ICANN|ARCHIVES

Internet Corporation for Assigned Names and Numbers

>>JOTHAN FRAKES: Good morning, my name is Jothan Frakes. Welcome to the session on the domain marketplace workshop.
This is an informational session. It is a first initiative to bring together a broad representative cross-section of ICANN shareholders to share information about changes and trends in the domain name marketplace.
The workshop's been organized in response to requests from the ICANN community for additional information on the numerous ways value is derived from domain names, and how the domain name marketplace has changed in recent years.
It's also been designed in response to questions about the appropriateness of some of these activities. And the workshop is also being organized to help ensure that policy decisions and potential contractual amendments are made with an understanding of the broader impacts that such changes can have on market activity.
The ability to create a common vocabulary and understanding of the marketplace are goals of this session.
I'm grateful for the opportunity to be the moderator for the session. I'd also like to commend ICANN for undertaking this workshop.
My challenge will be that we are attempt to go cover a great deal of ground in a compressed period of time. In a three-hour span. It's quite ambitious. We will have opportunities for public comment at times, but for time's sake, I request that the opportunity to address is compressed, that people are concise with their questions, and I also ask the participants who are panellists to please keep their presentations to five minutes so that everybody can make it through their presentation. We have a lot of ground to cover today.
I would encourage people to be concise and to forgive me in advance if I cut you off.
That said, I would like to turn over the mike to Mr. Tim Cole from ICANN, the chief registrar liaison. And he is going to describe the life cycle of a domain name and the history and evolution of the domain name marketplace.
>>TIM COLE: Thank you, Jothan.
Thank you, ladies and gentlemen.
I probably -- That title was probably more ambitious than I can really attempt to cover in the brief period of time that I want to take here.
I want to allow the greatest amount of time for our speakers and for interaction, because I think this is going to be a really interesting panel.
What I have up on the screen is the same as the handout that you have available to you.
And I encourage you to refer to this handout throughout the presentation, because the different speakers are going to be speaking about activities that take place along this continuum of the life cycle of a domain.
I am not going to go into tremendous detail reading all of these different periods off to you.
But what I would like to do is highlight a few key things.
As most of you know, a domain name can be registered anywhere from one to ten-year period, and anytime during that period, it can be renewed for additional periods of one -- up to ten years.
Upon expiration of the name, there are a variety of stages that a domain name can go through that may permit reregistration -- renewal or reregistration of the name by the original owner, by the original registrant, I should say, and/or, based on different registration agreements, there may be other options that can be pursued in terms of changing the registration during those periods.
What we're going to focus on today are three different -- three different spots along this continuum.
The one that we'll probably spend the most time on, after we talk about -- well, let me back up one moment.
The first thing we're going to be talking about today is the monetization.
We're going to have a brief session that's going to talk about how people make money off of domain names and in what ways they -- the interest in having these registrations is enhanced, so that we will get some sort of context for why these different activities take place.
The first thing we want to talk about today -- we're only going to do it briefly.
I think we may have -- it may turn out that we will want to have a more in-depth workshop in the future that will go into greater -- into greater depth into these concepts.
But we're going to talk briefly about the monetization.
Then we're going to talk about some of the activities that are generated as a result of the interest in that monetization.
The period that we're going to spend the greatest amount of time on or at least focus on today is the add-grace period, which is the first five days of a registration.
During that period, a name can be registered, but it can also be deleted and any registration fee refunded during -- from the registry to the registrar during that five-day period.
So what has evolved is a -- many people are using that period to test a name or the term has been coined to taste a domain name to determine whether or not it has monetization value.
So many thousands or millions of names may be registered or deleted during that period and only a small portion retained.
We will also be discussing the other end of the continuum, when a name is deleted.
We'll be talking about the various practices that have evolved to catch names as they drop off of the registration rolls and out of the zone files and so forth.
And people will be talking about some of the practices that have evolved to enhance the ability by registrants and registrars to capture names as they're deleted.
Finally, we will also be talking about the period between expiration and deletion, and some of the practices that have evolved with regard to registrar/registrant agreements that permit a variety of activities to take place during that period to possibly resell the name or to find out ways of either redeeming the name or getting the name into new hands during that period.
I'm not going to describe each of these in great length, although I think this is a very -- it will be very helpful to understand all of these different periods.
And I'm hoping that the different speakers will speak briefly to them as they do their presentations.
And from that point, we will move on to the first speaker, who is going to -- or I'm going to turn back over to Jothan, who will introduce the next speaker.
And I hope that this will be helpful to you.
We can go into greater detail in future workshops in terms of the actual history and monetization and so forth, but this is just a brief overview to give you some context for this upcoming workshop.
Thank you.
>>JOTHAN FRAKES: Thank you, Tim.
The next segment we are going to discuss is on monetization.
And there's different views, different perspectives.
The first speaker is a gentleman who I have deep respect for, very involved in the domain industry, is Mr. John Berryhill, and he will be talking a bit on pay-per-click and some of those trends.
John, I hand the microphone to you.
>>JOHN BERRYHILL: Okay.
Sorry about that.
Having been asked to go first, I'm going to assume for the moment that the participants, Jothan and Tim, have already thrown around some terms which may be unfamiliar.
And my goal is to simply define some of the terms and just provide a glimpse into some of the practices that we're going to be going into in greater detail on the panel.
And I want to briefly describe the market forces which drive speculative interest in domain names and then focus at the end on the issue of current popular concern, the so-called add-grace period registrations or domain tasting or various other terms.
Sometimes the practices that occur in the domain market are evolving.
They evolve faster than ICANN develops policy.
And it often turns out that by the time a policy is developed, the practices have moved on, and sometimes the terminology doesn't straighten out.
The classic model that people have of domain speculation is that I may find an interesting domain name, perhaps I may have registered the name of some, you know, common household item or a term of current interest, and then I then seek to sell this domain name at a higher value to someone else.
That classic picture of what goes on in domain speculation is largely not relevant to the market forces from which value is derived from domain names today.
The largest force occurring now is what's called domain monetization.
It turned out that in the late 1990s and several years ago, there were people who had accumulated large portfolios of domain names, thinking that, well, some of them I can resell at a higher value.
The question arose, what can I do with these domain names in the meantime?
And one of the things that people found was that if I have a domain name such as lipbalm.com, a common household product, or photography.com, these names have value simply to provide links to merchants who deal in the products denominated by those sorts of products.
So the types of pages that one will find at a monetized domain very often people say, oh, well, this domain name is not being used.
Well, actually, it is being used.
And these links that you see on a page such as lipbalm.com or photography.com earn money simply been accumulating the traffic that comes to them.
There is a point of view that says, well, these are just, you know, ugly little Web pages with a bunch of advertising links and no real content.
The other way of looking at it is that there are generic terms such as lip balm or photography which, if, for example, "lipbalm" was registered by one vendor of lip ointment who was in a geographic location remote from you, it may be someone's blog about how they love lip balm and all sorts of fascinating information about lip balm.
But if you are an Internet user who used the navigation bar, as many people do, because of the way that certain browser software operates, as a search bar, it's probably more useful to you to find links to relevant vendors and links to more than one vendor who has products that may be of interest to you.
This type of advertising, which you'll hear a lot, is commonly referred to as PPC advertising, or pay-per-click advertising.
Many people call it "paper clip" advertising.
And the way that pay-per-click advertising works is that there are fairly large companies, such as Yahoo! overture or Google, which will consolidate bids on key words by which people can seek to have placement within search results.
I know that many of you are familiar with Google on the right-hand side of the screen, you'll see a set of links apart from what are the organic search results, and these links that appear are paid, and their placement is determined by current bids.
So I used the overture bid tool here as an example.
And for a search term such as "lip balm," the current high bid is 76 cents.
That means that that advertiser will pay 76 cents for each person who visits lipbalm.com or any of thousands of other Web sites that may be advertising limb balm.
That advertiser will pay 76 cents per click to their site.
And you can see that the bids go down from there, 75 cents, 74 cents.
And nearly all search terms that people can type that relate to everyday products or services have a click value, have a maximum bid value.
One of the ones that is of particular concern and advertisers are also concerned about whether or not all these clicks are genuine, these bids can go as high as 25, 30, $50 a click for the word "mesothelioma," which, as a tribute to the legal profession, is a disease of the lung.
And the highest bidders for that term are, surprisingly enough, not doctors, but lawyers.
The common criticism levels at these sites is that they are simply content-free Web sites which earn money through advertising.
And as you can see, we have an example of yet another content-free Web site which earns money through advertising.
That is, there is no original content which is placed at these sites.
And all they do is they seek to earn advertising revenue from people who type things into a space on their computer screen and hit a search button.
And to compare the two models, what the domain registrants are doing, the so-called domainers or domain portfolio owners are doing, is simply taking advantage of the fact that the browser search bar -- or the browser address bar is used nearly as often as a search bar is used in order to locate things that are relevant to various key words.
So where we interface with ICANN-relevant issues is the various practices that domain registrants have developed in order to obtain domain names for this purpose.
And the truism of domain names particularly in dot com is that all the good ones are taken.
There are several sources where people obtain domain names at this point.
There is still an active secondary market which has become fairly predictable in that the value of a domain name is determined by some multiple of that domain's annual pay-per-click revenue.
For example, there was a portfolio of nearly 150,000 domain names which attracted 19 million unique visitors per month, which was sold at eight times annual revenue, at $164 million.
The next category are expired and abandoned domain names.
Many people register domain names either they can't determine how to make money with them or they lose interest, whatever.
There are 22,000 domain names which expire each day.
And a number of practices have grown up that are variously called drop-catching.
And these practices depend largely on registrar post-expiration procedures, which relate to the ICANN domain deletion policy.
There's also just blind registration, a term that may be of interest to you.
On the way back from the Luxembourg meeting last year, I noticed that there was this term printed on the side of all of the cigarette packs in the Netherlands, and I thought such a popular term might have value.
And apparently it is the Dutch smoking warning that is printed on the cigarette packs.
I was very disappointed to find that as a domain name, it only gets three visits per month and will not earn its registration fee.
To avoid the speculation risk in blind registration, the add-grace registration or the domain-tasting procedure takes advantage of a fact that domain name under certain circumstances can be registered for five days before a permanent annual $6 payment is due to VeriSign to register that domain name.
So what has been happening as nearly all of the dot com names which are worth $6 a year in pay-per-click revenue have been registered, the question is, how does one find, you know, domains that will earn their keep.
And the process is sort of like mining for various minerals where one takes a large amount of ore and, through various refining processes, saves that portion of the ore which is worthwhile.
There are a number of sources that people use to search through large numbers of strings.
These can be from search engines, dictionary algorithms, and so forth, that people are using to test.
Contrary to some statements that have been made, the financial impact, if -- and I built one model here -- if one registers two and a half million domains a week over a course of four weeks to test 10 million domains with a 5% keep rate, one is required to deposit with VeriSign $15 million, on which VeriSign earns the interest on the float.
The registration fees resulting from the kept domain names out of that would result to $3 million.
The 25-cent-per-domain fee to ICANN would result in $125,000.
So the notion that this is somehow a free process ignores the fact that even with a small percentage of domain names kept, a fairly large amount of revenue is derived both at the registry and at ICANN.
>>JOTHAN FRAKES: Thank you.
John.
And I'm sorry if you have more slides.
You ran a little bit past the five minutes.
I would ask the other guests that your presentation does show and frame the discussion.
So if the rest of the speakers would like to keep it to five minutes in respect to the people from the public audience who would like to comment and also from each of the other panelists.
Our next speaker is Mr. Josh Meyers, the general manager of Yahoo! search marketing many of you know as overture as a previous name.
And with that, Josh, I'll hand it to you.
And thank you.
>>JOSH MEYERS: Good morning.
Thank you, Jothan.
So I'm Josh Meyers, a senior director in the Yahoo! publisher network group and also the general manager for a product at Yahoo! called domain match, which is our product which is geared to serve this particular channel and the practice of domain name monetization.
So I thought I would just talk a little bit about Yahoo!'s role, how I think we're unique relative to other participants in the -- in the ecosystem today.
The Yahoo! search marketing group is the group within Yahoo! that's responsible for basically paid search and advertising products.
The Yahoo! publisher network group, of which I'm part, is responsible for distributing all Yahoo! products to publishers across the Internet.
So, for example, some of our partners include CNN and ESPN, as well as other well-known publishers, in addition, of course, to distributing all of our products across our own network, which is Yahoo!
The Yahoo! partner or the Yahoo! publisher network mission is to build the most trusted and valuable distribution network for advertisers, publishers, consumers, and developers.
And the way that we view domain really is just as a single channel for us to distribute our products and participate.
The products that we distribute include search, paid search, and all the Yahoo! consumer products that you can think of, such as Yahoo! shopping or, more recently launched, Yahoo! answers.
So I think one thing that is very unique about Yahoo!'s participation in the space is that, really, Yahoo! has a very significant interest in balancing the interests of all the parties that are involved.
So when we come and look at the space, we're not only looking at the value driven to domain owners, but also the value for advertisers, as John was pointing out, many of our advertisers are willing to pay fairly significant fees to us in order to generate qualified traffic in order to sell their products and services.
And then, of course, we also have a very strong interest in users.
That's the largest traffic site on the Internet and the most registered users on the Web worldwide.
Yahoo! clearly has defined quite a bit of value for users.
And we wish to continue to do that.
So I think, obviously, we have a very significant interest in all parties.
And within domain match, the product that we've built, is really designed to provide balance as well as value to publishers.
So the main features of the domain match product are to match the traffic.
So we might receive millions of different unique domain name requests.
And, of course, as you know now, the domains do not have any content on them when we receive a request from a domain.
So the goal of our product, then, is to take that domain name, which could be anything from Mr. Berryhill's lip balm or a cigarette warning in another language, and we have to try to decipher that into something meaningful that we can match into a set of content and/or paid listings.
The next step in the product is also to provide filtering.
So this is about balance.
Yahoo!'s very concerned about any kind of domains that we would not want to associate our brand with.
We do not want to be associated, nor do we support or condone any type of trademark-related squatting activity, as it's referred to.
And all of our contracts prevent and our partners or our publisher partners, from carrying any types of those domains or bringing them to Yahoo!
And one of the features of our product is to provide filtering to try to catch that anywhere we can, in addition to our contractual protections with our publisher customers.
And then, finally, the product provides reporting.
So it's very important, obviously, for our customers, our publisher customers, to have reporting.
They want to be able to see how many visits domains are receiving and what the rate of conversion of users or visitors into revenue is so that they can understand the economics of their business.
So, really, just in closing, emphasizing that, really, Yahoo!'s involvement is focused on three main parties: Consumers, advertisers, and publishers.
And we are really every bit as beholden to each one of those groups.
And that we actually share a vision, I think, with a number of our publisher partners, and we'll work with them to capitalize on this, which is to really maximize the value provided through each one of the domains that we service.
If you can imagine the lines blurring between what is a parked domain and what is actually a Web site, as John very -- as John pointed out, there are -- there's a Web site on the Web called "Google" that does very little in the way of providing content.
It provides valuable information.
And each domain has the potential to do that.
And Yahoo! has the most valuable collection of assets to help our publishers continue to add more value through domains.
>>JOTHAN FRAKES: Thank you, Josh.
Our next speaker is Mr. Roberto Gaetano from the ALAC, who we're grateful to have with us today.
Roberto.
>>ROBERTO GAETANO: Thank you.
First of all, I would like to clarify what is the scope of my talk today.
The at-large advisory committee is -- the way it's functioning in the ICANN structure, it takes care of the users, but also of the registrants, meant as registrants individual registrants that are not registrants for commercial purposes, simply because those are the two categories that are -- don't have any other constituency in the ICANN structure.
So I would like to make, in my very short speech, a distinction between those two categories.
Because they have a different -- they are affected in different ways by the domain name market, and specifically the way the recent use by the registrars of all the possibilities that are given in the contract.
First of all, I would like to say that, by and large, the user community, both registrants, but also end users, are benefitting by a certain, I would say, activity and flexibility and capability by the registrars to invent new form of business and by the competition that they are having, because, generally, that keeps the prices low.
And so whoever is going to register the domain name has benefits from this situation.
However, there are moments in history, I would say, that -- where the use of mechanism in -- with a little bit too much creativity is also creating some problems.
For instance, in the use of the grace period, I think that the affect of this on -- and I'm -- for the time being, I'm talking only about the registrants -- is creating a lot of confusion, because if you have your domain that is expired, then it's grabbed by -- and used and released, you have difficulties in knowing when a domain can be reacquired.
So it creates some -- I would say bad situation for the registrants.
This is not intrinsic in the fact that there's a creative use of the domain grace period.
But the -- is created by the fact that there is no basic rule.
We have a kind of a wild west situation right now where there's not a lot of clarity about the rules of the game.
And, for instance, I know that this will be addressed separately in -- during the day by a specific presentation.
But, for instance, we have -- we look with a certain interest in some attempts to put a little bit of order in this, like the proposed -- the so-called long-tail proposal.
But I think it would be inappropriate to talk about that now.
But looking, then, at things from the user point of view, I mean, the person who is surfing the Web without having registered domain names, I think that what -- first of all, I was struck by a consideration made by John, the distinction between a Web site with content and Web site without content.
From the user point of view, I don't think that there are Web sites with contents and Web sites without contents.
There are Web sites who provide the service and Web sites who don't provide any service.
And I think that from the user point of view, the most important thing is that we find what we want and we find that with the minimum number of hops, and we are not then routed to places where we don't want to be routed in the sense that we really, what we were looking for was a different thing.
So in this situation, the fact that we have this use of the grace period for having temporarily storing Web sites, the problem is not whether those Web sites are empty or full of contents.
The problem is whether those Web sites will provide me a service and will give me some useful information, either by routing me somewhere that is useful for me or not.
And I think that, again, in this situation, in the current situation, we have again the wild west, because I don't think that there is in the interest of whoever is getting a domain name and is hosting that only for a few days just to taste whether the name has value, maybe this registrant is more concentrated on the value of the name itself rather than the service that is provided to the user.
And again, so we are looking forward to get together and to find a solution with all stakeholders within the ICANN process so that we can kind of rationalize this situation and taking into account the good and the bad things for all the parties involved.
Thank you.
>>JOTHAN FRAKES: Thank you, Roberto.
And I'm grateful you were able to join this panel. I know that Wendy Seltzer on the public comment forum had expressed an interest in having public comment. So it's good you are here with us today. Thank you for the views you share.
Our next speaker is Ms. Sarah Deutsch, and she is the vice president, associate general counsel at Verizon communications.
Sarah, I hand you the mike.
>>SARAH DEUTSCHE: Thank you. Is my screen showing?
>>JOTHAN FRAKES: Rob hall, thank you for your help there.
>>ROB HALL: Don't thank me yet.
>>JOTHAN FRAKES: There you are.
>>ROB HALL: It's there. Can you see it.
>>SARAH DEUTSCHE: Yes. Although there is nothing long with a little healthy monetization, I am going to talk today sharing I think Roberto's and other's concerns about the dark side of monetization. And this unfortunately has nothing to do with lip balm or photography, but does have to do with the exploitation of trademark that victimizes consumers and businesses and the goodwill of the trademarks. And most of the problems are actually arising from the practices of domain name parking.
And let me say. Yes, here see for example our customer wanted to go to our Vcast service which is one of our wireless services but has found out all it needs to know about go daddy's services. And also from issues involving typo-squatting and domain name tasting.
So the practices are bad enough in their generic form but I am going to focus today on what I find very troubling and that is that some in the registrar community for various reasons are registering or taking ownership of typo-squatted variations of trademarks in their own names.
So many of you may be aware that various entities in the monetization food chain have already been sued. So this past March, Louis Vitton sued Overseed.net as marketers of landing pages or other revenue generation services for trademark infringement and cybersquatting, and Neiman Marcus and Bergdorf Goodman recently sued the register Dotster. According to the complaint, Dotster registered numerous variations of the plaintiff's marks in its own name and then listed no match in the WHOIS database for each of these infringing domain names. And unfortunately for us, one of the exhibits to that complaint showed that Dotster had actively exploited many famous brands including over 120 variations of our Verizon trademark as domain names. And we have written to Dotster and have only just heard that we will be able to get all of those domain names back.
This April, we also learned that eNOM had 144 variations of our Verizon trademark and our superpages trademark, which is the name of our online directory.
And eNOM cause called their domain name tasting service club drop. So we wrote to them about two in particular because we were in the process of selling our superpages business and we needed those domain names to provide back to the purchaser of that business.
So they wrote back to us and told us that apparently their customers had defaulted on the payment and registration fees and therefore, they had placed the names in eNOM's name.
So you can see here that we did a WHOIS search for one of the other infringing names, a typo-squatted variation of freeverizon.com, who it shows eNOM is the registrar, and then lists the registrant's name, their first name as master and their last name as account.
So it's difficult to determine if the owner here is eNOM or if it's some evil villain's nickname, you know, master cylinder, master account.
When we first complained to eNOM, they told us we had two choices. One is we could purchase our names for $160 plus a $29.95 registration fee or they would let the names be deleted. And deleted means it would be dropped back into the pool and auctioned off to the highest bidder.
And you can see here that they have a very active practice of optioning Verizon trademarks.
We then wrote back again informing them that their activity, under the clubdrop program, was a violation of our rights, and this has been since April, and we wrote back again in June after the Dotster case was filed.
So I heard on Friday that we would be able to get two of those names back, and I had a very good discussion here today with Paul Stahura, and I am very confident we will be able to figure this out and work this out to get, we know at least three other names of our trademarks that are in eNOM's name and we are not sure who owns the various others.
So I have spent most of my career at Verizon dealing with liability issues.
And although until now, I think the registrars have done an excellent job protecting themselves from liability, as you have heard today there are so many entities in this food chain who are getting a cut from some of these new monetization schemes. And I think this places everyone in that food chain at an increased risk of liability.
So registrars in the U.S., and I can only speak to U.S. law, I believe would face liability under, for traditional trademark infringement as well as the anti-cyber squatting act for their bad faith intent to profit as, quote, one who registers, traffics in or uses a confusingly similar variation of a mark. And you can see here the various elements that can be triggered to show bad faith.
The cyber squatting act also requires that the registrar must have a reasonable policy prohibiting the registration of identical or confusingly similar marks. It's hard to see how a practice of domain name tasting, parking, typo-squatting would be consistent with a requirement to have a reasonable policy. Moreover, this same law says that the registrar is not liable for damages only absent a showing of bad faith intent to profit from such registration or maintenance of the domain name.
In eNOM's case we heard they believed from their counsel they were immune under the Ford versus Great Domains case. However, the court in that case held that Great Domains was only as auctioneer who does not transfer or receive for consideration the domain names names that are sold over its Web site.
>>PAUL STAHURA: I would like to respond to that.
>>JOTHAN FRAKES: That's very reasonable, but can you hold that until Sarah has completed, and then you can respond.
>>SARAH DEUTSCHE: There will be plenty of time for fighting here this morning.
>>PAUL STAHURA: The facts are not correct, and it's not -- you know, fair.
>>SARAH DEUTSCHE: I think we should in courtesy let each continue because we can each interrupt each other with facts that are not correct.
>>JOTHAN FRAKES:That's indeed the case but Sarah, you are running up against your time limit. Do you have many more slides?
>>SARAH DEUTSCHE: No. One more slide.
So two last points. One is that I think the concept -- there's a risk here that the concept of secondary liability under trademark law might be expanded either in the case law or possibly in legislation akin to what we saw in the Grokster case, and I don't necessarily think that's great for many -- all across the spectrum to see increased secondary liability for trademarks.
Finally, I would note that ICANN has an important role to play in addressing this issue. Section 3.7.9 of the registrar accreditation agreement states that the registrar shall abide by any ICANN adopted specifications or policies prohibiting or restricting warehousing of or speculation of domain names by registrars. And although some in the domain name community it appears would love to do away with this section in the next iteration of the RAA, I would at a minimum urge that ICANN put this section in here for a reason, and anticipating the possibility of misconduct and the need to act on it.
So I think at a minimum we need a PDP to study these issues and suggest specific policies prohibiting registrars from promoting and participating in the speculation of typo-squatted trademarks in the secondary market.
>>JOTHAN FRAKES: Okay. Thank you, Sarah.
Now I personally heard a few blurring of distinctions and I think because Paul Stahura was singled out here, would I like to give him the opportunity to re-address before we continue into the next segment.
Paul, go ahead.
>>PAUL STAHURA: Two quick comments. You mentioned two types of names, your letter does, and by the way, I was only made aware of this yesterday, so I didn't enough time to research this controversy. But your letter to us mentions two types of names. The first two names we think were not paid for by the registrants. So some registrars -- every registrar has these names. Some registrars call it seized names or lost and found or whatever. But for various contractual provisions, we end up with names. Like it was used in child pornography or whatever, so we seize it. So what do we do with these names? If they are trademarks, if we approach the trademark holder, then they think we are holding their feet to the fire. If we delete them they may end up in some other registrar which makes it very difficult for the trademark holder to get these seized names. So what we do is we hold onto them and when a trademark holder approaches us and wants the name, we say okay, pay us retail. I think one of these names was in RGP. So therefore, we said pay us retail plus pay us our RGP fee which is 160 bucks.
So I think the way we settled those two is we just gave them to you and so eNOM is out some money.
Now, the second class of names in your letter was 145 names that your, I feel, overzealous attorney says that we were auctioning. Now, we looked at our database yesterday and found -- I was confused by this statement of your attorneys. So we looked in our database and we found that -- we looked at -- he said it was on June 6th. So we looked at days before June 6th and days after June 6th of what names were in auction and we couldn't find -- we found one name that was in auction. And by the way, when we say auction, that doesn't mean eNOM owns it. It means we have demand from registrants who want to register this name, more than one. The name is about to drop, so we don't know which registrant to give it to, and so we get it for one of the two registrants and then we auction it similar to other systems like Snapnames and Pool.
We couldn't find these names in any auction, but what we think your attorney did is look at the list of names that is about to be deleted, which is a list, and I think we'll have another speaker talk about the states that the name goes through and time but it's a list that VeriSign publishes every day that says these, whatever, 20,000 names are going to be deleted in five days hence. So I think your attorney looked at that list that we republish and found these names, these 145 names that you say you have a trademark or a typo related to your trademark on.
When we looked at the list we could only find 102 that were related to Verizon or superpages.
So further investigation, those names -- none of those names were registered at that time. After republished the list, our clients, which we don't know if it's Verizon or could be VeriSign looking for a misspelling of their name. One client of eNOM said they wanted one of those names on the list so we registered it for them.
So out of the 145, one name was registered, and not by eNOM.
>>JOTHAN FRAKES: You know, I understand you would like to re-address. This may have to do with litigious matters. It also is diverting us from our aim which is to talk about the marketplace.
>>PAUL STAHURA: I do agree with her last statement. I wanted to say I think we could settle this misunderstanding, in my opinion.
>>JOTHAN FRAKES: And it is very easy to make a misunderstanding. Obviously, these are sophisticated issues, which part of the participation in this is to create a common dialogue so people can understand all the various complex issues.
Sarah, I saw you had a point and then I would like us to move on to an opportunity for the public, after we get an opportunity to pose questions to each other.
Go ahead, Sarah.
>>SARAH DEUTSCHE: Okay. Thanks, just a quick point, I guess. We did this research in April. So 145 names were showed as being auctioned off then.
>>PAUL STAHURA: That's not what your letter says. Your letter says June 6th.
>>SARAH DEUTSCHE: In any event, this has been going on for months. Apparently if someone has stopped payment on their credit card, I don't think that this is necessarily charitable service that the registrar is doing by taking it in their own name. What they are saying is someone hasn't paid us, and someone has to pay, and it might as well be you, the trademark owner.
I think Paul here is acting in good faith and I am confident we will work it out.
>>JOTHAN FRAKES: That's good to hear. I love happy endings. So speaking of happy endings we have hit the end of our monetization segment. And now at this time we have about ten minutes for individual panelists to pose questions to the other panelists.
And I did want to start that off and see, are there panelists here who have questions of the other panelists, just in the discussion? Pat Kane from VeriSign, you have a question, and Paul, can you turn off your mike?
>>PAT KANE: John, a question for you. On the monetization slide, the last slide you put up, what are the assumptions you are making around the deposits that the registrars have with the registry?
>>JOHN BERRYHILL: Pat, as you know, there have been a number of different guesses which have been published about the numbers that are involved in the domain tasting. We had been hoping to receive from you a set of numbers similar to those which have been posted by Bob Parsons and others to get some sort of handle on it.
But my understanding is that VeriSign does require a substantial deposit for the -- for the float. If I'm wrong on that, that's fine. But.
We had -- we have been expecting to receive the numbers from you.
>>PAT KANE: Well, just --
>>JOHN BERRYHILL: You can give them to us, or you can guess, but don't call me wrong if --
>>PAT KANE: I wanted to know what your assumptions were. Because there is a payment surety that is required.
>>JOHN BERRYHILL: Yes.
>>PAT KANE: It does not have to be cash, and largely for the activity we are talking about, registrars have lines of credit that they give to us. So the numbers you are talking about in terms of what the five-day float actually does for VeriSign is inaccurate.
>>JOHN BERRYHILL: I'm sure it is.
>>JOTHAN FRAKES: That's a fair clarification. Are there other questions among the panel? I would like to move to questions from the floor. If there are questions from the floor, we do have microphones on the floor. We will have this opportunity and another opportunity after we discuss both add-grace and the expiring name practices.
We do have a question from the floor. I'd ask that you please say your name and who you are here with, and then proceed to your question.
>>CHUCK GOMES: I am Chuck Gomes with VeriSign, and I'm really not advocating any particular position. But I think that a correction I believe is necessary with regard to the R or RAA, Sarah, that you mentioned.
ICANN does allow for the development of a policy on warehousing names, and the language you are referring to I believe references that. But to my knowledge, there has never been any policy developed with regard to that. That's just a point that can be checked.
>> That's correct.
>>JOTHAN FRAKES: Thank you, Chuck. Our next speaker or next question from the audience.
>>MARCUS FAURE: I am Marcus Faure and I am with core council of registrars and I have a question for Pat or Sarah maybe. Would you say that the extended revenue you receive through domain monetization pages or you don't receive the revenue from the pages but from the registrations of course, would you say that those cover the cost you have because you have an extended load on the registry? So it is a business for you or is it a loss?
>>JOTHAN FRAKES: That's a good question.
>>PAT KANE: Every registration that we do receive is a $6 per term, is what we receive. And based upon the increased activity that we continue to see both in registration and in resolution, we do have to invest in the infrastructure to handle the upgrades in CPUs. We are talking about a system that was not designed to do 1.7 million registrations in a specific day.
So there is increase in the SRS and investment that has to take place. Is it a loss or is it a gain is your specific question, and I think the analysis has to be done in terms of if you would take a look at developing a registration system to handle just that volume, and that analysis has not been done.
So I can't answer your specific question.
>>MARCUS FAURE: I expected that a little bit. Are you planning to do that analysis and are you willing to present the result of that?
>>JOTHAN FRAKES: I was able to get some information, and I'll be talking about -- I'll share that as I introduce the add-grace, which are some numbers that indicate just overall like rates of these different practices, which one might be able to work out the numbers in their own head. Is that fair?
>>MARCUS FAURE: Maybe if Pat can answer the question, if you are going to do the analysis or not.
>> Pat Kane: I will make that analysis probably not publicly available.
>>JONATHAN ROBINSON: Jonathan Robinson from Net Names. I had a question for Josh Meyers about the incidence of intellectual property terms or trademark terms within -- that are actually used for these things, because it strikes me that a significant part of the value of these dropped, catched or names is when they have intellectual property. So what specific steps do you take to ensure that Yahoo!, for example, or Verizon or any other preregistered names don't occur in these names. Because you see in eNOM's case the difficult position they are with where someone has registered 100 names or not including Verizon. What do you do to ensure that those names aren't given added value and more useful in that way.
>>JOSH MEYERS: Sure, thanks for the correction. So first I would just challenge a bit the contention that a lot of the revenue is tied up in names that have associations with trademarks. I do think that they would generate significant revenue were they monetized in many cases. So I think the premise of your suggestion is correct. However, the generic key word names, the things such as lip balm or other names like that, actually do drive very, very substantial amounts of traffic. You can actually use the tool that we provide on our site which is a search suggestion tool and you can actually type in specific search -- any specific search term. So you can type in lip balm or lipbalm.com and you can see on our network how many people search just even in the search box alone, not including the address bar, for the term lipbalm.com, or any .COM phrase.
So there is a substantial legitimate business in this regard that we participate in.
As it relates to the trademarks in particular, the main provisions that we have and protections that we have in place are first are contracts, so all of our contracts with our partners absolutely prohibit any monetization of trademarks.
We also have attempted to, and it's a complex problem, but we have attempted to build into our product the ability to filter for trademarks.
One of the challenges is the huge preponderance of different trademarks in different countries as well as various permutations and misspellings thereof. So it's very, very complex. They might also be surrounded by other words or it might include more obscure sort of variations on trademarks, such as some of those that Sarah is referring to. Special services that they have or different things.
So we make every attempt, and the business would be a lot easier for us if the trademarks were not part of the game.
So we could certainly use as much help as we can get both from our publisher partners to honor their contracts and make sure they are not involved, and from trademark owners to notify us when. We do receive complaints from trademark owners, we almost unequivocally pull them down unless the claim does not have merit and let that be settled between the publisher and the trademark owner.
>> That's helpful because the essence of my question is do you do anything proactive to filter names out? And I understand your answer to be in spite of the fact there is no universal trademark database, to be yes.
>>JOSH MEYERS: We do. And we leverage the data from our search business where if we receive any type of trademark complaint in our search business, we also apply that same trademark into our database from our domain business. But I want to be clear it's a complex problem, and any sort of policy work that could be done on the front end with registrars or really anything that can attack the problem from multiple angles would be welcome from our point of view.
>>JOHN BERRYHILL: you also have to understand that the prime users of advertising systems are trademark owners. And there are trademark owners who have taken the proposition that talking about domain risk, ICANN monitor -- ICANN register 50,000 domain names, and I don't know what they are doing for me. Or I can take the same amount of money and put it into PPC advertising with my trademark and that will automatically guarantee that those sites that are variations of my trademark will link to me, because the ultimate result is getting my customers to come to me. And one can use the typo-squatters as a method of outsourcing domain registration cost and legal cost if what you are interested in is having your customers associate your product with their Web site. And there are trademark owners who have made that decision.
>>JOTHAN FRAKES: We have a redirect from Sarah, and then one more question from the audience. And then we'll need to move to our next segment for time constraints.
>>SARAH DEUTSCHE: I would just note that kind of decision should be left in the hands of the trademark owner. We would prefer that when our customers type any variation of Verizon they come to us to begin with. We don't want them to go to a parked site for go daddy, we don't want them to go to a site for competitors. Maybe they will ultimately find their way back to the real Verizon site but we will end up paying twice, once to have the sponsored link to begin with and once when they click on the wrong place to get to the right place. So that's an individual decision.
>>JOTHAN FRAKES: I have a comment by Bruce Tonkin, and then -- apologies for making you wait there.
>>BRUCE TONKIN: Just a general comment speaking on behalf of Melbourne I.T., but like Yahoo! we also try and build filters at various points in our process to filter out trademarks. It's an expensive process because there's not really an authoritative database to get those sorts of things and then when you look at misspellings, we do try, if it's a relatively long name, we might do a key word search on it so if it's something like Verizon. But if you have something like apple, try to use that as a key word search. Apple is a very generic word to start with.
So I want to emphasize, it's not trivial building filters, and they are not always exact. And I'm sure Yahoo! could attest to that.
But certainly I think, just speaking on behalf of our own registrar, but I know in most others, we do respond to complaints. So if we haven't picked it up in the filter, if they can properly identify what the issue is, we'll look at it.
>>JOTHAN FRAKES: If there is two things I could comment on this personally from my own experience, registrars seem very agile in hoping to solve these trademark issues, and one place that would add a lot of horsepower to trademark interests/benefit might be a central repository of trademark terms that registrars or other search engines could benefit from in order to responsibly address these things in a proactive manner.
We have one more question from the floor.
>>SABINE DOLDERER: Yes, my name is Sabine Dolderer. I am working for the dot de registry, that's the registry of Germany, and we see a lot of interesting business models and we obviously see also within the dot DE registration that there are business models not only in the .com space. And we see that a lot of these possibilities which are their business possibilities are there of our using special features which are provided for free by the registries, which is on the one hand side, I think the five-day add-grace period and which in the case of (inaudible) we don't have these type of add-grace period. It's our whois server. And the cost actually from the registry perspective, a lot of money to provide these service to those who are having this business. And it's not only money from the registries. It's actually money from the existing domain name holders and the questions what have they from providing all these services to these type of business models.
And my question actually to the panel, why do we need these five-day add-grace period? I understand that quite well if you buy a washing machine that you want to have it at home and test it and give it back if it's not working. But why do you need or want to need that for a domain? And is it really important to have this five-day grace period in the gTLD space or is it much better to abolish that and help focus therefore much more on the real needs?
>>JOTHAN FRAKES: Sabine, I appreciate the question because
>>JOTHAN FRAKES: Sabine, I appreciate the question because it's a perfect segue into the next topic.
And I'm certain you see this as the second largest domain registry in the world, that you see a lot of this.
Does anybody on the panel want to field that question really quickly, the comment on why do we need an add-grace period?
John Berryhill hasn't had a response.
Rob has hasn't had an opportunity to speak.
>>ROBERT HALL: Perhaps it would be more appropriate, I'll be happy to answer it paragraphs after we've gone through the topic with them.
>>JOTHAN FRAKES:Sabine, some of that will hit the light of day as we go through the next topics.
Is it a quick question, Nigel?
>>NIGEL ROBERTS: It's very quick.
It follows on from what Sabine says.
My name is Nigel Roberts, I'm the registry manner for .GG and .JE in the channel islands.
The five-day grace period clearly has a purpose.
But it's very clear that domain name tasting is an abuse of that purpose.
Does the panel think that domain name tasting is a theft?
>>JOTHAN FRAKES: I have one hand showing, and then we'll use these questions as an opportunity to segue into the next topic so that we can maintain our agility and meet this aggressive timetable.
Tim Ruiz had his hand up and looks like John Berryhill also has a response.
>>TIM RUIZ: I'll say yes that we do believe it's an abuse of the add-grace period.
And that's what I'll be addressing, explaining why we believe that in my segment coming up.
>>JOHN BERRYHILL: I'd say the elephant in the room is that, you know, VeriSign is not going to say one way or the other.
Apparently, VeriSign is unable to determine whether or not it's making money.
And probably it will become defined abusive by them when they cease to.
The problem is that it is a contractual provision, and, you know, when you have technical policy dictated by people like me, by lawyers, you end up with legal doctrines such as the interpretation of a contractual provision that allows you to do an activity which is in the middle of a contract that then says, "These are all the terms of this contract," is that the purpose, if any, of the add-grace period is lost, and it becomes an irrelevant question as to whether or not someone is using the add-grace period in accordance with its intended purpose, because when one signs a contract, all one is left with is the words on the page.
And, you know, hopefully, the -- this is the beginning of the discussion, at least on this particular provision of whether there are impacts that would require a revision to the contract.
>>JOTHAN FRAKES: Sarah had a comment and then Tim had another comment.
And then I'd like to transition into the add-grace to maintain our agility.
>>SARAH DEUTSCH: I would just say that I think the practice, the purpose of the add-grace was very different.
It was intended to correct mistakes that you might have made or -- during the five days.
It's now been exploited into something completely different.
And unless and until we can figure out a way to extract trademarks from the equation here -- and I think that they're inextricably linked, this is an unpalatable situation.
>>JOTHAN FRAKES: Tim Ruiz and then I think --
>>TIM RUIZ: Just want to clarify something John touched on about it being a contractual arrangement.
In reality, the current COM agreement, the language surrounding the add-grace period, which is what we're going to be talking about, where it talks about that a domain may be deleted during that period, it says "and a credit may be issued to a registrar."
It doesn't say a credit must be issued, it doesn't say a credit shall be issued.
It says a credit may be issued.
That may be arguable, some say it's ambiguous.
But it's very clear to me that the registry has an option as to whether or not they provide that credit.
>>JOTHAN FRAKES: Roberto and then John.
>>ROBERTO GAETANO: Yeah.
I would like to be very pragmatic here.
There was a policy, the intent of the policy was a certain purpose.
After the policy came the contract.
I agree with John.
Now the contracts are making the rule, and we have to stick to what the contracts are saying.
However, the policy-making process didn't disappear.
So there's nothing that prevents us from revising the policy and coming with a different policy, and then at that point in time, we'll have different contracts.
>>JOTHAN FRAKES: Jon Nevett.
>>JON NEVETT: Thanks.
Just follow up on Tim's point.
It's important to note that the new dot net agreement, and the proposed dot com agreement, change that provision a little, Tim.
And now provides that the registry operator shall have the right to charge registrars a fee for disproportionate deletes during the add-grace period.
So there's clear language certainly in dot net and in proposed dot com that the registry, if they were so inclined, could take action on this.
>>JOTHAN FRAKES: Okay.
So, you know, part of this is to educate the audience.
And we eventually have a lot of people who are very sophisticated in the art of domain names here at an ICANN conference.
But what seems to be being entertained is, you know, nuances about what is called add-grace.
And I'll use this as a segue into the next segment, which is about the add-grace period.
There's an add-grace period, it's a period -- if you'll refer to -- if you have a copy of the handout that Tim Cole had mentioned -- is a period following the -- right at the beginning of the domain creation process where there are five days in which a registrar, after submitting a domain addition, can delete the name with no financial consequence.
And it sounds like whether or not there is a financial consequence is in discussion here based on what Jon Nevett mentioned about the contractual amendments.
And what -- what I'd like to do is take -- I was -- I'm grateful that I was able to receive some numbers on what has been seen in the com/net registry from VeriSign.
I'd like to read those.
And now that I've introduced what add-grace is, we'll have different people talking about that.
What -- The numbers that I see, it's taken from a sampling from May 1st to May 31st, 2006.
During that time, there were 616 different registrars that added one or more domain names into the com/net SRS system, which is the automated system that people register domains through.
Of which 502 took advantage of the add-grace period, which means they deleted a domain name during the add-grace period.
That's 81%.
322 registrars had a delete rate greater than 5%, meaning that more than 5% of the names added were deleted.
87 registrars registered the same domain name more than once, meaning that you add a domain name, you delete a domain name, and you subsequently add it again.
This accounts for an overall of 5.2 million gross adds.
30 million unique names were added to the SRS in May in com/net.
Of those, 38.6 thousand names were registered six times or greater.
Now, if that's correct, I think I've heard the term -- I don't want to throw inflammatory terms out, but I've heard the name domain kiting used.
And that would seem more like domain kiting, where somebody's registering, deleting, and re-registering the name multiple times.
Sounds like somebody's doing that more than six times or getting a name essentially for free.
The add-grace period for registrars sending a large number of registrations at one time, 55 registrars registered names at a rate above a specific threshold through the autonomous system.
17 of these registrars had a delete rate of those names less than 1.3%.
29 of the registrars had a delete rate greater than 97%.
18 of those registrars account for 98.1% of all the measured behavior.
And I believe these numbers to be accurate.
So add-grace is an interesting topic.
There's a small number of registrars, apparently, who are participating in this.
There's -- I'd now like to pass the microphone to Tim Ruiz from go daddy, who is the vice president of corporate development and policy planning for the Go Daddy group.
