Home / Blogs

To our readers: Does your company offer DNS or DNS Security services? CircleID has an opening for an exclusive sponsor for our DNS topic. Gain unparalleled results with our deep market integration. Get in touch: [email protected]

Rationale for New gTLD Applicant Auctions to Resolve String Contentions

As Ray King emphasized in his post on private auctions, “[m]ost importantly, let’s get moving—private auction makes business sense, provides clarity and speeds the process for everyone.” In the “get moving” spirit, I wish to answer a number of questions that I am frequently asked about private auctions in general and the Applicant Auctions in particular.

Q: You know little about the domain market, why should we work with you instead of well-known industry players?

A: In contrast to others offering auction services, our expertise is indeed in the field of auctions and not in the domain industry. We view this is an advantage for two reasons.

First, we are independent. We do not have decade-long relationships with individuals and companies in the industry, and after the applicant auctions, we will again focus on other auctions, not the domain space. We value our professional reputation extremely highly, and have nothing to gain from favoring any individual applicants.

Second, we are auction experts, and especially expert in auctions for this high-stake setting. We have worked on hundreds of high-stakes auctions involving tens of billions of dollars in revenues. We have learned from this, and devoted our careers to making auctions efficient and fair. Large countries conducting auctions that are essential to their economies hire us to design and implement their auctions. Examples include telecommunication and electricity auctions in the US, the UK, Canada, and Australia.

Despite our lack of experience in the domain industry we are expert at quickly understanding the features of an industry that matter for design. We have built this expertise from decades of work on auctions in many industries and countries. We are quick studies and know the importance of discussions with market participants.

Q: Ok. I understand that independence is important. Still, isn’t having the detailed industry knowledge of a long-term broker valuable too?

A: In many settings, brokers play an important role in bringing sellers and buyers together. Their detailed industry knowledge and network enable them to facilitate matching of sellers and buyers. To create maximum value, bringing together the right buyers and sellers is one of the most important aspects of an auction if these parties don’t know each other. However, in this setting matching of applicants has already occurred and is known by all. The set of applicants competing for a particular string was established last spring. This makes the traditional brokering skill not applicable in this setting. What is needed in this setting is independence. Cramton Associates does not have any ongoing relationship with any applicant. Yes, we are talking with many applicants and this is an important part of our process, but we do this because our goal is to create value for the applicant community.

Q: Why are you proposing one auction design rather than leaving that up to the applicants?

A: Choosing an auction design is not a matter of taste, or of favoring one bidder over another. There is a whole field in economics concerned with determining what the best auction is for a particular situation. Thousands of scientific papers have been written on the subject and much has been learned from decades of study.

We have invested considerable time and money in developing a design that is efficient and fair in this setting. We have validated the design mathematically and in experiments. The main elements of the design have been tested in similar settings in the field and proven to be highly effective and robust including in auctions over the past twelve years involving assets valued at millions and even billions of dollars. Most recently we conducted an extended mock auction with actual gTLD applicants as a further test in a realistic setting. We are convinced that the design we propose benefits everyone.

The best auction in this case is one that does well for all applicants, not just the buyer. The two hallmarks of a good auction are efficiency and fairness. These are the objectives—together with transparency and simplicity—that the design team pursued in selecting the auction format.

An efficient auction ensures that the bidder who values the string the highest gets it. This maximizes the “size of the pie” that belongs to all applicants. A fair auction in this context splits the “pie” equally among the applicants. Our proposed auction does just that.

If instead the applicants decide the design on an individual basis or perhaps select from a menu of options, the selected design would likely favor one or more applicants. One bidder could strong-arm another into accepting a design that is unfair to them. In any event there would be a complex negotiation among applicants for a particular string about which design to go with. A much simpler approach is for the independent market facilitator (CA) to select the best design given the agreed objectives. In this way, the applicants get a well-designed and well-tested auction that has benefited from extensive expert knowledge, as well as the comments from all applicants.

The selected design has been shown to achieve efficiencies of over 98% with PhD students playing the roles of applicants and motivated by cash rewards for successful bidding—this means the auction mechanism realized nearly all the potential value, which occurs when the highest-valuing applicant for the particular string wins and this occurs for each of the 87 strings auctioned. When the same test was done with actual applicants with little knowledge of auction theory, little preparation or training, and less motivation, the efficiency was over 95%—nearly as high.

The use of standard rules and contracts also greatly facilitates the careful vetting of the auction rules and contracts. Poorly written contracts and auction rules can lead to law-suits and other costly delays. These pitfalls are avoided with a standard design developed and tested by auction experts and reviewed by stakeholders. Draft Bidder Instructions and Auction Agreement are available for your review and comment.

One of the most basic reasons for a standard design selected by an independent expert is that it avoids the difficult negotiation among applicants about the choice of auction. Reaching unanimous agreement in this negotiation among interested parties is problematic, especially given the conflicting goals of each applicant.

Q: Ok, but for .myTLD, suppose all three applicants agree that they’d rather do this differently. Do we need to find someone else?

A: Probably not. We are very interested in hearing about your alternative ideas and discussing them with you. We are open to customizing the auction for particular domains, provided there is a good reason to do so. Since we began exploring this problem we have learned a great deal from you, the applicants, about reasons for different structures. We have quickly adopted your good ideas. Here are a few important examples: the switch to string-by-string commitment, the introduction of multiple auctions (our current plan is three—March, June, and September—which of course is subject to ICANN sticking to its schedule).

Q: Why do fees matter if the fee is paid by the other party?

A: In any transaction, it doesn’t matter who pays the fees, as long as all parties agree beforehand. If you know you have to pay a 2% fee if you buy something, you will bid less, and if you know you will only receive 98% of the amount you are asking for, you will ask for more than you otherwise would.

Similarly, in the applicant auction, it is irrelevant whether the fee is paid on the buy-side, the sell-side, or both. All that is relevant is the size of the fee, which is why we have worked hard to keep our fees as low as possible, and dramatically less than the historic fees for auctions and brokering in the domain space. For example, suppose, as is the case in our auction, the winner pays the second-highest bid. If the buyer pays no fee, but sellers have to pay a fee out of their share, then all bidders will bid slightly more than they otherwise would to increase their chance of winning and not having to pay a fee. This means, in particular, that the second-highest bid will be slightly higher, and the winner has to pay more. In equilibrium the effective fee is split equally between buyer and seller.

Q: So what are your fees?

A: With the move to multiple auctions, we have simplified our fee structure. For the First Applicant Auction scheduled for March, our commission is 1%. The buyer pays the second-highest bid and the sellers collectively receive 99% of the buyer’s payment (less any fee of the settlement agent, which should be negligible). This low fee is intended to motivate participation early by those for whom the transaction costs are especially important. For the Second Applicant Auction scheduled for June, our commission is 2%. And for the Final Applicant Auction our fee is 4%, subject to a floor in overall fees of $1 million and a ceiling of $5 million (over all auctions combined). The shift to multiple auctions has significantly increased our costs and therefore the ceiling, which originally was $4 million is now $5 million ($0.5 million per additional auction).

Q: How much should I bid in the auction?

We cannot answer this question for you. However, the results of our testing should provide some guidance. We recommend that you or an expert you hire take a close look at the following materials:

Detailed research paper explaining the auction and bidding strategy

Research paper in presentation form

Video presentation of research paper presented at Microsoft Research, November 2012

The paper describes how to calculate the “equilibrium bid,” which is the optimal amount for you to bid to get the most from the auction. Please note that the calculation is based on idealized assumptions about each applicant’s knowledge of the range of values of the other bidders. Nonetheless, the calculations are helpful in developing an intuitive understanding of how to think about bidding and how aggressively you should bid.

Q: If I am bidding against deep pockets, should I bid more than what the domain is worth to me, so that I get more money if I am a seller?

A: This may indeed make sense, provided you also believe that the deep pocketed bidder values the domain highly. However, be aware that by doing so you risk winning the domain for more than it is worth to you, so you might end up with a loss. This risk has to be balanced with the opportunity to get a bit more money from the buyer. The optimal tradeoff is described in our research.

Q: I don’t want money. I want this domain. I will make the best use of it. I don’t want an auction!

A: We understand. But you’re not the only one who thinks like this. It’s not our place to decide how the domain will be best used. ICANN has determined that you’re not the only qualified applicant, and has specified the Last Resort Auction in the event the applicants cannot agree. In the Applicant Auction if another bidder values the domain more highly, at least you walk away with some cash to support your cause and your community.

Q: I’m a small applicant and cannot afford to hire an expert and take advantage of fancy training tools.

A: We understand the needs of both small and large applicants. We are developing a set of training tools that will enable you to practice bidding on your own time on our auction platform. These role-play simulations are the best way for non-experts to learn how the auction works. The training tools will be available at nominal cost to small applicants. Look for an announcement at @ApplicantAuc.


The Applicant Auction is a simple, fair and transparent method to efficiently resolve contention. Detailed materials on the approach are available at www.applicantauction.com. I urge applicants and other stakeholders to study these materials. We are now in the consultation phase of our process. CA welcomes questions and comments on our approach. The final plan will reflect the comments of all applicants, both large and small. Let’s get moving on this important step.

