Joseph Peterson takes an in-depth look at January’s domain name sales on NameJet.
January 2016 was a busy month for NameJet. Not only did the auction platform deliver its customary mix of expired domains and seller-managed inventory; it also handled the NamesCon auctions in conjunction with Right Of The Dot and (in many cases) a live auctioneer at the Las Vegas conference.
In presenting these January sales data, I’ve decided to exclude all domains associated with NamesCon, whether they were part of the live auction or not. This way, we can make apples-to-apples comparisons when looking at market performance over time. Inventory attached to the domain conference received extra exposure and included a big dose of nTLD domains. That’s perfectly fine, but such special circumstances obscure long-term trends. Better to write a separate article devoted to the NamesCon results than to mix them up with January’s regular auctions.
By excluding NamesCon, the statistics change in an interesting way:
- From 300 sales to 219.  NamesCon contributed 81.
- From $2.46 million to $1.53. Â NamesCon contributed $931k.
- From a mean of $8.2k to $6.9k. NamesCon’s mean was $11.5k.
- From an upper quartile (75th percentile)Â of $5.0k to $4.1k. NamesCon’s upper quartile was $8.8k.
- Median and lower quartile (25th percentile) are $3.4k and $2.7k, respectively. That’s with and without the NamesCon data – i.e. no change. By itself, NamesCon had a median of $4.1k and lower quartile of $2.7k.
This suggests that NamesCon auctions skewed toward a higher price range than regular NameJet auctions. In other words, some mixture of extra promotion, assiduous curation, and conference high spirits translated into bigger price tags. However, this mainly affected the conference “headliners”, since the lower half of prices for NamesCon and non-NamesCon auctions look similar.
Even with NamesCon results, NameJet fell about 20% shy of its December sales. However, that previous month did set the all-time record … and by a huge margin. More importantly, even without NamesCon results, NameJet nearly tied its 2nd best month ever, coming within 2.2% of November sales. Keep in mind, NameJet never broke $1 million until last July; but it has done so every month save 1 since then.
The reason for that uplift is easy to identify: China. But China must wait. This month, I’ve put the most active categories at the bottom and the smaller, less predictable categories nearer the top.
Domain | End $ | Domain | End $ |
---|---|---|---|
Oceans.org | 3050 | Apotheke.org | 2877 |
Contract.org | 2500 |
Let’s start with .ORG – 1.4% by count, 0.6% by revenue. Inside NamesCon, nTLD registries featured prominently at auction. Outside NamesCon, the only TLDs to cross $2k at NameJet last month were the usual suspects: .COM, .NET, .ORG.
Domain | End $ | Domain | End $ |
---|---|---|---|
Milford.com | 7500 | LaPlagne.com | 7055 |
PineHill.com | 5126 | Milford.com | 2500 |
Here’s a list of GEO domains. La Plagne is a ski resort; the other 2 match several U.S. cities and towns. No, your eyes do not deceive you; NameJet reported 2 different prices for Milford.com – both on the same date. As a matter of fact, I reported a third price a month before, during December. Apparently that $15.4k bidder didn’t pay, however, because no sale for Milford.com was included in NameJet’s December report. What then was the price: $15.4k? $7.5k? $2.5k? Until informed otherwise, I’d assume the lowest amount: $2.5k.
Domain | End $ | Domain | End $ |
---|---|---|---|
ePhoto.com | 37,000 | NPost.com | 3200 |
eVirtual.com | 2400 | eSeminar.com | 2352 |
Recently, I’ve heard bloggers say that prefix domains are a thing of the past. Maybe in the future they’ll feel dated. But right now market appetite for “e” + [keyword] names is demonstrably healthy. That $37k sale ranks 7th out of 200 during January. No categories outrank it but NNN, NNNN, and a minority of LLLLs.
Domain | End $ | Domain | End $ |
---|---|---|---|
58R.com | 7300 | TT11.com | 4601 |
HG267.com | 3376 | HG7767.com | 2679 |
Here’s your first glimpse of Chinese-style domains – a pair of short mixed-character items and a couple of numerical domains with that peculiar “HG” prefix. To date, NameJet has sold 23 “HG” numerics. At the peak of the Chinese surge last October, a single month cleared half a dozen of them above $2k. Actually, they first made the monthly reports in July 2014. And here’s a surprise: In terms of price, “HG” domains peaked earlier than you might think – April ($5.0k), July ($7.2k), and August ($4.4k) of last year.
Domain | End $ | Domain | End $ |
---|---|---|---|
Kabei.com | 10,099 | JieQin.com | 9000 |
JunGong.com | 8100 | XiSen.com | 7800 |
MaLing.com | 6500 | TouBai.com | 2420 |
PaoJia.com | 2033 |
Some of these appeared already in my weekly articles on the expired domain market: Kabei, Jun Gong, Tou Bai, Pao Jia. My assumption is that they’re all Chinese Pinyin, but that may be incorrect. If you have translations to add, comments are welcome. The highest Pinyin sale has been lumped with all the 4-letter .COMs later in the article: ReYi.com ($18.5k) actually means “heated debate” (çƒè®® / rè yì).
Domain | End $ | Domain | End $ |
---|---|---|---|
Cousins.com | 12,522 | Minty.com | 6888 |
Cathleen.com | 3800 | Prisoners.com | 2622 |
Sueno.com | 2526 | Scabs.com | 2300 |
Brasa.com | 2226 | Existential.com | 2101 |
Nilton.com | 2010 |
Next up: 9 single-word domains. Three you may not recognize I’ve written about already: Sueño, Nilton, and Brasa. Interestingly, the .ORGs we saw above all belong to this single-word category as well.
Actually, the top-selling LLL.com, LLLL.com, and .NET were all dictionary words. “BLT” ($65.5k) = a bacon-lettuce-&-tomato sandwich. DJ.net ($55.5k) implies music. And “TOOT” ($31.8k) is a word I wrote about at length.
Categorization can be a bit arbitrary. We could deprive those 3 short-domain categories of their 3 star sales, assigning them to a group of single words. That would raise sales from 9 to 15 and from $34.8k to$196.0k. Thus, depending on interpretation, single-word domains amounted to 4.5% / 7.5% by count and 2.3% / 12.8% by revenue.
Domain | End $ | Domain | End $ |
---|---|---|---|
BLT.com | 65,500 | TPZ.com | 60,900 |
QNE.com | 27,600 | NVL.com | 25,000 |
VDN.com | 23,800 | QVK.com | 23,433 |
JGU.com | 23,100 | AZG.com | 21,800 |
XZE.com | 19,200 | XVV.com | 18,800 |
QVJ.com | 17,600 |
Much as I enjoy the sandwich, BLT.com seems to have sold at a value not based on its scrumptiousness. After all, $65.5k is comparable to the price of TPZ.com ($60.9k), an acronym with no such popular meaning; and the buyer is Chinese. So I’d credit the LLL.com category. Altogether, LLL.com domains comprise 5.5% by count but 17.9% by weight.
Beneath that top pair, the remaining 9 LLL.com domains all contain either a vowel or a “V” – letters widely understood to lower the market value in China. While those “bad” letters do account for the steep drop-off in price under $60.9k, we’re left with something unexplained. Why is the price range among “non-CHIP” (CHInese Premium) LLLs so wide? Normally, China-favored categories sell at prices tightly clustered around a recognized market price. By comparison, $17.6k – $27.6k is an enormous spread.
I can offer 3 hypotheses:
- Market uncertainty leading to erratic buyer behavior.
- Hidden criteria that really do justify spending $10k more or less.
- Declining prices during the month of January.
There is some evidence to support the third hypothesis. The first 3 non-CHIP LLL.com domains to sell during January averaged $24.7k, but the next half dozen averaged only $21.0k. A least-squares regression suggests that (beneath the noise) prices were dropping $189 per day – a rate of $5.9k per month – during this period. Granted, it’s a small data set; and the correlation is fairly weak (R = -0.452). But the decline is plainly visible, whether or not it holds true generally or continues.
Many pundits with money at stake will tell you that the Chinese surge isn’t over, that this market sector will grow later in 2016. Perhaps. Neither they nor I know the future. But we ought to look for facts about the present. During December, 19 non-CHIP LLL.com domains sold between $24.0k and $55.6k. Their median then was $33.3k. A month later, January’s median dropped $10.2k to $23.1k. Let’s repeat: Prices dropped by 1/3 in a single month.
A theory I run across regularly says we’re seeing a lull in sales due to the Chinese New Year (February 8). Optimists proclaim that the Chinese surge will resume once that event (with its celebrations and symbolism) is out of the way. Yet that theory fails to explain 2 months of declining prices.