>>TIM RUIZ: Thank you, Jonathan.
As I stated earlier, godaddy does feel that using the add-grace period for domain-tasting is a problem.
That is an abuse.
There's been a lot of debate about the economics of that.
That's not our concern.
It's not our concern that somebody's getting something for free.
I think there are some legitimate questions that need to be asked in that regard.
But what our primary concern is, that it's creating customer confusion, that it's threatening consumer confidence in our industry, and that it's costing -- it's causing increased support costs for registrars, definitely for Go Daddy.
We've talked to other registrars just informally.
We'll let them speak for themselves.
But indications are that other registrars are beginning to see similar problems.
So the kinds of support issues that this is creating and the kind of confusion that this activity is creating I'll let registrars -- our customers speak for themselves by just sharing with you a few of the written complaints that we've received from customers.
This first one was directed to our CEO, Bob parsons.
And the customer says, hi, Bob, I've been doing business with you for many years.
On Friday, I checked the availability of something.com from your site as I always do.
It was available.
I had my meeting with the store owner and promised to buy the domain name for him today.
I ran the availability of it again this morning from your site and imagine my surprise when the domain showed as being unavailable and sold to some obscure registrar, and he names the registrar.
I do not believe this is coincidence.
I believe in my heart that you've got someone in your camp that's giving information to someone that's speculating on domain names.
I really hope that you -- you're the kind of person I've always assumed are you and this situation is one you take seriously.
Well, we did take it seriously.
We spent a lot of time investigating this situation, in fact, close to five days working with this customer.
Ultimately, we were able to get the domain name that they were afternoon because it was dropped after it had been tasted for five days.
Another comment, complaint, actually.
I've personally searched on several names which have absolutely no meaning whatsoever.
We assume no anybody else but himself.
All were stolen by some registrar and parsed out to their various subcompanies.
They eventually dropped within a few days.
Either someone internal to your company and/or someone who is in charge of the entire WHOIS system of queries and/or somewhere along the way some bot or worm is intercepting these queries.
Somehow, in some way, these queries are being intercepted and used to some registrars' benefit.
They know it, you know it, we all know it, ICANN knows it.
It's time to put the legal folks with the U.S. government on this.
This has gone too long and too far.
No resolution to the domain names that he was concerned about.
They all dropped, so I would assume that he was able to eventually acquire them.
And, finally, I used to be a proud go-daddy customer.
I registered and bought both a domain name and a hosting space there.
Since I searched for my second domain acquisition in the WHOIS box at Godaddy's site, even after entering my customer I.D. and password five days ago, and tried to make it effective on Tuesday, it came that the dot com I queried and found available the previous Saturday was already registered, delegated, and active since the day after by, and it names some registrar.
I was registering my boss's company name.
He was trying to understand what all that was about.
He doesn't speak English, doesn't trade with credit cards, so had I to accomplish the task.
And he couldn't believe his trademark was registered.
Obvious questions: What, why, how?
What can we do?
Fortunately, in this case, we were able to help the customer get the domain name.
So these are the kinds of things that we're seeing on the rise.
We're not seeing, you know, hundreds of these on a daily basis.
We're seeing probably tens on a weekly basis, but much more than we've ever seen before and related to this activity in the add-grace period.
As this activity continues to grow, we suspect this type of issue is going to grow, this type of confusion is going to grow and our long-term concern is it's going to erode confidence in our industry and what we do for a business.
Those are our concerns.
We feel, you know, there's been some discussion about policy development processes to solve this.
For us, that would be an unfortunate step to have to take, not because we don't have confidence in the policy development process, but, quite honestly, it's resource-intensive, it's time-consuming, and we feel that the tools already exist to put an end to this activity in the add-grace period.
And, again, this isn't about a problem we have with monetization.
We don't feel if we stop this activity in the add-grace period it's going to solve all the problems we're talking about here today.
But it will solve, we believe, the problems that we're seeing right now with customers being confused and losing confidence in what we're doing.
>>JOTHAN FRAKES: Thank you, Tim.
And thank you for taking exactly five minutes.
I'm very impressed.
Our next speaker is Rob Hall, who is the CEO of Momentous.
Rob.
.
>>ROBERT HALL: Thank you, John.
My name is Rob Hall, and I'm the CEO of Momentous.
For those of you who may not have heard of Momentous, we own the largest registrar group in the world, also own pool.com as one of our subsidiaries, which is the largest secondary market entity in the world.
I want to talk today about balancing community interest and innovation.
So I want to start by actually doing something some of you may find odd for me, which is to defend the ICANN staff, because recently, on the blogs, they've been secured with having come up with this add-grace period as one of their inventions.
And that's simply not true.
We have the luxury of being been a registrar since before ICANN existed.
And in the old days when network solutions was the sole registry and many of us were resellers of them, this five-day period existed for registrants and resellers, and now registrars.
So this isn't something that the ICANN staff made up or in any way was inappropriately created.
This grace period's been in existence for almost ten years.
And was simply transferred into the ICANN contracts when they came up.
There's also a lot of fact and fiction.
And I want to try and clarify.
And I fill me role here is an education of what we're seeing in the marketplace.
This is not about someone getting a free domain.
The term taste-testing is typically about people trying to find domains worth more than the registry costs.
So there's a group of domains out there that people feel are worth more than $6 a year being monetized through the various monetization of pay-per-click services and they're trying to find them.
They're very sophisticated.
It may have amazes me at how sophisticated these people are.
They're buying less from tool bar providers, getting your mistypes, figuring out what these names are.
That's one of the primary methods they use.
There's very many different business models these registrants are using.
The add-grace period serves many purposes.
It was created for various purposes.
We've seen innovation use it for other purposes.
And we may continue in the future to see it used for -- at unpredicted purposes.
And that leads me to business models.
Various registrars have different business models.
We should be different.
Tim at Go Daddy should be different than us.
That's the entire purpose of competition and innovation, is trying to get us different so that we can serve different communities and different market segments and offer various products at different prices, even, to the community.
So through competition, it leads to innovation.
I look at perhaps the various registrars models now, and you've got your traditional registrar, network solutions and Go Daddy, although -- and Tucows, although they are even different, discount, full-price, and a reseller model.
You've got the acquisition registrars, which we saw over the last couple years, used by companies like pool and snap names and eNOM.
I'm always amazed when people innovate.
And I take my hat off to Google, who went through the accreditation process to become a registrar and has never registered a name, but instead found value in the information that only an accredited registrar can get, so they derive value from being a registrar in a totally different way and a new and innovative way, and I take my hat off to them to figure that out and to use that to improve their products which are not domain names but search.
Let's talk about the size of the market.
I think there's a lot of scary numbers being thrown around.
We've been able to ascertain, and I'm glad the VeriSign statistics seem to match exactly.
There's approximately 18 registrants involved in this practice en masse.
This is not thousands and tens of thousands of people doing this.
They are doing huge numbers, but we're talking about roughly 18 different registrants or unique groups of people doing this.
My personal view is this will self-limit itself pretty quickly.
So you're seeing last month I think it was Bob parsons had speculated that there were 200,000 new registrations in May due to the add-grace activity.
We ought to be asking ourselves, is that a good result?
We have new people registering new names.
ICANN was created to foster competition.
We've seen the zone file increase rapidly since the beginning of competition.
And along comes a new product that is seeing more and more names registered and more and more people start to participate in the ICANN process.
It's unfortunate that I think we also confuse a lot of the different things that are happening, and one, and that's something very serious we need to be careful of.
I've seen even on this panel today confusion with the expiry market and deletion market, which are not related to the add-grace period up-front in the first five days.
So what we're typically seeing is people take large lists of names, go and attempt to register them for five days, this is not the reregistration necessarily of deleted names nor the expiry market.
And I'll leave it to my compatriot John to speak to the expiry market later.
I should also point out, none of our registrars or pool.com participate in this taste-testing or allow a registrant to give us a list of names that we will delete four days later.
What I'm trying to tell you is what we're seeing from our clients and what's happening in the market.
The caution I have is too often ICANN seems willing to guess at a sense of the community.
And this typically leads to controversy.
After a few discussions I've had here amongst my panelists, I'm hopeful that this will not happen here.
But I want to touch basically on some of the problematic solutions that we've seen.
So I want you to keep in mind that everybody has individual motivations.
So the registries certainly are starting to look at this activity and say they want their cut.
And, again, I'm going to do something odd for me, praise VeriSign for a second.
Back at the Tunis meeting, Chuck came to the registrar meeting and said, you know, we've seen this odd activity.
We saw a million names registered, and only 800,000 of them were kept.
And after we explained this new phenomenon that was happening, and was it really a problem, because 200,000 names in $1.2 million was kept by VeriSign, it didn't become as much of a problem all of a sudden.
As the number of names increase that were tested, we're still seeing about the same number, about 200,000 month after month kept from this activity.
It's important to keep in mind these are names that would never be registered other than that.
So some call it a technical issue.
Some call it a policy issue.
I'm amazed at the people that point fingers at irrelevant examples.
I think the classic one was the very serious issue of a rape crisis center name being voluntarily deleted by the registrant.
It was then reregistered months after it had deleted and expired even by someone who put up a copycat site of it and a bit of a sick joke site, unfortunately, and somehow this was used as an example of why taste-testing or the add-grace period was being abused when it has nothing to do with the add-grace. It's simply an inflammatory example of a different behavior that we need to educate and probably -- and perhaps solve a different way.
And I just want to make sure that we're clear in talking about the specifics of what's happening in each different period.
Because lumping them all together and causing fear, uncertainty, and doubt, will not help us get to a solution.
Clearly, there's market demand for this.
Clearly, there's a market out there that's been created that people want to find these names.
I've heard various schemes such as price fixing the wholesale price, but I don't believe it's the answer.
I think we need to carefully look at this.
And I'm concerned about applying a knee-jerk solution to an (inaudible) activity already labeled as a problem by some.
So the grace period was thoughtfully implemented and should be thoughtfully assessed.
Quick solutions are more likely to cause problems or -- it's interesting to me that people often label something that they didn't predict when they made a rule, or behavior that they didn't forecast as being a problem immediately because they don't like the change that they didn't expect.
I've seen rampant speculation, guesswork, and scare tactics that I don't feel are appropriate or conducive to finding a solution.
I've heard words today, abuse, exploitation, theft, a legal word, kiting, you know, to try and focus -- Tim did not use it today, for the record, by the way.
But I've heard these words used.
Kiting's a criminal act.
I mean, what they're doing is, I feel, inflaming something when what we need to do is educate people as to what it really is.
And along with encouraging innovation, the community needs to make decisions and find solutions based on facts.
And I think that's what I'm so happy to see this panel doing, is hopefully presenting and educating the community.
Because only after we understand what it is, what the market demand is, and what's happening, can we work together to change the behavior if we deem that appropriate.
But it has to be an entire community.
It can't be just the registrars or just the registries or just the domainers or the intellectual property constituency all saying different views on what this is.
We need to come together as a community and say, "Is this a problem?
And if so, how should we address it?"
So, in closing, I think the best thing we can do is continue to educate each other of what this is.
I'm happy to.
My -- I've put my E-mail address up here.
I'll answer questions today, and I'm happy to answer them after today.
But the most important thing to me is that if we all understand what this is and what's happening and what the motivations of the people doing it are, then we can start to address it and find solutions and move forward together as a community.
I think that's very important is to not get ourselves into a situation where we've developed a new policy or a new process that actually has the reverse effect to what one may want.
Thank you.
>>JOTHAN FRAKES: Thank you, Rob.
And I appreciate that you're making distinctions and clarity on these various matters, it's very important.
Thank you.
Our next speaker is David Maher, who's the senior vice president of PIR, the public interest registry, the registry for dot org.
David.
>>DAVID MAHER: Thank you.
At the outset, I would like to make it clear that I am unalterably opposed to fear, uncertainty, and doubt, and to knee-jerk solutions.
Having said that, I would like to talk about the experience of one registry, dot org, public-interest registry is the manager.
We have found in the past few years a dramatic increase in the use of the add-grace period.
We see some, but not all, registrars registering literally millions of names in a month or even a week with a deletion rate of up to 99%.
One obvious question is, well, why don't you just take the money that you get from the remaining 1% and be satisfied with that.
Well, there are two reasons why we think there is an issue here.
First is the technical burden on the system.
At the present, we can handle the traffic that we're getting.
Afilias, who is our back-end service provider, is doing a great job in meeting the demand that we're currently experiencing.
But if this practice continues, and if it spreads, which seems to be happening, because it is increasing, we believe there will be some technical impact that will put a real burden on the system, with resulting costs.
But that's speculation at this point.
Second, dot org serves primarily the noncommercial community.
We're an open registry.
We always have been.
But we're identified, really, with noncommercial users.
Many of these users are relatively unsophisticated, who, for one reason or another, allow their names to be deleted, or to lapse.
And what we see happening, and I emphasize that this is a growing trend, is that some names that are popular or seem to test well, are then picked up and used for antisocial purposes.
And Rob referred to one instance which may not be related to this particular problem, but what we're seeing is a trend of it happening.
And we do believe, even though Rob may not agree, that it is related to the increasing use of the add-grace period with 99% deletes.
We are doing two things about this.
The first, and I think the most important, and here's the one area where Rob and I certainly agree, and that is an educational program.
We think -- and we're launching it very soon.
We think that the users of dot org and the noncommercial community need more information about what happens if you allow a name to lapse or if you -- whether advertently or inadvertently.
People should be aware that names are being picked up and misused.
Second is the -- a proposal that we brought to ICANN which it turns out is reflected in the dot net agreement.
And that's the opportunity to charge for, quote, excessive deletes, unquote.
And I think that that -- if and as implemented -- could go a long way to cutting down this trend to the rapidly expanding number of deletes compared to registrations through the use of the add-grace period.
Thank you.
>>JOTHAN FRAKES: Thank you, David.
We have a question period set up for this.
But I do see Steve Crocker at the microphone who would like to comment.
>>STEVE CROCKER: Thank you very much.
I chair ICANN's Security and Stability Advisory Committee, and we've watched this add-grace period phenomena for a while.
I also was prompted to come up in response to Rob's comments about the particular incident involving the Syracuse rape crisis center, since that was, indeed, features in -- as one of the incidents in the report that we presented Sunday afternoon.
And, first of all, I completely agree with you that there was an issue of the -- primarily of the unintended consequences because the registrant, the rape crisis center, did not anticipate that that name might get reused in a way that would -- that they would find completely uncomfortable later.
And it was -- and I think the point that you made about getting the distinctions correct is a very important one.
I want to agree with you and emphasize it.
So in the report that we presented and we're about to release, the formal version in a day or so, when we finish the process, it was entirely focused on education toward registrants as opposed to suggesting that there was misbehavior by a registrar or the registry.
There's a much larger policy question as to whether names should be held out of service for a long time, as the way phone numbers are, for example. But that's -- that treads into a much more complicated territory, and, as I say, the reason why we focused our advice precisely at registrants and to increase awareness and education.
So I think we're all on the same page with respect to that.
And I appreciate you pointing that out.
The other aspect which David has spoken to, I think pretty strongly, and which other speakers have spoken on, is that as we've watched this use of the add-grace period, although there is revenue that comes from the residue of the ones that are retained, whether that's 5% or 1% or a tenth of a percent, there is, indeed, some revenue that flows to the registry.
However, the other quite obvious very large aspect of this is that there is a non-trivial amount of resources being consumed on the registry side for the ones that are not kept.
And that's unreimbursed.
Now, why should we care, say, speaking from the perspective of the Security and Stability Advisory Committee?
I don't care from a business perspective.
We heard early on during this process some complaints and yelps of pain from the registries that there was burdening their systems beyond their capacity.
So we listened for a while and we said if that's really true, then that does suggest that there's a potential stability issue here, and it's kind of odd, because it's really a business matter between registrars and registries.
Why don't you just rebalance that equation.
And more YELPS of pain, we can't do it because the contracts are in the way and because this is part of ICANN bureaucracy.
We said maybe we should take a look at that.
When we actually focused our attention, the registries said, oh, no, don't help us, please don't help us.
We want to continue, they didn't say they wanted to continue.
They just continued to complain, but said it's not really a problem, we don't want to piss anybody off, so we'll just continue with this.
I really don't know what the actual facts are.
But, you know, as we all understand, the add-grace period was put in as a consumer protection to reverse the very small number of mistakes.
5% is probably a more than generous figure for what the actual legitimate mistakes are.
They're probably down in the 1% range.
And then to scale that up so that it is a hundred-fold bigger, so that it is the dominant mode of registration, you know, how much overhead, latitude is put into -- there's a technical term -- headroom is put into the design of these systems to accommodate the excess load of registrations has got to have been exceeded many, many, many fold.
So the registries, VeriSign and PI are and Afilias, have put -- and those are the ones primarily affected, I believe -- have put quite a lot of investment into that.
But it is in some sense unreimbursed except for that small residue for those additional investments.
Whether or not that's acceptable or not is really for the registries to say.
But I think that it's clearly part of the larger picture that whatever additional revenue is going on, there is a very disproportionate additional investment that has had to be made, and operating costs that go with that investment.
And that's just part of the facts of the matter.
Whether that's viewed as proprietary and nobody wants to put those figures on the table or not is, you know, a business matter.
But clearly from an engineering perspective, you just can't scale up any portion of a system by 100 fold and then say, well, it's just more of the same and that's the contractual provision.
Thank you.
>>JOTHAN FRAKES: Thank you.
I saw a sea of hands up here on the panel to respond to Steve.
I'm just going to go from right to left across, starting with who started over here, who was it, Pat?
Okay.
>>PAT KANE: Steve, I would just contend that from the -- the discussion or what you said about we didn't ask for -- we said don't worry about it, go away, we don't want your help, it was more of a situation of presenting it and we had to solve our own problem, which we did.
And then when we were asked if we wanted any help, we said, no, the problem had already been solved.
So thank you.
>>JOTHAN FRAKES: David, I see your mike on.
Do you have something to say?
>>DAVID MAHER: Oh, no.
>>ROBERT HALL: Steve, I thank you.
I want to comment on two things.
First, I think there's a difference in how registries are handling this.
As I said, I complimented VeriSign on -- you know, they already were maybe fortunate enough to have a technical system in place, because they had to deal with the large volumes and the drop-catching, they may have been able to handle this better than some other registries.
And as the attention turns from .com in the secondary market to other TLDs, such as dot net and dot org and even the country codes, those registries are going to have to come to grips with the fact that they may need to scale their technical systems up to handle the demands of their clients.
The other thing you commented on, and I do want to address Sabine's question, too, but the other thing you commented on was a possible five percent threshold or something similar like that, and we have to be very careful here, and again this is one of the reasons I think it's best for the community to come together. One of the drawbacks of doing something like that as opposed to just having a hard number would be that it may allow is someone like go daddy, I'll pick on my friend Tim because he is sitting beside me, who has such a high level of volume of typical adds through their business to offer a service to domainers that a small registrar couldn't because they have such a high proportion of other business.
So we want to be careful that whatever solution is presented doesn't somehow create inequality or inequity amongst registrars based on their size. And I'm sure we could find that if that's what the community wants to do but we have to be careful not to all of a sudden create different classes of registrar because right now we are all fairly equal and treated equally by the registries.
To Sabine's comments about .DE, yes, I think the country codes are going to start to see this. As I said, I think the dot com market will play itself out very quickly, there are only so many names there, and once they are owned by someone, you may get people flailing away but they will quickly get discouraged that there aren't any names worth over what the registry cost is.
It's interesting that dot de sells names, as I understand it; on a monthly basis, so you have probably already been seeing this activity without realizing it because you have a lower monthly cost, so....
>> (inaudible).
>>ROB HALL: You are already seeing it?
>> (inaudible).
>>ROB HALL: You offer no grace period. And I think what you are going to see is you will see more of it in DE because what you have is rather than a big up front one-year cost you can buy domain for one month at a time. So it's interesting if we went out and presented to the domainers a one-month product like what you have in com, I suspect you would find them saying oh, a dollar for a 30-day product? I might take that.
So there are other solutions. I don't want to present it as a possible solution, because I think that's for a bigger group than just myself.
But I think you certainly will see this move into country codes as the domainers turn their attention to other domains as the com database becomes less and less available for registration.
>>SABINE DOLDERER: Rob, just to be very clear, we see these in dot DE already and we see it more than just one year so it is nothing that is completely new to us. But on the other hand side, of course we have a slightly different model and we have also slightly different ways of announcing -- of not announcing actually, of deletion of the name. So you don't get the information about which domains will be deleted in the next couple of hours or seconds in our domain system because we don't think that it's something -- we owe it to our customers so we don't publicly say that this customer will not (inaudible) the next couple of minutes and hours because we don't think that's something he has paid for us when he registered the domain initially.
So we actually protect his rights there, which I think is also very important for consumer protection, also trademark protection.
But we see actually that around about 20% of deleted domains which are published by a list by Google are already reregistered also under the dot DE domain space.
>>ROB HALL: And I think it's interesting, if I speak on behalf of what pool is seeing in the drop catching marketplace, we are certainly seeing more and more attention paid to the country codes, such as CA is certainly a huge one that's coming on in the deletion market. UK and DE have always been strong. But they are getting stronger and stronger as the registration or the expiry market in com moves the registration upstream from the deletion. So less and less domains are deleting in com than ever before and the market is drying up so to speak. So we are seeing strictly in the drop catching names, not the add/drop but more and more people turning their attention to other TLDs.
And I hope Nigel, I hope that touches on your question, too. I don't know if .GG will see the volume of registration that a DE or a com might. But it's possible that you would see some add/deletes but I don't know about your grace period. I can't comment on whether you have a grace period or whether that's a grace period issue.
>>JOTHAN FRAKES: We are going to, for time's sake, the intent was to take public comment after the next set of speakers to make sure we are cover all this content.
Part of why I'm here, I and the executive producer of our comment called the domain roundtable and part of what we are attempting to cover here in three hours is something we took three days to cover and barely scratched on. So this is very aggressive and I would request that the folks standing at the mike, you will have an opportunity. I have a few more speakers and then we will take public questions.
And it sounds like to take -- wow.
If you can have some just brief responses, we have one from Tim from go daddy and from Roberto Gaetano, and then I'd like to manufacture on to our other presentations.
>>TIM RUIZ: I will be as brief as possible. I just wanted to respond quickly to what Steve was saying about the agreements.
And just that again, our opinion is that most of the registry agreements don't prevent them from taking some step or taking action to put limits on the add-grace activity. Add-grace period activity. It would be interesting to discuss that further.
The other thing I wanted to just respond to was the fear, uncertainty and doubt. And we're not trying to raise fear, uncertainty and doubt. What we are trying to do is deal with the fear, uncertainty and doubt that the activity in the grace period is creating for our customers. And the comments I read from our customers, those are their words, not ours. And I think they are evidence that there certainly is some fear and uncertainty and doubt being created there.
We are not against the business model of domain name monetization. We are against this activity as a way of getting there. We think there are other solutions, we think there are other ways to do it. And hopefully we can move on to something quickly and not take months -- talk about this for months and years and find out we're still here. This activity isn't abating. It isn't going to go away. It was strong a year ago, it's even stronger today. We don't see any evidence at all that it's going to abate on its own. So it's something we are going to have to do and take action if we want to see it stop.
>>JOTHAN FRAKES: Roberto.
>>ROBERTO GAETANO: Yes, I would like to link my comment to a comment Rob has made in his presentation that ICANN sometimes is making policies without a clear understanding of what the constituencies say.
So I would like to go on record that the At-Large Advisory Committee has a clear position against what it considers an abuse of the grace period by the domain name tasting because it creates confusion to the users because it utilizes resources that could be used elsewhere, because -- for a number of reasons. And that basically, the bottom line is that because the end user who is surfing the net is having a service that is by and large lower than what it would receive in absence of this.
>>JOTHAN FRAKES: Looks like we have one more comment from Berryhill and then I would like to move on to the presentations for time's sake.
>>JOHN BERRYHILL: I wanted to mention as I think Tim meant to, the same contract that provides for add-grace -- you know, the add-grace testing, or the add-grace period, does provide VeriSign with a very broad power to define and prohibit practices which are determined by VeriSign to be abusive.
And it's odd that the conversation tends to suggest that VeriSign is bound by the one provision and not the other.
And I would suggest that the reason we're seeing a move into other TLDs is that there is something of a hubbert exploitation curve with dot com and it may be a self-limiting behavior.
>>JOTHAN FRAKES: I have what I am assured is a very short point from the audience. Please introduce yourself.
>>IZUMI AIZU: My name is Izumi Aizu, also a member of At-Large Advisory Committee, trying to sort of convey the voice of individual Internet users. It's mainly a question to Rob. I appreciate your comment that we really need to solve this as a whole community and I believe that end users around at-large is one of the constituencies, although we don't have any vote on the board or the GNSO council. What kind of conversations might be stifled to Rob or might kind of problems might happen if we take quick solutions or quick actions for the end users or for the individuals who are (inaudible) corporations who are not really speculating in the domain name business. I don't quite understand. If you could explain, these are sort of the possible problems if we take hectic action on this.
On the other hand, so far in Japan, I haven't seen anything like this from the registrars, but it might happen any day, taking example of 18 registrars.
So there is a need to effective action. What prevents us from doing that or what other problems you are trying to address?
>>JOTHAN FRAKES: Rob, before you respond to that, I know we have a policy section toward the tail end of this and I would like to maybe get through the next segment so we can take public comment on the expiring names and let Bruce talk a little bit toward policy. And then we move forward.
Thank you for your comments.
Do you have a fast comment?
>>WENDY SELTZER: Yes, please, it's Wendy Seltzer, also from At-Large Advisory Committee, also trying to bring some public Internet user perspective to the discussion. And to raise the point that sort of the naive economic sense could be that if money is changing hands, everything is okay, and the money will flow to the place where value is and helps us sort things out. Money isn't always a perfect signal, especially not when you're talking about balancing the public's interest against those with money in the game.
The public has a clear collective action problem even if they are trying to say that domain tasting is confusing to them. I've heard from several individuals and small entrepreneurs who have had problems obtaining domain names because they were caught up in a ISOC of added, dropped, added, and these folks were never able to pick up the name.
Other individuals hit -- are confused when they hit search pages, and these pages don't necessarily add to their Internet-using experience, even if, when the user clicks through an ad, it does make money for the person showing the parked page.
So stability from an Internet user's perspective of seeing all of these names churning around and these parked pages popping up and down I feel is also something we need to be considering here. And that's adding to the ALAC's concern that this is not something that helps the Internet user and should not be studied and kept alive in perpetuity but should be ended.
>>JOTHAN FRAKES: Thank you for your comments. Obviously there's tons of different things and different ground we can cover. Would I like to let John Berryhill respond to that, and I absolutely must insist after that comment that we do move to the next segment.
>>JOHN BERRYHILL: Just a very tiny comment. I mean, the beauty of Internet ingenuity is that money changes hands even if nothing happens.
For example, if you type an unregistered domain name into a browser bar, something will happen. What happens is going to be determined by your browser software publisher, or whatever plug-ins or toolbars you have installed on that browser. For example, if you install Microsoft Internet explorer and you type in an unregistered domain name, you will go to MSN search which is a monetized search system.
So part of -- the arm wrestling match going on here is saying okay, we are going to stop add grace testing and monetization of vast numbers of unregistered domain names is putting a dollar into the pocket of Microsoft.
So someone is going to get that dollar for traffic that is seeking a destination, whether the domain name is registered or not.
>>JOTHAN FRAKES: We have another comment from the panel.
>>JOSH MEYERS: I think I would just comment that the user behavior is at the core of the issue, and that typing in domains into a browser is an activity that, at least in the foreseeable future, is unlikely to stop.
And so through Yahoo!'s participation, we're able to shape what the pages are going to look like and how much value that they add which I think is extremely valuable.
>>JOTHAN FRAKES: And thank you for the comments on that.
The next segment to this session which we're doing in the best time we can, we started talking about domain name practices with the -- if you'll review your handout, the -- we were talking about the front end on that first segment. Now we are going to be talking about things that occur after the actual physical expiry of a domain name.
Our next presenter is Jonathon Nevett, and he is talking -- he is the vice president of chief policy council for network solutions.
We are going to be talking about the expiring name practices which is the handling of names as they approach expiry or deletion and how it's evolving and what changes are coming with the evolution of the registration experience from the registrant and we are going to hear registrar and registry perspectives on this, and with that I hand it over to Jon Nevett.
>>JON NEVETT: I would like to echo what Dave Maher and Steve Crocker said that there are clear signs that there are unintended consequences with domain name re-registrations. Essentially, all domain names, whether they expire or delete are subject to re-registration by someone outside of the control of the former registrant.
So if you take that as a given, an educational campaign is very appropriate, and we would support that, and we should educate registrants that -- of that fact and encourage registrants to re-registry their domain names.
Especially if they are registered with network solutions.
Essentially, less there be any doubt, we are incented to have our customers renew their domain names. We send four e-mails prayer to expiration. We send postal mail if the e-mail is bounced back.
