ICANN/DNSO
DNSO Mailling lists archives

[ga]


<<< Chronological Index >>>    <<< Thread Index >>>

[ga] Re-post of WLS Speculator-Specific Price Survey Results


Title: Message

My apologies folks, didn't realize the attachment exceeded the 10K limit. 

The link to the PDF document is at http://www.snapnames.com/wlsproposal.html

The text-only is below:

DOMAIN NAME WAIT LISTING SERVICE:

SPECULATOR-SPECIFIC PRICING ANALYSIS

 

January 22, 2002

 

I.             Research and Analysis to Date
 
As the technology provider whose license to VeriSign underlies the proposed domain name Wait Listing Service (WLS), SnapNames is keenly interested in understanding and responding to any concerns or questions regarding WLS.  That is why, over the past few weeks, SnapNames has posted fairly detailed factual and empirical analyses of some of the conjectural theories raised in response to the proposal.  These "white papers," which are all available at www.snapnames.com/wlsproposal.html, include:
 
·        "Domain Name Wait Listing Service - Retail Pricing Analysis" (providing specific historical data correlating retail price with system efficacy - $1 for each 1% increase in efficacy; and demonstrating why $69-$99 per name is well within the price range that registrars could appropriately charge customers for a 100%-effective WLS, given that customers already pay $69 for a service that is only 70% effective);
·        "Greatest Good vs. Benefits of the Few" (demonstrating, through empirical evidence, that the WLS target customer is a "mainstream" purchaser; 99% of SnapNames customers hold 10 or fewer SnapBack positions; 92% hold only 1-2; 54% of total orders are held by customers having 10 or fewer; and answering specific questions about so-called "defensive" WLS subscriptions, etc.); and 
·        "SnapBack Subscription Distribution" (disproving conjecture that WLS would only be appropriate for names already in the expiration cycle, by presenting a breakdown of SnapBack positions by time-until-expiration that demonstrates positions to be NOT correlated with time-until-expiration).
 
 

We next demonstrate that while some registrars have claimed the WLS will increase the average cost per domain name, they have provided no evidence to support that claim, and, indeed, the evidence is to the contrary:  registrars don't bill out deleting names at $6 plus margin; after figuring in the costs of their preferential access, speculators pay $40 to $80 per name excluding time and labor.  Accordingly, for the higher-value names for which it is intended, the WLS will not increase the cost of names at all.  The WLS will only ensure that mainstream customers have the same opportunity as these speculators to get deleting names at around the very same price - and that mainstream registrars will also have the same opportunity to sell names at such higher prices as the registrars who currently cater to speculators alone.

 
II.                           Alleged Problem:  WLS, if Implemented at a Market Price Approximating What Mainstream Consumers are Willing to Pay, Will Drive Speculators to Increase Registry Load Through "Hyper-pinging"
 
Without providing any factual responses to most of the issues covered in the above white papers, some critics of WLS nevertheless further go on to conjecture that the proposal would, if implemented, reduce the speculative purchases that many consider to be the engine of growth in the domain name industry.  In addition, these critics hypothesize that, if the WLS is priced at a market-appropriate level, it will somehow have a negative impact on Registry load as speculators decline to purchase WLS subscriptions on relatively low-value names, and therefore continue their practice of "pinging" - or, rather, "hyperpinging" -- the Registry to pick up names in the delete cycle.  No evidence has been advanced in support of this hypothesis.
 
In response to this claim, SnapNames has taken the time to examine the impact of WLS on the professional speculator, just as we did with respect to the mainstream customer  (See "Domain Name Wait Listing Service - Retail Pricing Analysis").  Based on our review of the available evidence, including extensive interviews with professional speculators, we observe that there is no evidence to support the proposition that WLS, if priced at retail levels in the $69-$99 range, will increase Registry load at all.  Furthermore, we find that, contrary to the impression that some registrars (and perhaps some professional speculators) have sought to create, the average cost of a secondary domain name obtained by capturing a deletion is in the range of $40-$80 today.  The average cost is decidedly not some variation on $6 plus.  Whether this evidence is considered conclusive, it does clearly point to the need to resolve any outstanding questions definitively during the WLS' one-year trial period.
 
III.       Thesis:  As Rational Economic Actors, Speculators Are Amenable To Paying Market Prices for WLS Orders on Valuable Domain Names, But There Is No Evidence That WLS Would Cause Them to Behave in a Way That Would Increase Registry Load.
 