That plot shows non-CHIP LLL.com sales for December plus January. If we remove the single $55.6k outlier (AUD.com = “Australian Dollar”), then we see prices dropping at a rate of $237 per day ($7229 per month) over the course of 2 months. This 2-month data set gives us more confidence (R = −0.633). Does Chinese spending always decline steadily for 9 weeks prior to the Chinese New Year?
Domain | End $ | Domain | End $ |
---|---|---|---|
417.com | 169,000 | 3568.com | 45,700 |
4728.com | 41,200 | 1251.com | 39,210 |
70777.com | 34,501 | 3405.com | 18,000 |
89911.com | 8600 | 13713.com | 5900 |
5755555.com | 3519 | 998777.com | 3400 |
688818.com | 2333 |
NameJet’s top auction during January was a 3-digit .COM. Considering NNNs sometimes clear $800k, that $169k sale – big as it is – isn’t jaw dropping. In general, prices for numerical domains follow the number of digits. Yet there are 2 instances of leapfrogging in the chart above: (1) 5N ($34.5k) > 4N ($18.0k); and (2) 7N ($3.5k) > 6N ($3.4k / $2.3k). Repeating patterns are the rationale.
It’s remarkable that domains containing the unlucky “4” rank 1st, 3rd, and 6th. This isn’t strictly chance. It’s seller behavior. Look back at the LLL.com chart. Why do 9/11 of the domains contain vowels or “V”? Overall, 45.5% of LLLs are CHIPs; yet we see that best-selling group appearing as just 18.2% of NameJet’s top results!
This vanishing-CHIP trend began during December when 19/23 of LLLs contained “bad” letters. So it’s quite real. Compare that to January 2015, a year prior. That month also saw 23 LLL.com sales; but back then 39% were CHIPs – matching the broad 45% we’d expect.
I could explain this discrepancy in 2 ways:
- Sellers have already sold their best domains to China. If they now have few domains left to sell, then NameJet will now disproportionately find itself selling the “dregs”.
- While prices remain near the 2015 high, sellers are liquidating some of their Chinese-style assets, beginning with the weakest subcategories. However optimistic their predictions, actions speak louder than words.
If you have another explanation for the prevalence of “4”, vowels, and “V”, then I’d be happy to hear it.
Domain | End $ | Domain | End $ |
---|---|---|---|
WineGrower.com | 10,746 | BeanTown.com | 7600 |
GreenRock.com | 7200 | BlueSquare.com | 5756 |
BikeWorks.com | 4957 | CurlyHair.com | 4900 |
ChinaTech.com | 4712 | Cable Companies.com |
4100 |
PrimeMortgage.com | 3450 | PayRight.com | 3315 |
SoulMan.com | 3100 | United Technologies.com |
3100 |
IsItDown.com | 3000 | StyleLab.com | 2988 |
BaseCapital.com | 2967 | ShopGreen.com | 2655 |
VideoInsight.com | 2650 | HangTough.com | 2611 |
SunProperties.com | 2601 | Korean Restaurants.com |
2566 |
Business Automation.com |
2408 | ALXNet.com | 2351 |
DietRecipes.com | 2320 | SnapBox.com | 2300 |
SodFarm.com | 2200 | ExpressPizza.com | 2000 |
Malpractice Cases.com |
2000 | PennyLoafers.com | 2000 |
Among these 28 multi-word domains, you’ll see a mixture of
- Invented “brandables” – PayRight.com, SnapBox.com, VideoInsight.com
- “Veristic” phrases – WineGrower.com, CurlyHair.com, KoreanRestaurants
- Colloquial phrases long used already – BeanTown.com, HangTough.com
Some of the brandables can act as upgrades for existing brands, which tends to raise the price. In this regard, I’ve already written about several: BlueSquare, ChinaTech, IsItDown, etc.
It’s worth noting that NameJet sold more multi-word domains during January than LLL.com and numerical domains combined: 14% by count if only 6.7% by weight.
