According to stats release by Dnpric.es today, “After falling for three years in a row, the average price of reported domain sales soared back to 2010 levels”
Here are the last eight years of reported sales according to Dnpric.es:
| Year | Total sales | Average price | Number of reported sales |
| 2014 | $60,800,227.98 | $4,668.32 | 13,024 |
| 2013 | $109,261,899.19 | $2,512.98 | 43,479 |
| 2012 | $112,514,132.70 | $3,364.66 | 33,440 |
| 2011 | $123,138,738.12 | $3,911.77 | 31,479 |
| 2010 | $235,485,073.34 | $4,557.04 | 51,675 |
| 2009 | $134,176,425.31 | $2,413.51 | 55,594 |
| 2008 | $196,158,730.58 | $3,411.40 | 57,501 |
| 2007 | $205,314,433.77 | $7,232.69 | 28,387 |
So as you can see according to the stats the Average sale price of reported domain sales fell in 2011, 2012 and 2013 from the previous year.
2014 so far, marks a substantial increase. Of course this year is only 1/3 over and but so far the numbers are promising.
not in my book. sales prices increasing every year since 2009 and topping this year so far.
Some inaccuracies on the site like showing Golf.tv selling for $600,000
Q3) There is an old myth that has gone around for a long time, it stated the domain Golf.tv was acquired for $600,000. There was controversy around the name that a gentleman from South Korea won the name but the original DOT TV Corp never awarded him the name. Did you pay $600,000 for Golf.tv ?
Igal: No I did not pay anywhere near $600,000 for Golf.tv. I cannot disclose the price but it was not $600,000.
http://tldinvestors.com/2011/07/a-chat-with-igal-lichtman-mrs-jello.html
How about the median price? Any word on that?
Yes, million dollar plus sales can skew the average. It would be better to report median sales prices.
DNPric.es does a good job of supplying raw data on past sales. And I don’t dispute the underlying idea that sales prices are back up after 3 years of a recession. It’s also true — as people have pointed out — that real median sales figures would be useful.
However, I don’t think that reports deriving from DNPric.es — or any other repository of past sales data — should be accepted uncritically. Under solid scrutiny, such reports will fall apart as inaccurate or inconclusive.
That’s not meant to disparage the work being done by people who prepared these figures. They’re measuring the data they have. Great! But drawing conclusions from that data is problematic. And I suspect that most people won’t realize just how fraught with inaccuracies such studies generally are.
Let me explain a bit.
DNPric.es or NameBio.com or DNSalesPrice.com, et al. — They all have large banks of past sales data. Raymond Hackney has pointed out that some of those data points are probably erroneous. But most are legitimate. So let’s assume, for the sake of argument, that 100% of their numbers are sound. Even so, most domain sales are not reported. Fewer than 5% of sales I’m involved with get reported, and I assume that’s typical for domain investors.
So clearly past sales databases are incomplete. Incompleteness isn’t a problem so long as the statistical sample is representative of the population of domain sales as a whole. Random samples are best. But there is nothing random about which domain sales get reported:
Certain market places report long lists of sales.
Other market places report nothing.
Private sellers tend to report nothing.
Domains in certain languages and ccTLDs are reported less frequently.
Sales below a lower threshold are often excluded.
High sales are frequently covered by NDAs.
After all those distorting influences, the sample of reported sales cannot be expected to accurately depict the whole picture.
That would be OK. We could simply understand that other sales are out there and discuss the trends in reported sales, right? Wrong! You see, the factors that affect which domain sales get reported are not constant over time. They vary along with domain prices:
Maybe NDAs are more common in 2014 than they were in 2008.
Maybe more sales of a particular kind are occurring through a market place that reports its sales, due to marketing done by that market place or contracts signed for distributed listings.
Maybe one year another market place suddenly began reporting sales so that its numbers are mixed into the sample in 2014 but not in 2010.
Maybe the threshold for reporting sales changed from $10k to $2k or from $1k to $2k, causing previously lower or higher numbers to be included or excluded.
Remember, DNPric.es and similar compile sales data from many sources. When a new source appears, its data gets added to the pile. When a new source vanishes, its old data remains there to be counted for some years but not for others.
Yes, it would be nice to report median sales data. But you won’t get the true median from any site that compiles sales data from multiple sources — most of which have lower reporting cutoffs.
This data is extraordinarily rich and varied and useful. The sales figures themselves seem to be mostly correct, which is very important. But the data is not a representative cross section of what’s going on. And that makes drawing conclusions about year-to-year trends impossible.
Even with a site such as Sedo, which reports mean and median sales by extension for each quarter, there are serious problems.
On the downside, Sedo is just Sedo — not the entire market. So they do better selling some kinds of domains than others. And they’re sales volume is lower than what’s reported in DNPric.es from multiple sources (including Sedo as a subset); so that means the fluctuations from random chance at Sedo are higher, especially when it comes to the mean being affected by a 7-figure sale or two.
But on the plus side Sedo’s data could be thought of as drawing comprehensively from roughly the same population of domains, year after year. That would make its sales trends more reliable, one would imagine.
Wrong as well! From year to year, Sedo has changed its network of market place partners. And Sedo has adjusted its marketing, as companies do, reaching buyers better in some areas and at some times than others. So not only is the inventory changing but the variable we’re attempting to measure (sales prices) is subject to the performance of the venue itself rather than just the strength of its inventory. Furthermore, Sedo will gain or lose market share relative to competing market places. And that won’t necessarily happen evenly across the domain sales price range.
All these factors, taken together, are a NIGHTMARE for any statistician. So I would caution anybody — domain investor or otherwise — from accepting reported numbers about domain sales. Even with accurate data and no malfeasance at all, knowing the truth is probably hopeless.
Correction: I meant to say that I’d caution anybody from accepting reported numbers about domain sales TRENDS or domain sales EN MASSE.
Thanks China.
One week on and Sedo just confirmed this bit too: