That funny ol' estibot

[Commerce]

I like to keep up with the value of names as does everyone else here because most all of us have a vested interest in such things in one way or another....

Specifically, I was interested in two letter .com valuations, as that cute little robot rodent Estibot, might see it (after all that resource seems to be getting a lot of press around here at DNF, so why not see what it has to say and put the little beast to the test). I selected 10 names for grins and comparison. Here are some results.

1) AA.Com - Even at $5 million, I'm not sure these owners would agree that this evaluation would fly.

2) DE.Com - Parked or not this name is clearly not ueber alles at $450,000.

3) LI.Com - While the price of the recent sale is about the $550K bot evaluation, the new owners might feel somewhat stranded at that value.

4) GM.Com - Would not drive them wild to find out their name is only $220K.

5) ON.Com - Their owners should be turned on to find out their name is worth $740K.

6) TA.Com - The investment firm associated with that name would probably never even entertain investing in their own name valued at $130K (they only do deals in the $60+ million plus range).

7) TI.Com - Even at $340K, its owners won't think that number adds up.

8) TZ.Com - The other end of the 26 T names their owners would probably be hard pressed to have to travel (at any speed) to a bid like $130K.

9) WM.Com would probably be surprised to find their name is worth "garbage" for a two letter .com at only $110K.

10) ZZ.Com is clearly not "Top" at $150K and one might guess that their owners would not sell it for all your hugs and kisses too.

In retrospect, it looks like for this rather unique set of names, some names clearly get different treatment than others and it makes one wonder how that happens? Well trust a robot to figure it all out for us!

My take? With only 676 of those type of names in this world of millions of domain names, each two letter .com is least one in a million. Especially so given the three letter sales out there today (even in a bad economy). Take your value from there.

I guess this goes to show you what folks with means, a taste for rare names and some well crafted scripts can produce. Not bad, but I think that the bot needs to consider adding another 0 on those valuations for the undeveloped names in the class we reviewed, never mind the ones actually doing something in brand and commerce.

What's your experience?

-Commerce

[sashas]

Commerce, there are some things that Estibot can't take into consideration. You say that GM.com is worth a lot more than Estibot's estimated 220k, but thats because you associate the company General Motors with GM. Estibot cannot do that.

Same thing for DE.com. For you, DE.com stands for Deutschland, but Estibot doesn't know that. It appraises the name just on the data it has, which cannot take into account the cultural associations and brandability of the name.

Sure, ZZ.com is worth a hell of a lot more than 150k, cause its so bloody brandable. But to Estibot, its just a repeating "bad letter" two letter .com..nothing special.

In terms of pure metrics, I think Estibot does a fair job, hell of a lot better than any other automated system out there

[Commerce]

sashas,

Actually, with a little more engineering, I think it could do quite a bit better job in such things that you say it does not do.

Were I to design such a bot, minimally, I would maintain a cached site home page to do a bit of basic checking. For example, from such a page I could determine if the page were parked, alive, active, etc. Clearly it is doing some of that for certain select sites, so there are some things it can do (or at least can be added manually or tweaked by its owners).

Case in point - eBay.Com - arguably not the greatest domain name in the world (of course only if looked at from the standpoint of a bot doing neutral evaluations), yet it comes up with an interesting answer relative to a search of something like uBid.com.

I can cite a variety of other seemingly odd discrepancies in the system which makes me more than a bit suspicious of this bot.

I guess before I put a whole lot of faith in the automation, it would make sense to clarify the metrics by which a conclusion is made. Almost certainly, end users will begin to grasp that they can use an automated tool to "get the value" of a name, which would be clearly not in the best interests of the community at large. The value of a domain comes from much more than simply one or two or even five metrics. That is why some folks make a business of doing name evaluations.

Heck even those inside the domaining community will doubtless use it as a starting point for evaluations for more massive "do you want to sell your domain name" email campaigns. It was bad enough getting those before such a tool existed, I can only imagine the exchanges that will come from those using such a tool. Can't you?

I agree with your point re ZZ.Com branding, but then, that applies to any two letter/number (or three character) .com name? Most humans can keep two or three characters in their minds for a period of time, especially if they are like. In other words, two letters or two numbers are easy to recall, nearly the same with three letters. With two letters, there simply are no "bad letters", so even discussing that re that class of names makes the person bringing up the "bad letters" argument look silly, if not entirely disingenuous.