SnapNames' research and historical analysis of its existing customer base (for the 70% effective SnapBack(tm) service) indicates that conjectural fears regarding an alleged negative impact of WLS on registry load are not only unsupported, but unsupportable.  Furthermore, accepting claims that WLS should be priced very low would have significant negative consequences:
 
               A.            Evidence Demonstrates Speculators' Willingness to Pay Market Price for WLS
 
Lest there be any doubt that speculative domain name holders do indeed purchase a WLS-type service (SnapBack) at market prices, approximately 17% of all SnapBack positions are held by persons holding 11 or more names:  that is, by modest-volume speculators.  Nearly another 29% are held by persons with over 100 names - large-volume speculators.[1]  In light of the fact that a full 82% of our customers hold only 1 SnapBack position, these statistics demonstrate amply that the type of service contemplated by WLS is suitable for speculators and mainstream purchasers alike.
 

From repetitive and still unsupported comments posted by some opponents to the WLS proposal, one might have an impression that speculators, as a rule of thumb, pay fairly low prices for secondary names.  Some people (perhaps the same registrars that have successfully eviscerated industry-wide margins in the initial registration business) even suggest that the range for speculative acquisition of secondary domain names is $8 to $12 -including the $6 registry fee (!) - as if the preferential access to Registry connections came free, and as if the time and inconvenience of conducting a full-time speculation business had no value.  Rather than - and forgive the word - speculate, SnapNames actually conducted interviews to find out what speculators pay today (in monetary terms alone).  Our findings cannot be reconciled with recent claims that speculators pay particularly low prices for deleting names; our research indicates that the real range is typically in the neighborhood of $40-$80 per name.  The following are typical of the responses we heard[2]:

                                                                                        


 

  • "I don't operate scripts, I just use services and average $35 - but prices are going up."
  • "We have precise readings on our costs; overall we pay $14.23 per name.  If we count NSI batch drop names only our cost of acquisition is $28.30"
  • "Before they announced wls I was winning bids at an average of $50, now about $100.  A lot of people think this is the last chance to get names, so they're paying more."
  • "We paid $9.79 average over November and December 2001"
  • "I average about $80 per name."
  • "We pay a lot more than most people because we use all of our bandwidth to focus on a few good names each day.  Our average works out to $200 per name."
  • "I wind up paying on average $120 per name."

 

Even a cursory review of prices charged for deleting domain name services supports the proposition that actual prices are far above what WLS opponents will admit.  To wit -

 

-          SnapNames sells SnapBack on a FCFS basis for $69 per name

-          NameWinner claims the average auction nets a $50 price.  Some professional speculators say this number has been rising lately.

-          Scripters pay typically $9 to $15 per name, though these tend to be very low-value names that would not be purchased under WLS and therefore are not relevant to a WLS analysis.

-          IAregistry's model is volume-dependent; $8 to $10.50 per name, plus $50 per day ($1,500 per month if you use it to get at least one name per day) for 3 dedicated connections.

-          eNom's model is $2,500 per month and each client chooses how many names to pursue each day.  Some who go for high-value names only end up paying a very high price per name, but feel it is worth it.  Each subscriber currently gets 4 (or 1/10) of the dedicated connections

-          Several registrars (e.g. DotRegistrar, Signature Domains, AWregistry) serve a single or small number of clients, or their own portfolio.  Their prices, like their access, are not disclosed to the public.

-          Several registrars rent out their 40 Pool 3 connections to other players for a monthly fee; fees vary and are not generally disclosed to the public.  One offers $1,000 per month or a 50/50 split on winning auctions.

 

Weighting the results based on estimated market share that each supplier claims, or is believed to possess based on regression analysis, the average cost of a secondary domain name obtained by capturing a deletion is in the range of $40 to $80.  While some speculators use only one or two suppliers, others use a smattering of vendors in order to get a sufficient number of names to support their business model, and they bear significant (uncounted) labor expense in managing the process. 

 

In summary, actual prices paid by speculators on names to be subject to the WLS are several times higher than what members of the articulate, persistent, special interest group of their vendor registrars really charge them.  Of course, it is in the interests of such a special interest group to claim that names are sold -- and can only be sold -- at some variation of $6 plus, because if such a claim were to prevent the WLS from being implemented, the special interest group could continue to make what is effectively far more per name under its preferential arrangements, and to do so without fear of competition from registrars that also service the mainstream.