Domain | End $ | Domain | End $ |
---|---|---|---|
TOOT.com | 31,805 | REYI.com | 18,500 |
SSEE.com | 7500 | VGAS.com | 5800 |
SIHI.com | 5101 | DOAC.com | 4700 |
MEAS.com | 4200 | HSNC.com | 4195 |
MAGG.com | 3955 | BEGS.com | 3900 |
BDSD.com | 3399 | CACS.com | 3333 |
JDJW.com | 3300 | DBPW.com | 3209 |
XMBW.com | 3200 | JPHL.com | 3100 |
YWEB.com | 3060 | FAIS.com | 2950 |
WJLN.com | 2700 | QMCC.com | 2671 |
SCNY.com | 2622 | KMFK.com | 2601 |
YTQQ.com | 2600 | BDSD.com | 2500 |
JCNB.com | 2500 | MRGT.com | 2500 |
SWHT.com | 2422 | AXIE.com | 2400 |
FWMK.com | 2320 | XHDW.com | 2319 |
RRQM.com | 2301 | HKNT.com | 2290 |
ZYTP.com | 2260 | PHPZ.com | 2238 |
CPPY.com | 2210 | LYZW.com | 2201 |
JMCL.com | 2200 | JHRR.com | 2150 |
SBTF.com | 2124 | FHTQ.com | 2119 |
BQDS.com | 2100 | MJNX.com | 2100 |
TFKW.com | 2100 | YPTY.com | 2100 |
SPLK.com | 2098 | RNDP.com | 2085 |
KFMJ.com | 2050 | QBPL.com | 2010 |
SGXP.com | 2010 |
Aside from numerical domains, no category is more closely associated with the Chinese surge than LLLLs, for which the word “CHIP” was coined. During January, LLLL .COMs were the second biggest category – 24.5% by count. Yet they contributed only 11.9% of revenue.
However, the highest 2 sales don’t necessarily belong here. “Toot” is an English dictionary word, and REYI.com is Chinese Pinyin. (See above.) They’re not alone: Other words such as “begs” and “fais” can be found in the chart. Merely removing the top 2, however, reduces this category’s revenue contribution to 8.5%. In that case, their mean price is 1/3 NameJet’s mean price overall.
Here’s a summary:
Statistic | Dec 2015 | Jan 2015 |
---|---|---|
 Sold: LLLL .COMs |  142 |  49 |
 Sold: CHIPs |  107 |  34 |
 Sold: Non-CHIPs |  35 |  13 |
 Mean: CHIPs |  $3004 |  $2441 |
 Mean: Non-CHIPs |  $7668 |  $7477 |
 Median: CHIPs |  $2600 |  $2238 |
 Median: Non-CHIPs |  $3200 |  $4200 |
 % CHIP by Count |  75.4% |  69.4% |
 % CHIP by Weight |  54.5% |  46.1% |
Let me emphasize: These are not the real median and mean prices within today’s market. Remember, nothing below $2k is included in NameJet’s monthly reports. Lately we’re seeing many (perhaps most) CHIPs selling beneath $2k. Yet here we’re looking at averages of whatever rises above that threshold. By definition, these averages are always at least $2k. And they’re biased higher than reality. That upward bias will become more pronounced as the real average slips farther beneath $2k. Therefore, because of declining prices, January sees more upward bias than December did. And this effect especially distorts the median. So January’s real median price for CHIPs is quite a bit lower.
Here’s what you can see:
- The number of LLLLs / CHIPs / non-CHIPs selling above $2k has dropped by a factor of 3 in a single month.
- At the high end, non-CHIPs are not seeing any decline in prices. January’s mean resembles December’s, and its median actually increased.
- CHIP prices fell by 15-20% based even on the shrinking number of sales that do clear $2k. If sales beneath $2k were included, a more severe decline in prices would be observed.
- At the high end, non-CHIPs continue to outperform CHIPs in terms of median, mean, and max sale.
- If present trends continue, CHIPs will constitute a smaller percentage of LLLL.com sales at NameJet. Also, LLLLs themselves will represent a smaller slice of the pie.
Earlier in this article, I looked at non-CHIP 3-letter .COMs and showed 2 months of declining value. Why non-CHIPs? Because January gave us 9 sales in that area but only 2 with Chinese premium letters. No point doing linear regression on 2 data points!
Still, it can be argued with some justice that we ought to examine CHIPs, since they’re the letters China actually prefers. Fair enough! Here are 2 months of declining prices in 4-letter CHIPs. Looking at that scatter plot, you’ll notice 3 high-priced outliers. That’s because December sold a trio of triple repeaters: RCCC.com, MMMY.com, PPPJ.com. It would be unfair to make comparisons while those high sales are present in December but lacking in January. Setting them aside, the decline is less steep but far easier to see.