Here is another case - SS.Com - Do you think for one second that this domain has a worth of less than 7 figures from a branding standpoint alone? The bot does. There is always a domainer argument about the "quality of letters", "duplicate letters" and other techniques used to bash down value. It seems to me, excepting certain exceptions that are seemingly "adjusted", that the bot uses these instead of some obvious metrics it does not.

Recall is certainly a far more intangible metric which can be scored in a model with just a bit of work. Character count should certainly offer a heavier weight that it obviously does today.

Ultimately, my point, one can do whatever one chooses when programming an evaluation "tool", before trusting it, no matter what "bot" it may be, just make sure to understand what it is actually doing to come up with its answers and what the motivation of its operators may be.

-Commerce

[EstiBot.com]

Hey, thanks for testing the tool, and thanks for the comments.

When running obviously valuable domain names it's worth noting that the tool was developed with mainly the lower end domains in mind - that's what 99% of domainers are interested in.

Here's a quote from the estibot.com user's guide:

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.

Guide to interpreting the appraisals:
EstiBot valuation $100,000 and above: Congratulations - if you own this domain name, you probably don't need EstiBot to tell you that this is an immensely valuable domain name and could be worth hundreds of thousands, or even millions to the right buyer.

EstiBot is designed to give you some solid keyword data and a ballpark dollar value appraisal - do not take the dollar value literally, just use it as a pointer - if it's high, it means that the domain is at least worth a closer look. There are no guarantees that anyone will pay that amount of money for your domain. Do not make purchase or sale decisions based on this appraisal alone.

In effect, according to the user's guide, a valuation of >$100,000 means that the domain can be worth millions, so the LL.com appraisals quoted above are meaningful.

Estibot doesn't make any big claims as to the accuracy of individual dollar appraisals - it's meant to be a helpful "quick look" measure of the various keyword metrics - I personally use it to get a very quick idea. If the dollar appraisal is high, the domain is usually worth further research.

Estibot is first and foremost a suite of tools for domainers - the appraisal algorithm is just one of the features. It's a work in progress, though I believe it's statistically already by far the best of all existing auto-appraisers.

Ultimately, my point, one can do whatever one chooses when programming an evaluation "tool", before trusting it, no matter what "bot" it may be, just make sure to understand what it is actually doing to come up with its answers and what the motivation of its operators may be.

-Commerce

I always say, don't trust Estibot...and it's written in bold letters on the site as well... It can be a helpful tool, but it doesn't give the user the licence to give his own brain a vacation

More about my motivation for making this tool here:
http://estibot.com/blog/?p=3

[GAMEFINEST]

Don't take things too deeply, its not like you are paying for it

[cursal]

I always say, don't trust Estibot...and it's written in bold letters on the site as well... It can be a helpful tool, but it doesn't give the user the licence to give his own brain a vacation

Hey EstiBot, Your candor and transparency are appreciated.

BTW welcome to DNF.

[WebsiteTraders.Com]

hi Commerce and Esa,
lots of hard work needs to be done to get that sort of appraisal tool.......every tool has its pros and cons............some people still sell genuine stuff.......i am no affiliate to Esa....infact i pay what everyone else pays when i use it..........but i can see all the engineering, programming, analytic, statistical, research as well as monetory efforts thats gone into the development of that tool.........n as it gets older.....Esa is improving it

[Commerce]

EstiBot.com,

I have to agree with cursal. Both responding and having a level of transparancy are great things to do in any community. Even better to learn you are a DNF Member.

That said, estibot itself needs to continue to be modified and matured in order to increase the worth and value of its answers. When releasing a new resource, it is bound to have some issues out the door. How those issues get addressed go a long way to adding to the credibility of the resource.

Finally, I think it is important for people to see what you are all about and also think that your post went a long way forward in doing just that. In replying the way you did, it shows your own character and gives others the ability to understand where you see things as well as to help them learn about where estibot can be most useful (as a promising gauge of value, but not as an authority).

I wish you well with your tool. For the record, would you take feedback offline via PM here from DNF members to help you in improving estibot's results?

-Commerce

[EstiBot.com]

Thanks-

Of course, I welcome your suggestions&criticism and even positive feedback is welcome

Feel free to PM me here on DNF

[Raider]

It also doesn't work well for brandable domains.

Also, I would re-think paying for a sedo appraisal, from my own experience they lowball in a attempt to sell more domains. If you want an appraisal, try Moniker or AppraiseDomains.

[Onward]

Get off it...estibot is a good tool for basic evaluation....