 
               


B.            Even Assuming That Speculators Would Decline to Pay Market Prices For a WLS Position, Setting the Price Too Low Will Jeopardize Important Systemic Benefits of WLS

 

As outlined above, our research and experience lead us to be highly skeptical of any claim that speculators would not purchase a WLS at an appropriate market-based retail price.  But setting aside that unsubstantiated claim, there are still compelling reasons why introducing WLS at an artificially depressed price could have significant negative consequences that should not be risked.  These consequences include:

 

  • Ensuring that the WLS fails to benefit mainstream users, as intended.  A too-low price would give speculators a tremendous incentive to hoard WLS positions at the expense of the mainstream user who has an actual, practical application for the domain name, rather than a desire to arbitrage it to an even needier or more eager purchaser.  A WLS price that is established too low would therefore be tantamount to re-establishing the status quo by different means - which may be precisely what some parties hope to achieve through clamoring for an artificially low price.  Should the market, and market research, set prices, or should a few customers dictate them?

 

  • Ensuring that "all the good names are taken".  One common objection to the WLS is that all the "good" names will be taken.[3]  While not all good names would be taken if the retail price were market-appropriate, all the good names surely will be taken if the WLS is priced so low that there is virtually no financial disincentive to purchasing names solely for the purpose of arbitraging them to others.  Think of it this way:  if the shoe store sold its entire stock at $1.50 per pair, all the good shoes would be gone in very short order.  And "shoe speculators" would be sitting on mountains of shoes, selling them out of their garages for a price well above $1.50.  If, on the other hand, shoes were priced in the $69-$99 range, then the typical purchaser would probably look a lot like the current shoe store customer - i.e., like someone who actually needs a pair of shoes.

 

C.     There Is No Basis to Believe That an Appropriately-Price WLS Will Lead to Additional Registry Load.

 

The issue of Registry load is clear red herring.  While it is possible that WLS might have a positive effect - by decreasing hyper-pinging with respect to names that are acquired through the WLS -- there is no rational, articulable reason to link the WLS to an increase in Registry load.  If the WLS is priced at a level that consumers (whether mainstream or speculators) find unattractively high, then they would simply continue their pinging at current levels.  Too-high a price = no change on Registry load.

 

But if the price is set artificially low, then consumers would take WLS positions on more names.  Thus, it is conceivable that a too-low price could lead to even lower pinging, as more purchasers use WLS (it being cheaper than the $40-80 average cost per deleting domain name), and fewer ping.  But based on the tremendous volume of pinging even for very "low value names," and on the fact that new registrars continue to enter the field, it is unlikely that WLS will have a significant impact on Registry load in either case.  In other words, WLS will have, at best, a mildly positive impact, and at worst a totally neutral impact.  But nobody has yet demonstrated that there would be any negative impact.  Such claims remain wholly speculative, and as such, cannot form a valid basis for opposing a trial period designed to test such claims.

 


IV.       Conclusion

 

WLS is a ready-to-implement solution to a number of problems in the status quo, such as the absence of an ordered process for distributing expiring domain names, and the practical reality that most mainstream domain name seekers are "frozen out" of the secondary market by speculators who purchase preferred access to connections.  Despite a few unsupported and conjectural claims to the contrary, the evidence shows that, like all rational actors, speculators do pay the market rate for secondary domain names, for subscriptions to the WLS-like SnapBack service, and for preferential access to registrars' connections - on average, about $40-$80.  The WLS will have little impact on the price speculators pay, or the rate at which they buy; what will change will be that speculators will be joined by mainstream customers. 

 

Further, in the absence of actual evidence that speculators would respond to WLS in any way that negatively impacts the registry load, we urge that WLS be welcomed as a valuable additional service for registrars to supply to all customers.  We further urge all parties to understand that implementing WLS at too low a price would significantly hinder the service's ability to distribute secondary names to parties who will actually use them.



[1] Further detail on our customer breakdowns is furnished in the "Greatest Good vs. Benefits" white paper.  To date, SnapNames' interlocutors have declined to furnish equivalent information about their own businesses, and we renew our invitation that they do the same.

[2] In the interest of full disclosure, we note that several responses are paraphrased for brevity.

[3] Indeed, some parties have claimed simultaneously that "all the good names will be taken" AND that the price is set too high.  See George Orwell, 1984 (defining "doublethink" as the ability to hold two contradictory beliefs in one's mind simultaneously, and to accept both of them).



<<< Chronological Index >>>    <<< Thread Index >>>