Even with so many sales submerged beneath that $2k surface, the upper part of the ice berg has – for 2 months now – been visibly sinking. CHIPs have been losing value at a rate of $15.52 per day ($473 per month). If you were to draw a line along the upper edge of that scatter plot, you’d see a much steeper downward trajectory – roughly $1000 per month. Likewise, if sub-$2k sales were visible, we’d see most of them during January not December; and that would tilt the regression line more sharply.
Domain | End $ | Domain | End $ |
---|---|---|---|
DJ.net | 55,500 | YWY.net | 4400 |
ZHZ.net | 4300 | ZPW.net | 4300 |
RQW.net | 4200 | RXW.net | 4200 |
PGG.net | 4100 | YZN.net | 4100 |
ZMY.net | 4009 | ZKZ.net | 3866 |
ZLH.net | 3810 | DKX.net | 3800 |
WMY.net | 3800 | YXS.net | 3800 |
ZNH.net | 3800 | ZRX.net | 3800 |
ZSY.net | 3800 | SJQ.net | 3700 |
YZK.net | 3700 | ZJK.net | 3700 |
ZQC.net | 3700 | ZXL.net | 3700 |
ZJD.net | 3660 | DQY.net | 3600 |
RYF.net | 3600 | TQZ.net | 3600 |
YZL.net | 3600 | ZNB.net | 3577 |
DQH.net | 3510 | ZJY.net | 3502 |
MLQ.net | 3500 | QQD.net | 3500 |
YXD.net | 3500 | ZBK.net | 3500 |
ZCX.net | 3500 | ZDY.net | 3500 |
ZFJ.net | 3500 | ZGH.net | 3500 |
ZHC.net | 3500 | ZHK.net | 3500 |
ZJR.net | 3500 | ZJT.net | 3500 |
ZKD.net | 3500 | ZKJ.net | 3500 |
ZHR.net | 3420 | XKR.net | 3411 |
HRX.net | 3401 | LDJ.net | 3400 |
ZBY.net | 3400 | ZKG.net | 3400 |
ZNJ.net | 3390 | ZPK.net | 3390 |
ZCT.net | 3388 | ZLM.net | 3377 |
ZLY.net | 3377 | LYR.net | 3376 |
NYQ.net | 3360 | ZBX.net | 3350 |
ZCD.net | 3350 | ZQM.net | 3350 |
ZKX.net | 3319 | KPZ.net | 3312 |
ZDQ.net | 3310 | BXK.net | 3300 |
SFQ.net | 3300 | XMF.net | 3300 |
ZHP.net | 3300 | ZQD.net | 3300 |
ZQG.net | 3300 | ZQX.net | 3250 |
ZQZ.net | 3250 | ZRB.net | 3250 |
ZRJ.net | 3250 | ZTD.net | 3250 |
CQJ.net | 3243 | ZWR.net | 3220 |
KQN.net | 3210 | ZTB.net | 3210 |
LHQ.net | 3202 | ZTJ.net | 3202 |
YZM.net | 3200 | ZCG.net | 3200 |
ZJN.net | 3200 | YZS.net | 3110 |
RQD.net | 3100 | ZSQ.net | 3100 |
RBY.net | 3010 | KQF.net | 3000 |
QWM.net | 3000 | MAK.net | 2350 |
During the record month of December, 100 .NET domains amounted to 26.1% by count. NameJet sold less overall during January, making its 90 .NET items proportionally much more important. A staggering 40.9% of domains belonged to this TLD.
Price-wise, it seems every short Chinese category has been slipping; so this 40.9% category only contributed 23.9% of revenue. That’s slightly higher than December’s 19.9% … but with double representation.
By far the highest price was $55.5k for DJ.net. Not only was it the only 2-letter item, it’s a word with commercial relevance. No wonder it came in 13 times higher than the runner up! Of the remaining 89 items, all but the very lowest-priced – MAK.net – are vowelless, “V”-less, LLL.net “CHIPs”.
Are LLL .NETs the late bloomers of the Chinese surge? For the first 9 months of 2015, NameJet sold a meager 1-5 .NETs per month, averaging just 3. Then this category suddenly cleared the $2k reporting threshold with 22 sales in October, 48 in November, 100 in December, and now 90 during the first month of 2016.
