Hey guys, so this is pretty basic. Trying to get a general overview of how you all do it. Basically in my college class we are tasked with designing a very basic web indexer. Something i am interested in playing around with is how to rank a domain on value. Here is what i currently have. I'm having a hard time determining variable values. Suggestions would be great. (# of backlinks * A)+((Month/year created - current date) * B) + (Dmoz * C) + (Yahoo Dir * D) + (PR * E) + (TLD * F) = domain value Few notes: Domain value is an arbitrary number not its actual worth, a way of ranking domains in essence. TLD is extention: .com would get a value .net would get a value and so on Yahoo and Dmoz would either return a 0 or 1. ------------ So lets hear it, how do you rank the value of a domain. Looking for the most basic to most complex ideas. If i hear some good ones I will share my results Please note, this is for a college class not doing this for money, so fire away. I want to make it the best i can. Edit: there should be no human involvement, just the computer doing calculations.
That's tricky. I honestly don't trust or put much faith in any automated appraisal software. There are variables to be considered but there are so-many non factorable variables that automated systems make a lot of mistakes and misvaluations. I'll go ahead and list some variables, though, later on.. when I got a minute. bbl
This cannot be automated. Only way is to put it up for auction and let the demand for the name determine it's value as of that time. Even then so many potential bidders may not be involved in bidding that the final value (i.e. price paid) will be below what it's true value was at time of sale. But, I wish you luck with your domain-valuation tool.
I have tried doing regression analysis with a lot more variables and didn't even touch domain extension because it's so wild (only used .com). Most of other TLDs are highly speculative and I would say only .com (if any tld can be evaluated with any modicum of accuracy) is seriously apparaisable. also the strongest variable in determining domain value at least in my research was overture with extension scores. Link counts are difficult because different SEs do it different and I had some wild stuff that got negative coefficients on some SEs in determining price (which means there is obviously a problem if more links==less value at least in theory). Alexa was pretty worthless as well. If i were going to do it again i would look at search values (overture, keyworddiscovery, wordtracker), market (look at value of each market that it is in using economic indicators), PPC (ppc bids across multiple places, bidder volume as well as actual bids). in determining domain only value (not websites), i think those are probably some of the most important variables, stuff like links/age/pagerank/dmoz/yahoo (tested pagerank and directory values too, statistically worthless if i recall correctly) really doesn't make sense. thats my 3 1/2 cents on the subject and if you decide to actually pursue it... good luck. collecting and sorting that data is a pain in the... all the stuff i would like to look at is data youre going to have a helluva time trying to scrape or automatically parse (keywords are fun... ExpertsExchange.com or ExpertSexchange.com? Fun stuff for multiple possible keywords.
You should check out http://www.domainresearchtool.com/... Download the demo and start it up... They have a built in tool to calculate the value of a domain and you can check out the default and learn from that?
Thanks for the feed back! I really appreciate it. I've already kinda committed to this for class, so looks like I need to make it clear that Domain Value is a completely arbitrary value and that everything finally comes down to being human decision based on a bunch of other variables. I'll also stipulate that this will only be for .coms. I guess my final goal should be that this will rank domains by "statistical value" thus putting the ones to be human reviewed at the top. Thanks again for the feedback. Still open to hear more if anyone's got it.
Note: Your formula attempts to find website value not domain name value. Determining a value on a developed website is easier than determining the value of an undeveloped domain name. You are essentially trying to determine the amount of traffic the site gets as a means of estimating potential income. I would add Alexa.org rankings to your formula. Once you've estimated traffic and income you can then place a sale value of 1-2 years income on the website. A more complex analysis would try to find what category or keywords a website has and then assign that a weight (ie a site with higher paying keywords would be more valuable)
it's fun to see what sort of results come out. im not sure how youre approaching it, but the best R^2 i could get was around 44% i think (pretty worthless in terms of valuation). too much of a difference between end user and reseller purchases that throw any metrics out the window. also issue of full information in the market, it's not really an open/transparent market so people have lots of inside info. again, good luck. if you find anything interesting id be happy to chat about it or at least interested in seeing what sort of results/method you used. shoot me a PM when youve done a bit/finished if you want.
This formulae has nothing to do with valueing a domain for me. IMO you can't just create a formulae to value a domain name automatically. The value of a domain depends on the quality of the keyword(s), the extention, the development potential of the domain, the buyer eg. if it's a reseller or an enduser & also how much they really want the domain. Also, the traffic & parking revenue of the name.