We have -- many registrars have what is called an auto-renew product or service where a registrant could give us credit card information and then we would auto renew their registration. Some registrars do it as an opt-in basis, some registrars do that as an opt-out basis.
Once a domain name reaches its expiration date, different registrars treat the domain name differently.
We, after an expiration date, send another two messages to the registrant, encouraging them to renew their registration. We provide what we -- what I'll call a registrant grace period where we let the registrant have 30 days to renew their registration at the typical price that we're charging, and wean courage them to do so.
After the registrar grace period, there's what's called the redemption grace period. None of these grace periods are required on registrars. I should note that. That these are performed by registrars because it's as a service to our customers.
There's been a change in processes in the last year or so where many registrars are providing additional services for customers upon an expiration of a domain name.
For example, some registrars will auction off domain names that are expired and provide customers with a share of the amounts that are achieved or received at auction.
We do that, and we provide the customer the ability to choose not to have their name auctioned if they just want to let it delete. Of course, post deletion, with a drop-catch system, most likely their names will be auctioned off anyway, but this is an opportunity for them to get essentially a share of the revenue that comes in through the auction.
Some registrars are providing what's called an enhanced grace period where they will actually re-register the name on behalf of the registrant, and then provide a window of opportunity, it could be six months, could be three months, where a registrant at any point could redeem the name for the redemption grace period amount. And that amount is -- it ranges during the pure redemption grace period, that amount that we're charged is almost seven times the amount of a regular registration from VeriSign, for example. But they'll charge that higher amount. And essentially what some registrars are doing, not us, is monetizing the domain name during that period so they will have the benefit of the domain name and any kind of traffic that that domain name is getting, and then if the original registrant decides they want to have their name back, then they charge them $100 or $120 or $80, depending on the registrar.
Other names, usually if there's a lack of interest through an auction process, they will go through a pending delete process where the original registrant will not be able to recover the name. And then the name will be deleted.
One other thing we do and other registrars do is usually upon five days of the expiration date -- and people have to be clear, there is a different between expiration and deletion. There is a time period in between. But post the expiration or five days after the expiration, many registrars will essentially take down the domain name. So that that's another warning sign for a registrant, hey, your domain name registration has expired. Your domain name and Web site won't resolve, and it encourages folks to renew their registration.
And I'm hoping Pat will talk about the proposed product that we've talked about in many prior ICANN meetings called the central listing service where that's another opportunity for domain names, upon deletion, to go through an auction process. And that's a product that many of us are looking at and are interested in talking to the community about.
Thanks.
>>JOTHAN FRAKES: Thank you, Jon.
I don't know if we're talking necessarily as much about after-market. Unfortunately, there's a variety of different things we can talk about if in the domain marketplace and I don't know if we are covering ground on the after market or auctioning of names particularly in this session. It might be covered in future workshops.
But Pat will know what he's talking about and I think a good opportunity now to introduce Pat Kane who is the director of business operations and policy for VeriSign.
Pat.
>>PAT KANE: Good morning, what I am going to talk about this morning is what happens at the registry once a name finally gets to expiration from the standpoint of what we see at the registry. Every day at 2:00, eastern standard time, we actually go through a delete process to where all the names, all the registrations that are expiring are actually deleted and returned to available status.
The reason we do it at one specific time during a day is, previously, expiring names would delete over the course of an entire day and registrars, they call it drop-catching, at VeriSign we call it add storming, largely because we would be taking add requests all day long because there was no knowledge as to when those deletes would take place.
So on average, there are 60 million add commands per day for expired registrations.
When we were doing 24-hour a day operation deleting names we were actually seeing 120 million add commands for these expiring names. Most of them occur between 1:30 and 3:30 in the afternoon.
>>JOTHAN FRAKES: Can I ask a clarifying question about that? So this is surrounding the drop-catching type of behavior? Like when a name gets deleted, people send as many adds as possible in order to hit that magic moment of deletion and be the first one in? Is that what you are talking about?
>>PAT KANE: Yes. What happens is registrars have actually done an analysis on our deleting program, such they have estimated how many deletes will occur per minute, and when that will happen over the course of that minute. And so therefore, they send a bunch of adds, not really waiting for a response from that add, to actually acquire those domain names.
Does that clear that up, Jonathan?
>>JOTHAN FRAKES: Thank you, yes.
And for people in the audience, we're talking about the furthest end of the diagram where the name is released and available, I think.
>>PAT KANE: Now, during this period, over the course of the month of May, we saw on a typical day as many as 564 different accreditations are trying to acquire those specific deleting names.
Currently within com and net there are 651 production registrars. So there's a large number of registrars that are trying to acquire these names.
On an average day, we are deleting 26,000 names on average. The min/max on that is 21,000 to 42,000 which works out to be 2,276 attempts per deleting registration.
That constitutes 99.9% -- well, 99.9% of all those names are actually re-registered. So there's days when there are 100% of those names that are re-registered and days when 97% of those names are re-registered, but typically it's about everything that's available.
Of those names, what makes it through the five day grace period is about 1300 of them. That's a stick rate of almost five% which comes to about friction,000 add attempts per paid registration. So these names are having some service process against them and it is different from the massive numbers of registrations that we see during the add grace period, but still, there's the opportunity to not hang onto these names if they prove to be not viable within whatever service a registrar provides.
Thank you.
>>JOTHAN FRAKES: Thank you for your brevity. And it sounded like Paul Stahura had a quick question about that.
>>PAUL STAHURA: Just real quick.
You said 21,000 to 42,000 names per day were deleted? That does not include names that were registered during the five-day add/delete and then deleted; right?
>>PAT KANE: No, those are the number of names actually in the pending delete status that will be returned as available.
Names that delete during the add-grace period are returned immediately to available status.
>>JOTHAN FRAKES: Good clarification.
I see Steve Crocker once again at the mike.
I knows Dave Piscitello covered some security and stability issues in the SSAC meeting, and I think we may hear about that now.
Steve.
>>STEVE CROCKER: Thank you very much.
I think it's helpful to have -- to -- a little bit of -- to make this picture pretty vivid.
As the names are deleted, we have the registrars competing vigorously with repetitive attempts to -- by banging on the system, to try to register.
There are 600 odd registrars.
But there's actually only about a third of that many who are real companies.
And the rest are created for the sole purpose of having access for the purpose of banging on the system.
That is, all the registrars are supposed to be treated equally.
So much bandwidth is available to each one of them.
And so it becomes financially rewarding to go create a registrar pay the several thousand dollars to ICANN, go through the accreditation process for the sole purpose of having yet another thread that is banging on these systems.
I often try to explain this system to people who are unfamiliar with it, that what amounts -- it's a little more complicated than what I'm going to describe, but it's basically a simple probabilistic system of just taking as many whacks at it, since some registrars have more threads than others because they have a family of registrars, it's kind of a weighted system.
So this amounts to what is a weighted lottery system.
And if I were teaching a computer science class and I said, here's the problem that we want to solve, class, we have names that become available, we have a number of people who might want them.
We have to have some sort of allocation system.
We want to implement a weighted lottery system.
Please go off for your homework assignment and design one of these systems.
And I would expect that the kind of designs would come back would be that you would sort of buy a ticket and it would be put into a basket and at some appropriate time a ticket would be picked out of a basket, the way we do in all kinds of lottery systems.
The student who comes back and says, "I'll tell you what, we're going to implement this by having a spin-loop," which is a term of art in computer science, "Implemented in the slowest and most expensive possible way by doing it over the Internet, by having hundreds of systems banging away across the network in a grossly inefficient way, and putting maximum load on this one registry system.
And it's sort of easy to implement and everyone can do it."
My reaction if I were grading the paper is, first of all, that is a really bad design.
And second of all, you're very dangerous.
Please do not design systems for anybody else.
Please choose some other line of work.
So it is kind of an embarrassment in our industry that that is the mode in which these things are being allocated.
So from an engineering perspective, this is just outrageously bad.
[ Applause ] applause.
>>STEVE CROCKER: I think I made my point, so I will subside.
>>JOTHAN FRAKES:Did you have a question?
>>STEVE CROCKER: No.
A comment for clarification.
>>JOTHAN FRAKES: Paul Stahura and then Bruce Tonkin have comments on that.
>>PAUL STAHURA: Everybody knows that it's a crappy system.
It's the system we've got.
I mean, there have been other proposals.
WLS comes to mind.
CLS.
You know, all kinds of other systems.
A ratio method.
But one system that was actually implemented was a lottery, by dot biz during, you know, their sunrise period.
And they got sued for it and lost.
So I don't think a lottery is the right way to go.
But there are plenty of other systems that would be better, in my opinion.
But this is the one we've got.
>>JOTHAN FRAKES: And then Bruce and then Rob Hall had a comment and then John.
>>BRUCE TONKIN: I think my comment would really just echo Paul's.
Certainly the customers of the registrars -- of the registries, being the registrars, have suggested numerous solutions, none of which have been taken up as yet.
>>JOTHAN FRAKES: Rob.
>>ROBERT HALL: Steve, I agree, it's from a -- from a technical perspective, horribly inefficient.
The one thing it does do, though, and I would probably be open to discussing any new solution that kept innovation possible between registrars and kept the competition and the competitive market we see now existing, 'cause it's only through a that we've seen a decrease in pricing, we've seen different registrars implement different business models.
So I'd love to find a technical solution, and I have to again compliment VeriSign, because they've implemented quite an elegant one for the volume and the problem they see, the volume of registrations they see.
And now, in fact, I would encourage other registries to look at what they've done, because they are able to handle the volume and it is going to start to move up to other registries.
But the interesting thing is, from the acquisition registrar phenomena that's been created by people registering for registrars to get access to this acquisition market, whether it's lottery or some other means, you may still see that same phenomena or a new one you don't predict.
So, again, we need to carefully consider, make sure there's competition and innovation still allowed.
I mean, in the current market, as inefficient as it is, it's very possible that eNom will innovate in such a way that puts pool out of business in a month.
So that's a market risk and a business risk we all take.
But it's one that I think's important to keep the competitive nature of the market.
>>JOTHAN FRAKES: Thank you.
And then John, and then we have a presentation -- we have presentations to follow and then more public comment.
Thank you, John.
>>JOHN BERRYHILL: One thing that's worth knowing in terms of a marketplace perspective, and, of course, working with a number of people who put substantial sums of money into Rob's pocket and Paul's pocket and John's pocket, is that the notion that out of these 22,000 names that should expire on a given day, the registrars themselves have amended most of their registration contracts away from the default policy provided by ICANN for handling deleted names.
You will notice if you look at, say, the top several registrars' contracts these days, there are specific provisions dealing with transfers of a domain name after expiration.
So at least in that sense, the largest registrars have taken names that would otherwise be going into contention in the drop pool and moved them into other sorts of market systems for reallocating them, independent of expiration at the registry.
>>JOTHAN FRAKES: I wonder if those changes might be part of a source of confusion for people in the litigious field, such as the incident that Sarah and Paul were dueling over earlier.
It's certainly -- these changes do make for confusion.
And that might be a response to one of the ALAC questions, at least from my point of view, that these changes do create ripples, you know.
>>JOHN BERRYHILL: Tremendously.
I see plenty of situations where people have not paid their renewal fee, the domain name has gone into redemption status, which is, I believe, the source of one of the numbers we heard earlier, and registrants are put off by the redemption grace period fee.
And that's a fee that's passed down, you know, from the registry.
And, unfortunately, many of these people are told, well, if you wait a certain amount of time, it will become available again.
And, you know, really, in order to figure out how to pick that name up, you have to know, if it's a network solutions name, it's going through one mechanism.
If it's a registry.com name, it's going, you know, through another mechanism.
And that, you know, there are names that will never drop into the delete pool for contention because of various deals that registrars and after-market participants have made to route around the actual deletion cycle.
>>JOTHAN FRAKES: We have one more quick comment by Tim Ruiz from Go Daddy, and then I'd like to go in to give Paul an opportunity to present on long tail.
.
>>TIM RUIZ: I just want to point out, though, that if some of the other solutions that were manufactured, that, you know, forced that back 100% to the registries, such as WLS or CLS, isn't going to make it any easier for the registrant in those cases to actually get the domain name.
It's going to be just as difficult, just as expensive.
And if they don't understand the system and have the right people working for them, it's not going to be any easier for them to recover the name.
>>JOHN BERRYHILL: Oh, yeah.
>>TIM RUIZ: So nothing changes there.
>>JOHN BERRYHILL: It helps to know what you're doing.
And, Tim, there are people here who were at Washington, D.C., at the hearings held in Congress where Go Daddy was dragged through the mud as being a registrar that extorts people by charging $80 to renew their name after the first term.
And to sit there and hear a Congressional committee being told that when it's clearly an RGP situation, you know, there's just so much confusion that people do have to know what they're doing.
And I think that David's on the right track with, you know, just making sure that this information is put in front of their face and not buried in a registration agreement.
>>JOTHAN FRAKES: I'm hopeful that this workshop will put things forward.
We heard the term "RGP."
That is redemption grace period.
It is a period after a domain expires and is not auto-renewed.
There's a period called the redemption grace period, which lasts 30 days.
And during that time, there's a special -- a higher registry fee that a registrar must pay to renew the domain and pull it back into an active status that's what's being discussed in this final component.
I'd like to introduce Paul Stahura, from eNOM, who has a presentation on a concept called the long-tail proposal.
I'll go ahead and pass the baton to you.
>>PAUL STAHURA: Thanks.
Now that everybody is ramped up, it's a complicated system.
And I hope at least you walk away with the understanding that it's -- there's a lot of processes and it's very complicated what a name goes through.
But now that, hopefully, you understand some of that, at least, I'm going to introduce an even newer concept.
First I'm going to talk about the long tail, fat front, torso, what is all that.
Then I'm going to talk about a new type of domain name.
It has different parameter settings.
Then I'm going to talk about how that kind of relates to the add/deletes and name-tasting.
Then I'm going to talk about some benefits that this proposal has.
So here's a graph of the names.
In this case, it's dot com.
But this could apply to, you know, the whole name space.
I used dot com as an example here.
And they're ordered going from left to right.
And with the most-valuable names on the left and kind of the least-valuable ones on the right.
And you can see kind of it diminishes, obviously.
And the area labeled "fat front," that kind of grayed-in area, that's all the names that are registered in dot com, VeriSign, okay, times six bucks is the revenue that VeriSign gets.
So that rectangular area represents the value that VeriSign gets in registry fees.
The region above that is the value that the registrant gets.
For example, MSN.com gets a lot of value for their $6 that they -- essentially, that they pay to VeriSign, to their registrar.
And then there's other names.