In reality, this category began its upward price climb much earlier during 2015, underneath the $2k threshold. And it is now experiencing the market downturn along with other Chinese categories. In this raw scatter plot of LLL.net CHIPs for the past 2 months, you’ll see 1 conspicuous outlier (an $11.2k sale from December). Removing it softens the slope, but you can still see we’re on the downhill.
Even so, that’s a much slower rate of decline than we’re seeing with other Chinese categories. Losing $130 per month from a median LLL.net value of $3.5k is far better than losing more than $473 per month from a median LLLL.com value of less than $2.2k.
Perhaps because LLL.net seems like a “late bloomer”, having only lately crossed the $2k and then $3k threshold, buyer optimism remains higher for this category than for others. It’s within the budgetof those who can’t afford 3-letter .COMs; and it hasn’t been observed plummeting in value as fast as 3-letter non-CHIPs or 4-letter CHIPs, the comparably priced alternative.
Whatever the reason, Chinese domain categories don’t rise and fall in unison. Some lead; some lag. For the time being, LLL.net lags. In the heat of a bear market, some will view assets that exhibit lagging behavior as a safe haven. They don’t need to be right. In the short term, that perception becomes a self-fulfilling prophecy.
John says
Joseph here’s what i don’t get about Name Jet -you send them names-not poor but quality names to list and you never hear back from them? What’s with that? One of their staff off record once told me that they only want to sell names they have an equity interest in-do you find that to be true? Thank you.
Joseph Peterson says
@John,
Your experience of never hearing back from NameJet after submitting domains for sale is probably the norm. That was my experience with them for years. Never a reply of any kind.
About a year ago, they seemed to be opening up to more sellers. And I was then able to list 1 very small batch of dispensable inventory, some of which sold. Their new manager seems to be much more communicative than his predecessor – who never deigned to reply to the likes of me. So that’s a much appreciated change.
But since testing the waters that first time, I submitted again (several months ago) and received no reply. Whether they’re flooded with submissions or we’re submitting domains they’d rather not bother with, NameJet really should implement an auto-responder to say “We’re Full” or “No Thanks”. Something. Anything.
Lack of transparency and communication breeds discontent among customers. Currently, NameJet is a platform where domainers are encouraged to buy but discouraged from selling. Yet domainers see a minority of sellers listing there regularly – often offloading domains worse what you might have submitted. Resentment and suspicion are the inevitable consequences.
Since NameJet’s rationale for choosing to work with some sellers and not others looks quite murky from the public standpoint, all sorts of theories will arise to fill the void. I know NameJet charges a commission like other venues, but that’s about it. Equity stake? Haven’t seen that.
NameJet really ought to be leading this discussion, managing relationships with their customers through messaging on site, via email, articles, etc. Instead, the company’s habit of silence leaves the door open to all sorts of speculation about motives and under-the-table deals.
NameJet is doing quite well in terms of sales. In terms of PR, not so much. They underestimate dissatisfaction. 1 reason Flippa did so well was due to companies like NameJet and Sedo dropping the ball. Sellers felt a pent-up need to list inventory at auction, yet they were blocked at every turn by the largest market places. Flippa was the major exception, a venue where they were assured a listing and could take charge of the process without depending on overtasked employees to maybe get around to telling them No.
John says
Excellent response Joseph so thank you. Ive seen many “e” names sold on Name Jet so we gave the eMobileVideo.com -imo a great name for the right company-ZERO reply. It’s just a matter of time before people understand that Name Jet is there for only one reason to offload names to the retail public-period not very different than Wall st. selling penny stocks to unsuspecting retail customers.
Joseph Peterson says
@John,
It’s NameJet’s prerogative to decline to sell your domains or mine. Just wish their policies on seller listings were more transparent and equitable, and I especially wish that they’d respond to customers more consistently. Regarding the latter issue, a few template responses would be adequate.
The comparison with penny stocks doesn’t sound right to me. NameJet handles a mixture of retail and (primarily) wholesale transactions. They cover nearly the full market spectrum from $40 to 7 figures. What they do they do well.
Krishna says
Wonderful research work from Joseph Peterson. Thank you.
thelegendaryjp says
Nice report, informative and helpful.
Meyer says
I am not taking a position about the future trend of LLLL .coms. However, just showing Dec and Jan numbers are not a true indicator. You would need to include, at least, Oct and Nov. numbers.