By Peter Cramton, Chairman, Cramton Associates; Professor of Economics, University of Maryland

Learn more about Applicant Auction for Top-Level Domains / on Twitter @ApplicantAuc.

Visit Page

Filed Under

CircleID Newsletter The Weekly Wrap

More and more professionals are choosing to publish critical posts on CircleID from all corners of the Internet industry. If you find it hard to keep up daily, consider subscribing to our weekly digest. We will provide you a convenient summary report once a week sent directly to your inbox. It's a quick and easy read.

I make a point of reading CircleID. There is no getting around the utility of knowing what thoughtful people are thinking and saying about our industry.

Co-designer of the TCP/IP Protocols & the Architecture of the Internet


Well, quote:" Q: I don't want money. Alexander Schubert  –  Jan 8, 2013 6:32 PM

Well, quote:

” Q: I don’t want money. I want this domain. I will make the best use of it. I don’t want an auction!

A: We understand. But you’re not the only one who thinks like this. It’s not our place to decide how the domain will be best used. ICANN has determined that you’re not the only qualified applicant, and has specified the Last Resort Auction in the event the applicants cannot agree. In the Applicant Auction if another bidder values the domain more highly, at least you walk away with some cash to support your cause and your community. “

“Hey little tree-hugger; take the cash;  plant some trees and leave the deforestation to the big guys”. “Wham bam, thank you mam”.

Rest assured that none of the CTAG members will consider your “services”. Still plenty of other “flesh” that can be sold on the meat market.

To some applicants this is NOT about “cash”. its about their identity on the Internet. I assume that is hard to understand for a cash-guy.


You are shooting the messanger Michael D. Palage  –  Jan 10, 2013 10:55 AM


I share some of your concerns regarding how some gTLD strings will be allocated according to the current applicant guidebook. In fact I predicted as much almost two years ago, see http://www.circleid.com/posts/new_gtld_auctions_and_potential_unintended_consequences

However, I am a little confused at your anger toward Peter. All he is doing is proposing a solution that is open, transparent and equitable based upon the parameters provided in the current applicant guidebook.  I have been impressed by his research and his detachment to the underlying “widget” being auctioned.

While I do believe there are important public policy considerations that will touch upon a number of new gTLDs, ask yourself this question.  If a really good actor doing good things with a gTLD falls behind in its quarterly payments to ICANN, do you think ICANN will be a tree hugger? NO it is going to draw down on the applicant’s letter of credit (originally only intended for registry failure but now available for ICANN to draw on unconditionally).

One of the reasons I have been extremely critical of ICANN over the past couple of years has been certain leadership that was focused almost exclusively on revenue. When I was on the ICANN Board (2003-2006) I saw a dangerous trend where compensation packages to key executives were in part justified by external experts according to the budget of ICANN.  Thus an inherent conflict of interest to keep growing the size of the organization.

With the hiring of Fadi, I again have hope that ICANN might be able to get back on course to serving the mission me and many others have invested substantially over the last decade.

Just my two cents


Professor Cramton is undoubtedly a rocket scientist Norman Payne  –  Jan 10, 2013 11:04 AM

Professor Cramton is undoubtedly a rocket scientist when it comes to auctions but in the case of the new gTLD program, his rocket explodes on the launch pad.  The idea of a private auction being “better” for loosing applicants may make sense in a university classroom but it makes absolutely no sense in the world of business reality. 

Losers don’t lose. 

These auctions are actually bad for winning applicants.  As an applicant, why would you ever put yourself in a situation where by winning a bid, you actually fund a competitor who will use those proceeds to compete with you at another auction or through marketing another TLD?  If you win, you expect your competitor to go away, not come back and beat you up with you own money.  Try explaining that decision to your board of directors or investors.

Where is the incentive?

There is zero incentive for the Google, Amazon and even Donuts to come to the table.  For them, this is a long term game.  Google and Amazon have billions in cash flow and see no reason to rush this.  Donuts has 150 uncontested strings.  They need what’s left of their cash pile to get those off the ground.  If one party sits out, your auction doesn’t happen.  That is unless you are crazy enough to proceed with an auction that doesn’t include all participants.  Then again, if you decide to do that, you deserve what you get - a second auction.  Never heard of someone being happy with winning the same auction twice.


Winning the auction and operating the TLD is the point of these auctions.  As long as it is done legally, why should anyone else care about fairness?  Luring in applicants who have no expectation of winning under the guise of fairness is disingenuous.  Giving smaller applicants who have spent as much time energy and effort on their applications some of the crumbs from the auction as a consolation prize is not fairness.  If you are a deeper pocketed applicant, giving your competitor cash in exchange for losing, is not fairness.  The fairness racket is all about herding weaker applicants into a pen to be dispatched with as soon as possible.