I checked Seattlebakery.com.
It was registered when I checked it.
Probably not worth as much as Google or Yahoo!
But still worth something more than $6.
Then we have the little label -- the little area named "torso" in between A and B.
That's the area that is names that are not registered but are worth more than $6.
So those are the names that these add/delete people are trying to figure out which ones those are, and then they monetize this little triangular region here.
That's their profit.
They pay VeriSign this much, $6, and then they get this much from PPC most of the times.
Another interesting thing speaking of PPC.
I did this graph, you know, I think -- names towards the fat front.
Most of their value is not because of PPC.
But most of the names towards the long tail side, I think most of their value is because of PPC.
So, then, proceeding on to the right, we have the long tail part, which is those names are worth less than $6 and are unregistered.
So there's really no point to registering those, you know, economically, because if you monetized it with a PPC site, you know, from Google or Yahoo! or whoever, then you wouldn't make back the registration fee, so it doesn't make economic sense.
And then I have some names there like Seattlebakedgoods.com.
Last time I checked, that was not registered.
That's probably because it's worth less than $6.
And then I have one way on the right there, ztodayfortoast2.com does not have much semantic context, so it's worth even less.
One of the trends in PPC is it's increasing.
More advertisers are coming online, people are bidding on clicks.
So that curve especially towards the long tail are going up.
These names are more valuable, esoteric names.
And there's other reasons why the more esoteric names are more valuable.
Because if you're on Yahoo!, you're probably going to pay more for a click for Seattlebakedgoods than maybe for Seattlebakery, or just Seattle, because if you just buy Seattle, nobody's going to convert to your bakery goods product.
But -- so you'll pay more for the more esoteric name.
Anyway, PPC is increasing.
That's why that dotted line is there.
So now I'm going to talk about -- this is a review.
For dot com names, everybody pretty much knows what these different parameter settings are they registry, okay.
So for a dot com name, the add-grace period, it's five days.
The registry fee is six bucks, ICANN fee is 25 cents.
When those fees are incurred at the exit of the grace period and there's other parameters, like the registration period, usually the minimum is one year for dot coms.
And here's a big fact.
Availability during registration for dot coms, currently, they're not available.
Once you register them, it's obvious to all of us, thinking about it, nobody else can get it once you have it.
So they're not available during registration currently.
Name servers for each name when you register a name, you're allowed to put any name servers on it.
And WHOIS, everybody knows it's required.
And there's other parameters, like whatever the RGP length is and so on.
So a proposal is to create a new class of name.
So still a dot com name, but it would have different parameter settings.
For example, you know, I have the -- I call them class I names are the current names that everybody is familiar with.
Class II names is this new type of name.
So the ones in blue are the changes.
So, for example, the add-grace period, it's five days currently, with the class I name.
I'm proposing a new type of name with 365 days of grace period.
So that means that if you were to delete your name during the period, you would get your money back.
And you wouldn't pay -- it's -- it's exactly like the dot com name, but its period is longer.
I'll get to all the implications of it in a second.
All these other fees are the same, still costs 6 bucks, 25 cents.
This other one is the same.
You pay those upon exit of the grace period.
The registration in the current type is one year.
But in the new type, it's up to one year.
So it might be less, which I'll get to.
This is a very important part which I have bolded here.
It's available.
So if a registrant has a class II name, it's still available to any other registrant as a class I name.
So if, you know, Joe registers ABC.com as a class II name, Mary could take it away from Joe immediately and register it as a class I name.
And, you know, writing some notes about the comments that everybody was making, and Tim had, you know, some of the -- some of his issues with taste-testing was that he had three comments from his customers, and all of them were, my feel, is because those names that other people were tasting, his customers were confused, because they were unavailable when they wanted to get their name.
Like, they would try a name.
It's available.
Then they didn't register it.
They tried to register it a couple of days later.
It's unavailable.
If it was still available during that period, there wouldn't be any complaints.
They could just go ahead and register it.
Another important difference is this class II name has not any name servers.
It has specific name servers.
So when you register it, you can't set the name servers on it.
They would default to a specific name server.
The add-grace period threshold, there would be none in the class II names.
As soon as you register it, it's registered.
And -- but there would be a change if we implemented these class II names, there would have to be a change in the class I name to prevent some gaming, which I can get to, too.
But, basically, we would need to get rid of the add period -- add/delete grace period for a class I name by -- I have two examples here.
Maybe we charge a fee.
Or maybe you have to keep X percent of them, and so on.
The WHOIS is still required for both name types.
And then, of course, there's the other parameters that would be the same.
And then we get to rev share.
So currently with a name, you get to set any name server on it.
You can park it or you cannot park it.
You can put up your Web site.
You can do whatever you want with it.
But with a class II name, you don't get to pick your name servers.
They would be defaulted.
And in the name servers, it would point to a monetization page that would be provided by the registry.
The names -- so, therefore, the names would get monetized by that page, and the money flowing from the monetization, from the ad network, like Google or Yahoo!, would go to these parties -- and I propose these rev shares.
And I propose the rev shares -- now, this is just a guess, but I propose 55% for the registry, 5% for ICANN, and 40% for the registrar.
And why these shares?
Rev shares?
Is because, well, you know, if, on average, the names that are registered as type II names -- or class II names, if, on average, they're worth 5 bucks, less than 6 bucks, 5 bucks an average, 5 times 5% is 25 cents.
That's the same as dot com.
I made some other assumptions and came one these percents.
Of course, all this revenue would be after the ad network gets their share.
Speaking about the specific name servers, all these class II registrants, they would consent when they register a name to specific name servers.
In those name servers, they would point to the site like Google or Yahoo!
On the search site, it's centralized.
So that minimizes some complexity.
In other words, if we didn't allow that and allowed, you know, the registrants to point their name to any -- you know, a multitude of monetization sites, it would be very hard to figure out -- to collect that revenue, A, and to figure out if they're trustworthy and reporting the right revenues.
With one site, the quality of that site for users can be maintained at a high level.
So from the advertisers, I bulletted this because it's important, too.
From the advertisers to the ad network, to the registry, to the registrar, and the registrant, all those parties have an incentive to provide the best, utmost quality user experience for the user at that site.
Another benefit, I guess, of having one site is the ad network, the total revenue from the ad network would be maximized because the traffic would be aggregated as opposed to spread across a multitude of sites, other reason is there's less development costs.
We don't have thousands of people developing this site.
Just one.
And then also, this offering remains competitive, even with one site.
Because different registries could have different search sites.
They could, you know, play the networks off against each other to get the best rev share.
The different registrars, they would offer different rev shares to the registrants.
You know, they would offer part of their rev share to the registrants, depending on, you know, how much money they wanted to make and other factors.
Just a couple observations.
If tasting was a problem, this new type of name may be the solution.
Even if we -- with just type I names, if we change the pricing on it or, you know, maybe did some other things to get rid of the -- or try to minimize the add-grace period tasting, it would still exist.
Just the threshold would be higher, I guess.
Another thing is, existence of these new type of names would eliminate this need for tasting during the class I add-grace period.
Because you could just get a class II name.
Why need to taste it?
And then I also mentioned that class II names requires a modification to the class I add-grace period to eliminate it, to remove some gaming, which is a detail, but still needs to be handled.
So if we could do, like I said, a couple different ways to get rid of that grace period, you could keep "X" percent, you could charge a small fee, or you could just make the period zero.
Another observation, all the class II names, they all have a registrant.
So here's my last slide, some benefits.
It opens a large, in my opinion, new market for competition.
This is the market where names are worth less than 6 bucks.
Opens that long-tail area up to competition.
Another benefit is these names are not locked up.
So it solves kind of the problems that Tim is talking about, some of the other people here, like David Maher, are talking about, you know, deleted names being picked up and used for weird purposes, you know, deleted names being tasted after they're registered.
They can be registered as class II names.
So they're not locked up.
Another one is, the registrant cannot hold the trademark holder's feet to the fire.
He can't call up a trademark and say, I've got your name, pay me a thousand dollars for it.
Because the trademark holder can just register it for $6 as a class I name.
So all the taste-testing that's going on now with the trademarks, the trademark holders could just register them themselves.
Just immediately take it away from the taste-tester.
Another benefit, all the players' incentives are aligned to provide the best user experience.
Right?
The ultimate goal -- the goal is to ultimately help them get to where they want to go.
And another benefit is reduces the load on the registry by removing the add/deletes incentive, you know, to monetize that torso region.
And also eliminates all the name rolling or, I guess, they're calling it kiting, you know, the repeated add, deletes, add, deletes that happens.
And I think I already mentioned this part, the class II name's viability solves the "I tried to register a name, but it was being tasted, so I couldn't," complaint.
And I'd like to -- you know, if you need to ask me questions, here's my contact information.
>>JOTHAN FRAKES: Thank you, Paul.
And the long-tail proposal, I didn't read the introduction, but it's a proposal that's been introduced to address a variety of different monetization activities and the add-grace behavior.
It was meant to describe the proposal and kind of solicit some feedback which we'll hear right at the end here.
I did want to give Bruce Tonkin the opportunity to speak on some of the ICANN policy implications. After I have some questions on Paul's long-tail presentation.
Are there any questions from the audience? We would invite to you come to the microphone and step in.
>>ROB HALL: John, is this in general or do we have a chance to respond to earlier questions?
>>JOTHAN FRAKES: We will make our best effort here. We have the GNSO coming in in 15, 20 minutes.
>>ROB HALL: I understand the timing but we had earlier people we put off and I want to answer their questions so I want to make sure we have time to do that as well.
>>JOTHAN FRAKES: Did you have a particular point you wanted to respond to from an earlier comment?
>>ROB HALL: There were two questions from the floor earlier. One gentleman asked me what innovation would I see being effected. And I think that deserves a serious answer.
>>JOTHAN FRAKES: Can it be brief?
>>ROB HALL: I can do it quickly.
The challenge of predicting what innovation will be effected is just that, it's innovation. And we have seen many, including the add-grace period, the entire secondary market someone said they didn't predict what would happen with it and that was innovative.
I could give you an example perhaps of what I think would happen if one of the solutions being proposed was implemented and that's some people said perhaps we should just add a fee to this. Price fix it and add a slight fee to it.
What would happen immediately is you would cause companies like ourselves that don't participate in this type of market to say perhaps that a legitimate product now and we should ramp up production and take that to our hundreds of thousands of registrants and offer it as a product. And you would actually see the problem increase tenfold as opposed to do what you want with it.
So you would see innovation around the new structure and new rules. And my original point was to say, I'm not saying we shouldn't act or there shouldn't be change. I'm saying you have to be considerate of how the market will react because the demand is not going away. And you don't want to introduce new unpredictable challenges to the ones you currently have and at least understand.
So as long as we are all on the same page and we understand what it is, then we can move forward and I think that's why this education session has been so valuable.
The other question there was they were commenting on confusion will the domains added and deleted and added. And I can assure you for the last ten years being a registrar, we have been having to educate our clients on the fact that yes, domains exist at some points and don't at others. And typically it's a year term that we see this in. And there's been confusion for ten years about what's the term, when does it expire, when does it delete, why is it available sometimes and not others?
All we are seeing now really is that length of time shorten so we have to do more education. But this is an ongoing thing.
The more the registrants get educated on what can happen to their domain name and the timing of it, the better.
>>JOTHAN FRAKES: Speaking of shortening time lines, we are running up against a 12:30 limit so I do want to be able to get to the people from the floor and have time for Bruce Tonkin to speak and make a closing.
Once again, we have from core.
>>MARCUS FAURE: Hi. My question is did I understand correctly that you could turn a class two registration into a class one registration at any time?
>>PAUL STAHURA: Yes.
>>MARCUS FAURE: Then I don't know how this addresses the trademark problem because as soon as you have tasted the name and you see that it's valuable, you will probably want to turn it to class one.
>>PAUL STAHURA: I see your point, but Sarah said trademarks and tasting are linked. I agree with that. Trademarks and domain names are linked, and it's hard to unlink them no matter if you got a name during a tasting period or after the five day grace period. They are linked. Trademarks and domain names, and it's a complicated subject. But we have a lot of rules, laws in the states about registering trademarks. So if you were to register a trademark as a type one, whether you tasted it first or however you got it, you have to deal with the trademark holder. It's against the law in the states to do that.
>>JOTHAN FRAKES: Sarah had a comment also.
>>MARCUS FAURE: So you are saying your proposal would not address that specific problem?
>>PAUL STAHURA: What I am saying is someone was tasting a name and someone else had a problem with it, whether it's a trademark problem or just wanted it like Tim's customers do, they could register it as a type one name, classical name I call it in my presentation.
>>SARAH DEUTSCHE: I think trademarks by their very nature are well-known. Therefore, they drive traffic. So what this system is doing, it's going to create an artificial feeding frenzy where you are forcing people to register and things will be automatically registered in class one that are trademarks. And trademark owners, their only recourse is it's a forced sale. They have to buy everything by every cybersquatter in order to deal with this issue, which is not a great result.
>>PAUL STAHURA: They could also get their listed trademarks as type two names.
>>JOHN BERRYHILL: Was that the response of the trademark community when Site Finder was running? Because every domain name in the Dotster suit, every unregistered possible string containing Verizon was being monetized by VeriSign when Site Finder was running and there was not a single trademark action that resulted from the monetization of the total unregistered name space.
>>SARAH DEUTSCHE: The trademark community was complaining very mightily about Site Finder as well.
>>JOTHAN FRAKES: Okay.
Next in our queue is -- we have another question?
>>CHUCK GOMES: Chuck Gomes from VeriSign again. I have several comments based on a variety of comments by different speakers, from here and from the platform.
First of all, I would like to clarify how the add-grace period came about because I was directly involved in that. It was first of all not a policy development process. When the shared registration system was introduced in 1999 so there could be competition at the registrar level, there was no five-day add-grace period.
What we discovered several months later in responses from our customers, which at that time were mostly test bed registrars, some of you on the platform, is that registrants would sometimes make a typo and there was no recovery for the registry fee under that scenario.
So we actually proposed to registrars at that time and to ICANN that we be able to introduce that grace period.
It was not part of the first contract for com, net and org.
In the renegotiation that occurred in 2001, it was incorporated as part of the contract. So there really was no policy development process.
The initial intent was for typos and to allow mechanism to deal with that.
Now, I am not advocating that that's the only valid purpose of the add-grace period.
For example, I believe that many registrars use the add-grace period as a time to deal with fraud. That was not anticipated when we added the add-grace period. That's an example, I think, of a very legitimate use, in my opinion, of the add-grace period.
Next, I wanted to address a comment that Steve Crocker made. I think, if I heard him correctly, say that registries were opposed and then registries were okay. I don't think that's accurate. The registry constituency, to my knowledge, has never taken a position on this.