Also, take into consideration the Chinese philosophy of accumulating cash just before Chinese New Year.
Joseph Peterson says
@Meyer,
Suppose I throw a ball up in the air. Suppose it went up and is now coming down. Suppose I want to talk about how far and how fast it’s dropping. Well, then I must begin my measurement from the highest point, the peak. It makes no sense to go farther back and measure from a moment in time when the ball was still traveling upward. In that case, I’d be measuring from 10 meters a few seconds ago to 10 meters high right now. Then we’d assume the ball is simply floating 10 meters up in the air! We’d miss the whole reality of a ball reaching a peak height of 25 meters and then dropping.
In other words, just showing December and January is really the best indicator of the decline we’re looking at. Up or down measurements should start or end with the highest point.
Meyer says
However, we do not know if Dec. was an outlier. 50 – 60 – 100 – 70 is still an upward trend.
Joseph Peterson says
@Meyer,
An entire month with 383 transactions is unlikely to be something we can dismiss as an outlier. It’s 1/12th of the calendar year. A single auction might be a fluke, but nearly 400 auctions?
December wasn’t a meaningless aberration. It was a big part of a meaningful trend. I can see no reason to throw its data out. Right now, including December shows a decline; and that will upset many investors. But 45 days ago, if I had thrown December’s data out, they’d accuse me of suppressing the good news about peak sales!
Let’s look at your sequence:
50, 60, 100, 70
You can view that as UP to a PEAK then DOWN.
Or you can throw out that 100 as meaningless. Then it looks like a steady UP.
But in this case, you’re asking me to throw out not a single value … but 383 consecutive auctions. And with no justification. See the difference?
During the article, I mentioned a few cases of outliers that I excluded. But my rationale was this:
1. In every case where I excluded an outlier, it made the decline in value seem LESS steep.
2. I didn’t throw out data simply because it told a displeasing story. I excluded a few misleading high sales that would have forced me to make unfair apples-to-oranges comparisons. For example, AUD.com didn’t sell simply as any old LLL.com. Its drastically higher value was due to a special meaning (“Australian Dollar”). Since December’s highest LLLL.com sales were of a type not found during January (triple repeaters), I set them aside so that we could compare 1 consistent asset class month to month.
3. The tiny number of individual auctions I set aside can still be viewed in scatter plots that I included as links. You can see for yourself that they’re way off on their own apart from all other data. Visually, they seem detached from everything else going on. That wasn’t true of the 383 December sales, which fit in well with trends observable in November and January.
Another important point: December wasn’t a single blob. You can see the decline INSIDE the December data. That’s why the linear regression fits as well as it does. We’re not just looking at a decline from December’s blob to January’s blob. We’re seeing fairly steady decline throughout those 2 months, day by day, week by week.
For all these reasons, December can’t be ignored. We don’t learn what’s happening by covering our eyes every other month.
Meyer says
All I am saying, is that, you can not say something is a trend by using 2 months worth of data as 2 units of measure. Yes, your numbers say Jan. is down from Dec. But, that is not a trend.
I am not saying to throw out Dec. numbers. It just needs to be shown with Oct. and Nov numbers (4 units of measure) to give a better indicator of a trend. As I have said in the past, rarely do things go up in a straight line. .
Joseph Peterson says
@Meyer,
The notion that we’re talking about “2 units of measure” is your invention.
So it’s impossible to talk about a trend that lasts 2 months? Because the number is just 2?
Fine. Then let’s talk about a trend that lasts 62 days. Now it’s a number bigger than 2. Maybe a trend that lasts for 383 + 200 = 583 consecutive auctions > $2k. Now the number is even bigger.
Your argument can be used to prevent all human beings from talking about any trends ever. We might try talking about a trend that lasts 2 years or 2 centuries. But, after all, that’s just “2 units of measure”. Therefore we can’t know anything about trends!
Let me repeat: If you’re going to talk about a decline, you begin the measurement at the peak – not before. We’ve had roughly 2 months of decline. So that’s all there is to measure so far. In the same way, if I wanted to measure the rise in prices during the Chinese surge, then I would go as far as the peak price and no farther. In that case, we’d stop at December and ignore the decline since.
You can’t gauge the height of a mountain by measuring from midway on the up slope to midway on the down slope. 1 end must be the peak.