I hope Professor Cramton and his team sent Donuts a jumbo fruit basket for Christmas in thanks for bringing them to the new gTLD party.  It’s full of people who have been beaten down by ICANN delays, who are beyond impatient, and running out of cash.  When you are in this position, you make irrational decisions and participating in one of these auctions, most certainly can be called irrational.

Professor Cramton is not a rocket scientist - he is an economist Michael D. Palage  –  Jan 10, 2013 3:16 PM

Again people are wrongly trying to shoot the messenger. Please review the video from Microsoft research where Peter is very forth coming about how he was approached by Donuts.

While I have had a number of philosophical differences with the executive team at Donuts over the years, in this case I think they have really done an excellent job in identifying Peter and bringing him into the eco-system.

What I do not understand is that your distrust/dislike of Peter’s proposal totally ignores the fact that if he did not exist it would just be ICANN’s auction of last resort where the losing applicant/bidder gets nothing.  At least in his model there are buyers and sellers.

I guess the only reason I feel compelled to respond is that I initially had similar reservations about Peter’s proposal, but the more I investigated it the more sound it appeared. In fact I will likely be making recommendations to a number of my clients to further explore this solution.

Dear Michael,no worries. I have no intention Alexander Schubert  –  Jan 10, 2013 4:38 PM

Dear Michael,

no worries. I have no intention to shoot Cramton. He seems to be a jolly good fellow. I just urge community priority applicants to not engage in private auctions - regardless of who offers the service.

In particular I am opposed to his rationale of ”....at least you walk away with some cash to support your cause and your community”. The problem is: He might be a brilliant business man and excellent auction facilitator. But it seems he fails to understand that for some this is NOT about business but about IDENTITY. Assume China offers to pay off the United States ENTIRE debt plus $US 25,000 cash directly to every US citizen - if the United States would be governed out of Beijing. Hey, great deal, no?


Understand the narrative here? This is NOT about the APPLICANTS and what THEY want. This is about the people out there. Yes, out there it is full of people and some applicants do ACTUALLY care about them. Community priority applicants for example. Some of these new gTLDs will create online lands for certain communities and that land is sacred to them. They want it governed by someone who cares for them, provides safeguards, gives back to the community, engages the community, understands the community.

But then again: Community apps seem to be the minority here and other applicants have other priorities.


new g TLDs create value for all Peter Cramton  –  Jan 12, 2013 5:05 AM


You have a good point that the new gTLDs are intended to create value for all of society and not for the applicants. This is especially clear for community and geographic applications. For these ICANN conducts a separate process to determine what is best for society. Our private auction has nothing to do with that.

The private auction is focused on those domains for which the community/geographic/trademark elements are not dominant. In those cases, the market economy addresses scarcity with prices. A well-designed auction has been shown over the centuries to be an effective way of establishing prices to resolve scarcity. You may not like prices and markets but they are a reality of our global economy and are just as relevant in China as in the US.

If ICANN decides that community/geographic/trademark factors do fully resolve contention for a string. Then the contention must be addressed in another manner. In these cases the Applicant Guidebook states that the Last Resort Auction will be used to resolve the remaining contention. The private auction is an alternative to the Last Resort Auction. If you or any other applicant does not want to participate in a private auction you don’t have to. Each applicant has veto power to take the string to the Last Resort Auction. Thus, the relevant question is: In the event that string contention is not fully resolved by community, geographic, or other factors, do you prefer a private auction (or other negotiation process) or the Last Resort Auction? And if private auction, which one? That is the choice.

To me, applicants have much to gain by making use of a private auction like the Applicant Auction that has been designed by experts to achieve the objectives of efficiency, transparency, simplicity, and fairness. Then the contention can be resolved more quickly and at no additional expense to the applicants, since collectively the applicants pay no more to ICANN, and potentially receive a refund depending on the timing.

I wish you success in your community application. If you think that you have a strong community case, then you certainly should wait until the community issue is resolved by ICANN and not participate in an early private auction. The private auction is not meant to resolve contentions in this case, but rather the cases where community/geographic/trademark factors are not dominant.

Peter Cramton

Comment Title:

  Notify me of follow-up comments

We encourage you to post comments and engage in discussions that advance this post through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can report it using the link at the end of each comment. Views expressed in the comments do not represent those of CircleID. For more information on our comment policy, see Codes of Conduct.



IPv4 Markets

Sponsored byIPv4.Global


Sponsored byVerisign

Brand Protection

Sponsored byCSC

Threat Intelligence

Sponsored byWhoisXML API

Domain Names

Sponsored byVerisign