There were some registries that are still opposed to this. There are some that have dealt with it, as Pat pointed out. VeriSign, after exploring various options finally just decided to deal with it and accommodate it.
Now, Tim made a comment that registries have the option of taking action. He is absolutely correct. John made reference to it later. And consequently, what we did last year, in one of our joint registry/registrar meetings and more than half of you at the table were in that meeting, I actually brought up the issue with the registrars and said, okay, where are you guys on this? Are registrars -- do registrars have a position?
And at first there was kind of leaning one way and then all of a sudden others spoke up. There was no consensus as far as I could determine. And we didn't think it was wise for us to take any unilateral action, nor would we probably have been allowed to do that, in that regard.
But we have explored many solutions since then. And as Bruce suggested, there have been many suggestions made for solutions. And we've explored many of those with many of you. Especially on the registrar side.
And once the policy -- the new registry service policy for introducing new registry services is completed, and a lot of work is going on on that, and it is close, okay, there will be opportunity then to deal with some of these issues, including proposals like Paul proposed, if the community supports that.
RGP was mentioned several times with regard to registry price. I think it's very important that you understand how the price was developed for that.
The basis for that, from the registry point of view, which is $40, and that's in the RGP policy, it was not established by us, it was not established by the registry constituency. The thinking on that was that if you make it too low, then you're going to create some of the same issues that we're seeing right here in the game playing and so forth. And I don't mean game playing in a negative sense, but if we set it too low, and this was conversation not just from me but from multiple people in the community.
Finally, let me compliment the organizers of this workshop for what I think was very excellent sharing of valuable information from various points of view and often different points of view, and we need that.
Thank you.
>>JOTHAN FRAKES: Bruce Tonkin has a quick response to Chuck, and then I fear Vint Cerf has the microphone.
>>BRUCE TONKIN: Chuck, just before you --
>>VINT CERF: This is a question for clarification. Do you want to come back here? Go ahead.
>>BRUCE TONKIN: It leads on to something I will be talking about next, but Chuck you made the comment that the policy was not complete with respect to registry services.
I know the board has voted on it. Do you mean the implementation?
>>CHUCK GOMES: I'm sorry if I said that, Bruce. I didn't mean to say the policy is not complete. The implementation of the policy is not quite complete. It's very close. And a lot of hard work has been going on so there's no criticism of anybody in that regard.
We were informed yesterday that it's very close. We'll see it in short order, I think, being able to be implemented, which will then provide a mechanism to introduce new services like this.
>>VINT CERF: Okay. Vinton Cerf with a question for Paul Stahura.
Would you say what happens at the end of the year to a class two registration? Is there a charge or no charge? It's a free thing or what happens?
>>PAUL STAHURA: It would be just like the end of a five-day grace period for type one. It would automatically be charged $6 and that's a regular name. You have to delete it.
>>VINT CERF: Essentially it's a one-year redemption -- add-grace period, which is what you said, but I wasn't quite clear.
>>PAUL STAHURA: Correct.
>>VINT CERF: Could I add one question which is irrelevant to the whole proposal, which I find intriguing by the way and it might not be Paul who answers it. It might be Pat Kane.
When a domain name has expired, and then it's re-assigned to someone else, what happens to the SOA record for that domain name as to its start date? Does that change automatically or does it stay the same or are there circumstances where a domain name changes hands but it doesn't look like it has if you are looking at its birth date?
>>PAT KANE: If a name is transferred, the date stays the same.
>>VINT CERF: Okay. So that's a problem for Google. And it's one of the reasons that we became a registrar, but it didn't help.
We were hoping that we could detect that something had changed hands and that, therefore, we should invalidate a lot of things sitting in the cache that referred to the former content of that domain name.
>>PAUL STAHURA: I can answer that. You can still do that because you can monitor the WHOIS and if you see the WHOIS change, then you note the date.
>>JOTHAN FRAKES: There's a service called domain tools --
>>VINT CERF: Let's take this off-line because it's clearly not part of the main line.
->>ROB HALL: The easiest way might be to monitor whether the DNS server changes or not. Because most often when a domain changes hand, the DNS changes as well to the new owners. You could certainly say hey, DNS changed. And it would typically be the only activity you will see this in is the 45-day period. So if you are within 45 days after the renewal date and you see DNS change, it's most likely a transfer. But we can take it off-line.
>>JOHN BERRYHILL: In short the create date has become completely unreliable because of the registry route-arounds, so you can't rely on the creation date anymore.
>>JOTHAN FRAKES: We have one more question or comment from the audience with John Klensin.
>>JOHN KLENSIN: John Klensin. Paul, when I look at your proposal, something springs out at me, and I just wanted to comment on it, especially since it relates a little bit to some of the earlier comments about trademarks.
It seems to me you have a potential for race conditions, as people decide they want to convert something from a type two -- from a class two to a class one.
And my experience watching this business over the last several years is the people who are trying to figure out ways to monetize this and thrash registries and utilize race conditions are far smarter about what to do about race conditions than I am. But what I have generally noticed is that every time in this situation we see a race condition, we see new and creative ways of using it.
And following up other comments which were made on the panel about doing things quickly and in a knee-jerk way, something which involves race conditions and has not been very, very thoroughly analyzed as to how this can happen gets scary at this point concerning things containing race conditions. And I will spare the audience but there is a general computer science principle there, too.
>>PAUL STAHURA: I call them fringe cases that creative people, I guess, utilize. But I only had ten minutes to present the whole thing. Maybe I could talk to you after this. I don't think there's that kind of condition. There is something that requires a type or a class one name to -- for us to get rid of the add-grace period in a class one name or minimize it, which I could talk to you about. But besides that, I can't think of another way of gaming it. And I'm pretty creative.
>>JOHN KLENSIN: The most obvious one is two holders of two putative trademarks recognizing a class two name as being potentially interesting and both going after it.
>>JOTHAN FRAKES: That's a good point.
>>PAUL STAHURA: Still first come, first served.
>>JOTHAN FRAKES: And can I encourage you to extend your dialogue to after the conference, maybe in side bar, because I do want to respect the GNSO's time coming in after us.
I have Bruce Tonkin with a brief note, and then we'll do our closing.
>>BRUCE TONKIN: Thanks, Jothan.
I will try to do this relatively quickly.
Firstly, I just want to identify a couple of ICANN policies that are related to the discussions we've had.
The first one is the one that Chuck just mentioned, which is the GNSO developed a policy for registries to request a change to their contracts to allow changes in the operation of the registry.
Essentially, this is a policy that allows a single registry to change its processes provided that the changes are consistent with ICANN policy and don't cause security and stability problems.
And I think as Chuck mentioned, this policy has been approved by the ICANN board. Its implementation requires setting up a committee to review changes for security concerns. There's some implementation work going on behind the scenes.
But that is certainly something that some of the registry operators, when they have said that they have brought something to ICANN, the correct way, if they are going to do it as a single registry operator wanting to make a change, is via this policy.
The second policy that relates is quite an old policy now. In fact, I think goes back to 2003.
This is called the expired domain deletion policy. And that says at the conclusion of the registration period, failure by or on behalf of the name holder to consent that the registration be renewed in the absence of extenuating circumstances, result in cancellation of registration.
Now, it's interesting to see how much I guess the environment has shifted since that was developed. At the time that was developed, the complaint was that some names were basically disappearing from the system and not being able to be used at all because some registrars at the time were not cans selling a name but also leaving the name effectively dark. And so it was just completely unused.
And then we're sort of saying we want to have access to those names earlier. So we want access to those names at the end of the 45-day period.
Since then, we have redemption grace periods, we have calls for potentially holding the domain names for longer periods because people want to have the opportunity to renew them and so on.
So certainly if, from a policy point of view, there's at least a policy there, it certainly doesn't seem to be particularly relevant now in that it seems rare now that names just disappear. They seem to be used in some way.
The other thing that's been suggested a few times is why doesn't ICANN put some sort of ICANN fees into this equation to basically change the economics of some of the practices.
The mechanism there is that registrars have variable accreditation fee provision in their agreement with ICANN, and currently that's implemented such that registrars pay 25 cents per registration or renewal, and they also pay a fixed fee component as well.
One of the suggestions has been that the 25 cents is only paying at the conclusion of the five-day grace period and there's been suggestion should that change.
The mechanism or process there, the registrars I guess collectively would need to agree to such a change, and registrars representing in aggregate two-thirds of all registrar level fees would need to approve such a change. I just want to make clear that the process, if that's something being suggested by the parties, would need to go through the registrars as a group and get at least two-thirds of the registrars in terms of how much they pay to agree to that. And ICANN would also need -- the ICANN board would need to judge whether those fees are reasonably allocated, whatever form of reasonableness you wish to use.
In terms of policy side, if any group wants to actually do policy work, then through the GNSO council they can propose at least the creation of an issues report, which may well form the basis for policy development work.
So any of these registry operators that are finding an issue can go through the registry representatives and request that a policy be considered.
Of course, registrars can do that, ISPs can do that, businesses can do that. And the noncommercial users can do that.
There are also additional members on the GNSO council that have been appointed by the Nominating Committee, so certainly I assume that any other individual could approach them and say we'd really like you to raise this as an issue.
Also, the board itself, if the board was deciding that it wanted to be proactive, the board can direct the GNSO council to initiate a policy development in any of these areas. So I just want to make clear that there's a very clear process for deciding to take this on as a policy work, if that's what the community wishes to do.
Before they do that, it's probably worth just reflecting on ICANN's mission as well.
And that is it coordinates the allocation and assignment of the three sets of unique identifiers, one of which is domain names. And coordinates policy development reasonably and appropriately related to these technical functions.
That's the mission.
There are a number of core values. I just pulled values that seem to be related to this workshop and one is that one of ICANN's core values is respecting the creativity, innovation and flow of information made possible by the Internet, and limiting ICANN's activities to those requiring or significantly benefitting from global coordination.
So I guess one of the differences here, the policy for an individual registry operator to request a change is now there. So an individual registry operator can make a change.
If registries collectively believe that a change should be made that applies to all registry operators and all registrars, then that would presumably fall into they believe it significantly benefits from global coordination.
The other couple of core values that seem relevant here is where feasible and appropriate, depending on market mechanisms to promote and sustain a competitive environment.
And introducing and promoting competition in the registration of domain names where practicable and beneficial in the public interest.
So certainly, I think, without arguing one way or another, it's not for me to say. I'm just speaking on behalf of the council. But those are the considerations the council needs to consider when deciding to commence policy development work.
>>JOTHAN FRAKES: Thank you, Bruce. And we've -- in respect to the GNSO, we are running out of time but I see a certain person who might want to comment. Paul Twomey.
>>PAUL TWOMEY: A certain lunatic would like to comment.
Can I just say three things. First of all, I would like to thank everybody who has participated in this workshop. I think it was outstanding. It was even colorful.
And no, Rob, I'm not taking you up on your opportunity to bash you up, as you put it. Rob suggested to me that it is my one chance where I could sit at the microphone and attack Rob up there on the table, as opposed to vice versa.
>>ROB HALL: After being up here for three hours, I am in deep appreciation as to what the board and you go through at these things, Paul.
>>PAUL TWOMEY: Secondly, I would like to reinforce Bruce's analysis there and bring people back to the instruments that are available for furthering of this discussion.
I would make the observation that the ICANN staff and office receive many of these complaints. I recognize many of those particular e-mails read through earlier in the presentation as things we have also received.
And having it resolved not necessarily by coming to staff or people saying you have to fix this. I have an e-mail coming saying I personally have to fix this, e-mails, and I think Bruce's intervention is useful.
Can I make two slight, relatively small comments but the first one is important, that in many respects ICANN is a steward for the users of the Internet's naming system, for the interest of those users of the naming system.
And I think as the market continues to evolve and change and new markets evolve, it is all part of promoting competition. These discussions become, you know, often one interest versus the other. But I think one thing that's certainly very much on the board's mind is this role of being a steward for the users of the system. And I think that has implications, potentially for things like well informed marketplaces and information where users actually understand what's actually happening. I think that's probably really quite a key thing.
Personally, I think that's a key thing.
Second observation I'd make, it's a little bit related to that steward issue. It came up in the discussions here, but we've seen it in other areas as well, which is the use of legal personalities to distance from undertakings that exist in present agreements. And I'm not saying that that's a -- I'm not saying it's a bad thing. Many businesses have different legal personalities and they use subsidiaries, and people have got resellers.
But I do think this is one of the things that's beginning to emerge if one has a view of that we have a role of steward for the users of the naming system. One of the things that is striking me at the moment in a number of areas, resellers, for instance, is an example. We heard other names about subsidiaries.
There is an issue in my mind, I suppose, of the spirit of which undertaking were made, special RAAs or other agreements, and the realities of how this particular legal personality is not a signatory to this agreement, and then there's lots of policy discussions around that. And I just wanted to -- in other types of similar -- well, in other areas, you find eventually these things sometimes come down to discussions of the group as opposed to the company.
And I'm just wondering about that issue in the back of my mind, about the distinctions between an agreement which is put in place with a particular player and then the mechanisms they use to actually do their business, and then you end up with some, from the user's perspective, it's -- you end up with, I think, from the user's interest perspective, you end up with disconnects between the made undertakings by one of the entities and the behaviors or other things done by associated entities. It's a personal observation.
>>JOTHAN FRAKES: Thank you, Paul.
Well, in closing, I would like to first of all thank all of the panelists who contributed to this session. It was very -- I'm very grateful personally for your participation, and I know it's added a lot to the session.
The session was meant to be an informational session, to educate people on what's going on in the marketplace. We did cover some good ground.
We covered it in a concise amount of time. As I said, the domain roundtable, we cover it in a three-day span so we were lucky to cover it in a three-hour span. I would like to thank John Berryhill, Josh Meyers, Roberto Gaetano, Sarah Deutsch, Tim Ruiz, Rob hall, David Maher, Jonathon Nevett, Pat Kane, Paul Stahura, and Bruce Tonkin. Thank you for participating in this session.
[ Applause ]
>>ROB HALL: And on behalf of the panel I think we would particularly like to thank Jonathan. He spent many hours pulling this together and marshalling time. So thank you, Jonathan.
>>JOTHAN FRAKES: I think Tim Cole had a couple quick comments. Okay. That's easy.
With that, we adjourn this workshop, and thank you very much to all the attendees.
>>BRUCE TONKIN: I will just put on another hat which is chair of the GNSO. There is a GNSO public forum about to commence. I'd suggest that will commence in ten minutes' time. So that's ten minutes to 1:00 that the GNSO public forum will commence.
(Session concludes)