John says
It’s really not that difficult-chart it out via sales volume/price etc -plug in 2 standard deviations from the mid point on both extremes upper and lower and you’ll see if there’s a trend or not-2 months or 2 years it won’t matter.
Joseph Peterson says
@John,
What you describe would show price distribution, but it doesn’t include Time as a variable. So it wouldn’t be possible to see chronological trends.
John says
??? Of course it would show time. If not i’m not sure what type of chart you’d be creating.
Joseph Peterson says
From your description, it sounded like the axes were (1) Price, (2) Sales Volume. Perhaps I misunderstood.
Jothan says
Joseph this was a very impressive analysis.
Nice work, man.
RaTHeaD says
this is so true. namejet appears to be a criminal organization but as far as i know none of the top people have been arrested.
Andrew Allemann says
“this is so true. namejet appears to be a criminal organization but as far as i know none of the top people have been arrested.”
Um, OK.
RaTHeaD says
man… you don’t seem to understand drunk people at all. i now bar you from this site and banish you to only posting to domain shane. i would really do that. unfortunately…i do not have such powers.
Joseph Peterson says
@Rathead,
Why do you think so? I don’t share that view.
Acro says
Thank you Joseph for the detailed, well-researched analysis. Quite possibly the *only* data-backed analysis of the LLLL / Chip market right now, that does not reek of hype and self-righteous bias.
We all have our favorites, but Rick Schwartz once said, numbers don’t lie. So here’s the numbers, folks, he might have added. 😉
The question is, with .com trends falling fast on a day to day metrics, do other TLDs stand a chance? ccTLDs even?
My personal opinion on Chips is that I can’t rely on the market’s whichever trend to make a few more bucks on a commodity that has been passed around like a kutchie at a co-ed party. Bought low, sold high, I’m out.
thelegendaryjp says
@Acro, good question. We cannot even compare the new tld’s to dot com because obviously they have “jumped” (chips and short stuff) in terms of reg and trade rices since it has just become a trend. We would need several more months to see what they move like. It will be interesting to see if a dot com down trend means the same for .whatever but considering they are still cheaper they may have room to grow.
steve says
@Joseph – outstanding analysis. One of the top Chinese domain buyers should hire you to “buy domains” for him/her
“The question is, with .com trends falling fast on a day to day metrics, do other TLDs stand a chance? ccTLDs even?”
@Acro – legitimate question. My .com premium portfolio has been losing value. On the flip side, I can buy premium .com domains for less than what people paid for them 2 years ago and launch products with these names.
Without the high LLL & NNN sales, mostly on speculation, things don’t look so hot.
Possible reasons: Apps, over saturated market due to the GTLDs, declining economy, especially outside the USA?
Joseph Peterson says
The liquid categories associated with the Chinese surge have proven incredibly volatile.
In a stable market where prices are flat or growing at a sustainable slow rate, there’s no real money to be made by trading assets like these. Sellers mainly just lose on commission. So even while prices hold steady, for those invested in a category that has leveled off, it’s tempting to divest assets from that boring area and re-invest in the “next” Chinese subcategory – something new enough and cheap enough to attract a fresh bunch of optimistic buyers. For this reason, prices don’t seem to stay constant for very long. Either prices rise while everybody climbs up the human pyramid, or else prices drop as people climb off.
In a falling market, things are quite tricky. If you see prices dropping, how can you make money by buying within that asset class? Sellers expect to sell at today’s market price. They’ve even grown accustomed to selling above the day’s price, based on rapid appreciation during 2015. Yet that won’t work once trends have reversed. Buyers anticipating further decline will offer substantially less than today’s market price. That way, they’d at least have a chance (based on the price difference) to flip very fast. A buyer might bet on a rebound, but he’ll want a discount to cushion the risk.
Accordingly, during a bear market, sellers will find that their “liquid” assets are far less liquid than expected. At least, they’re tough to move at market price while that price is plummeting. In order to restore liquidity and sell ahead of dropping prices, they’ll have to entice buyers by deliberately under-pricing the market. And this drags down the market value even faster.
But this is a weird space. News travels slowly when 2 disjunct markets and languages are involved. Different audiences hear different stories. As a group, domainers are prone to unrealistic financial hopes. Bubbles form and burst at different times. So we’ll undoubtedly see people fanning the flames of some new trend in Chinese liquid assets no matter what is happening to market value in the previous batches. Investors jumping ship will climb aboard another, and they’ll shout to recruit a crew.