According to a press release we received today the biggest domain name sales database ever built just just rolled out.
The database is at DNPric.es.
“”Currently the database has 315,461 domain name sale records totaling $1,135,277,224.00.
Not only you can lookup historic sales, the site also offers amazing stats section where you can fish for tons of interesting information.
Let’s have a quick look on it from different angles.
1. TLDs
While .COM is still the king with more than half of the sales (this correlates with the number of domain names) other extensions are doing surprisingly well.
For instance .ME, #9, topping even .tv and .co.
2. Brokers
Here we can narrow down the whole industry to five big players (Sedo, Moniker + SnapNames, GoDaddy+AfterNIC, TDNAM, NameJet, DomainNameSales), than another few dozens of boutiques.
3. Time
We can observe that overall average domain name prices are falling in the last few years. Especially those for .COM. And then, this year it started to improve.
4. Length
Interestingly, average prices for LLLLL names are higher than those for LLLL and LLL names.
Average historic price for five letter domain names (LLLLL.***) is $6,060.81.
That for LLLL is $3,641.76.
And that for LLL is $5,653.34.
It looks like names that are too short but not ultra short (two characters) have somewhat less value than those of five characters.
A paradox that intuitively I am still struggling to explain.
Can you?
The longest known domain name sold is xn--private-krankenversicherung-fr-selbstndige-6sd16h.de a.k.a. private-krankenversicherung-für-selbständige.de.
Well done Sedo!””
The database is available and accessible for free as of now.
Of course only publicly reported sales are included, which probably account for most of site developers issues regarding difference is longer domains selling for more on average than shorter domains.
The paradox could be that the 5 letters might have ‘meaning’…
Great data.
I wonder if they will report the 3rd Blockbuster DOT COM sale of the year:
“37 Dot Com Sold to Chinese Company for Nearly $2 Million”
as reported last night by DNJournal Dot Com.
Mike, when clicked, the link above is broken (points to Whizzbang).
How does this compare to tools such as NameBio?
Looks like NameBio, DomainTools and others list less sales for similar queries.
Acro
Link is fixed
Namely.pro is the developer of DNPric.es.
I like the interface & search options on DNsaleprice.com better. But it’s always good to have another source. Thanks for the link Mike.
How bout those #1 Gators
Try telling someone to go to DN Pric and see the look on their face.
This made my Friday…thanks MHB
It could also be that the a higher percentage of the “best” 4 letter domains had settled into permanent end-user possession before the domain sales aftermarket had developed, while the vastly broader field of 5 letter domains were still unsettled. Plus, there’s the issue of sales subject to NDA not being included, thus skewing the data by not including quite a few presumably disproportionately huge outliers.
Great to see another site with past sales records! The more sources we have to corroborate reports, the more legitimate they are likely to be and appear to outsiders — assuming independent data gathering and validation processes for each.
You know, until about a year ago, my primary project was building a domain appraisal system. I had hopes to create the first scientific methodology for domain name valuation, which would serve to legitimize the industry and bring in large-scale outside investment. The project is actually 100% do-able, but I’ve postponed it due to lack of access to data, the absence of backing (from a business angle), and the uncertainty of monetization. Yadda yadda yadda.
Seeing the confusion about the difference in average value between 4-letter and 5-letter domains reminds me how badly needed accurate information is.
The discrepancy is easily accounted for. Domainers have historically tracked sales of LLLL domains very closely. So a disproportionate number of LLLL domain sales have been documented — including sales in lower price ranges. Meanwhile, there is much less reporting on domain sales in other areas. Obviously, the inclusion of numerous lower-valued sales for 4-letter domains will cause the average price to drop.
If ALL 4-letter and 5-letter domains were tracked equally — including cases without sales — then we’d see the expected decay in average price per domain as length increases. But we’re only looking at reported sales, which are most definitely NOT a representative sample of the domain population. Inferences are dangerous and largely invalid.
Also, remember that multiple factors account for domain value simultaneously. Acronym value decreases as length increases. But semantic value can actually increase with length. For example, if you imagine the whole English lexicon as a sort of lopsided bell curve, then you’ll it peak at words of 8 or 9 letters. We also have 2-word phrases that are more richly descriptive than either word alone. For example, NewYorkCity.com RealEstate.com. So if you exclude acronym value, you’ll see semantic value peak with longer domains. These 2 effects are layered on top of one another — along with thousands of other factors. So it’s not unthinkable to see a double-peaked distribution either.
Think of it this way … We may be comparing sales of 4 and 5-letter domains such as these:
FOCU.com and OCUS.com and LOCU.com and LAUG.com and AUGH.com
versus
Focus.com and Locus.com and Laugh.com
There are 26 times as many 5-letter domains as 4-letter domains. But for each 4-letter string there are (roughly) 52 5-letter strings that differ by a single letter — any one of which could make the difference between AUGH.com and Laugh.com. There are 2.2 times as many 5-letter words in English as 4-letter words. And there’s a high probability of a word being found among the 52 overlapping-but-longer strings, which brings semantic value more to bear on the averaged price result.
Based on DNPric.es numbers, they have 48% as many reported 5-letter sales as 4-letter sales. If these two categories were selling equally (which they’re not, given the huge volume of meaningless 5-letter combinations) … and if these two categories were seeing sales being reported equally, then we’d expect 2,600% as many 5-letter sales as 4-letter sales. So the question remains basically this: How much of that discrepancy in reported sales is due to a difference in sales volume and how much of it is due to a difference in reporting volume.
Joseph,
Valid points. Also worth noting:
There are 1,774,224 LLLL combinations (456,976 if you exclude numbers and dashes). This is for each zone.
There are 65,646,288 LLLLL combinations.
Even if you focus on .COM alone, the secondary market simply is not big enough to accommodate this scale of speculation.
Keep in mind there are only 265 million domains registered as of the last Verisign report
Exactly. The question to ask is how many domain names can there be?
Namely.PRO, 10-15 billions – per one for each human on Earth and for each organisation, area, group, etc