love all the new tools... especially the IDN tools. Thanks, Esa, for making it.

 

[noy.com]

best among existing pool is not much of a something...

I will also join those thinking that this version is step backwards.

btw, valuate.com is paying client of estibot that probably sponsored the 2.0 version and they are obviously interested in lower pricing of most liquid assets, including LL.coms and LLL.coms, since they have "sell it" link next to valuation. And actually on estibot as well there is option to ask for sell help.

 

[Carlton]

Estibot seems unable to understand geo domains & geo-targeted domain names. Examples ...

NewYork.us $3,800
NewYork.co.uk $26,000

NewYorkCity.us $1,300
NewYorkCity.ca $3,800 (Canada's extension)
NewYorkCity.co.uk $13,000 (United Kingdom's extension)

 

[mobihead]

As far as automated appriasals are concerned estibot 2.0 is an amazing tool - Sure it is not 100% accurate and there are probably a very large amount of odd cases around domain value but it gives a decent computerised guess around the possible value to an end user taking a few key elements into account.

One thing I have found strange about estibot is that while it obviously has a fairly good (prob best around) set of equations in place it does not include learning technology or feedback directly into its process, hence it is not a learning tool. IMHO neural nets are the way to go for automated domain valuation - probably a great PHD thesis for someone with a possibility of having a business at the end.

 

[hakita]

Have a check at DomainAppraisal.net
Pros : It handles 12 langages + can appraise up to 20,000 domains name in one time.
Cons : does not support all tld. ( no co.uk, .mx, .ca ... support)

Difference with estibot: DomainAppraisal.net makes appraisals based on search volume + cpc, and not historical sell data. This difference has its pros & cons.

Compared to Estibot with the domain you listed:

Berlin.de
190,000$ with DomainAppraisal.net Berlin.de domain appraisal
21,000$ with Estibot EstiBot.com - berlin.de

Because DomainAppraisal.net handles searchs volumes in germany, it makes more credible appraisals for geo domains.

I likes estibot as it is usefull to me. But my more secure approach it to mix several appraisal tools results with human subjective considerations.