Transcription of the ALAC discussion on vertical integration 29 October 2009

PDF Version

Alan Greenburg: A couple of points before we have two people in the queue at least, maybe three. We're expecting Margie to be here any minute now. She is – I'm sorry. I – hello, Margie. I didn't see you come in. So we will take a break to let Margie give her talk and go on to that section, then resume on this. I ask people, when we do come back, to try to be a little bit brief. The buses leave in 45 minutes and it will be nice if we at least gave everyone a change to say something before we finished. Margie?

Margie Milam: Hello, everybody. I'm Margie Milam. I'm with ICANN policy staff. I think I saw you – most of you a few days ago. You guys have asked me to talk about the vertical separation issue. And I'll give you a brief overview of what's going on in the GNSO and what the debate is all about. And then I think we can open it up for questions to see what your thoughts are.

At the moment, as you know, we're trying to develop the implementation model for the new gTLDs. And as part of that analysis, there's a draft document that is the registry, the proposed registry agreement. So the debate is regarding what kinds of rules go into this registry agreement that deal specifically with whether a registry can own a registrar. One of them's an ownership issue. The second – and vice versa, whether a registry can own a registrar or a registrar can own a registry. Whether they could have affiliates, so it's not total ownership, but maybe partial ownership.

And the other issue that they're debating at the moment is whether – if there's no ownership restrictions, whether there should be contractual provisions that apply when a registry is owned by a registrar, or vice versa. And so that's what the debate is all about right now.

It started early – maybe a year go when the – when ICANN commissioned a report. It's called the KRA (ph) Report. It was a report that looked at the registry/registrar market to come up with recommendations on whether the policy should change on registry and registrar separation. And the KRA Report made some recommendations that the community is trying to analyze right now.

And one of the recommendations that the KRA Report made was that there might be – this might be the appropriate opportunity to list some of the restrictions on cross ownership. And one of the recommendations that the KRA report made was that perhaps this restriction of ownership should not apply to certain models, such as a single entity TLD where one entity is the registry and the registrar. And so that – so that's one of the exceptions. And there's also consideration of other exceptions, but the community is debating right now whether these restrictions are appropriate or not.

So what's happening right now is ICANN staff is hearing from the community on whether it is appropriate to lift these restrictions. Whether we can identify what's in the public interest because there's very heated debate on both sides of this issue. And I don't know if any of you attended the registry/registrar meeting yesterday, but – separation – but you could see that they're very polarized at the moment with respect to these issues.

And really, what Peter Dengate Thrush today has been indicating is that he really wants to understand how registrants are affected by this. Because on both sides of the argument, particularly between registries and registrars, they may have commercial interests for pursuing their position. And what ICANN is trying to understand is what is the registrants' view? What's best for the registrants? Does it help them to have these restriction lifted? Or does it hurt them? And that's really the analysis that the Board wants to understand so that they can make a decision on what to recommend in the implementation of the program.

So what is happening on the GNSO side is that Alan – no, not Alan. Mary Wong from the NCUC made a motion within the GNSO to request an issues report. And that's what's happening right now in the GNSO to analyze this issue. It doesn't mean that we're going to undertake a policy development process, but it means that we're going to present a report to the GNSO analyzing the issues and determining whether we should pursue a policy development process (audio gap).

That process is separate from the analysis that ICANN staff is undertaking to determine what to do in the new gTLD program. So we have two tracks going on right now. We've got the track of trying to understand what should be in the recommendation for the next Draft Applicant Guidebook. And that's separate from the GNSO request to understand the issue better. Does that make sense?

So that's where we are right now. ICANN intentionally did not include recommendations in the draft or the – in the latest draft of the Draft Applicant Guidebook because ICANN staff realized that there was so much diversity on this and no consensus on this issue right now. And is looking for input from the community on what helps – what's in the public interest, what helps registrants, what helps improve the marketplace?

And so that – I mean that's basically the status. And I guess I can open up to questions.

Alan Greenburg: Evan.

Evan: I guess through most of this I've just been confused about why there's such opposition to some possible business models. Like for instance, the one I keep thinking of is the Dell model. You want to go to Dell. They don't use resellers. They make the computer; they sell you the computer. Why are business models that will work in other goods, in other businesses, in other services, where a vendor – a manufacturer or the source of something can choose to go through a channel or they can choose to not go through a channel?

Why is this such a big deal here? If people don't like their business model, they don't have to buy that product. If people don't like the way a TLD distributes its stuff, they don't have to get it. Why – could somebody please here explain to me why there's this kafuffle in the first place? Why – if we're not allowing for a large number of TLDs to come in that have different business models, and the ones that work fine, and the ones that don't want to use registrars, if they live or die based on that model, they live or die based on that model. And the worst thing about that is ICANN has to figure out what do you do with a failed registry, but that has to be done anyway. And so what's wrong with that kind of approach to things? If a registry wants to be a registrar, forget the ownership issues. Why – I don't even understand why there's a debate allowed on why this kind of business model should be disallowed.

Alan Greenburg: I have two speakers. Margie, do you want to address it?

Margie Milam: I mean we haven't taken a position as ICANN. I can tell you what the arguments are. Would that help, Evan? Because – and I heard Jay Scott this morning explain why the – from at least the IPC perspective, they thought this is an issue.

In the analogy of the Dell model, what causes concern is that you have the producer selling to a distributor. Right? And then you have the possibility that the producer takes ownership of the computer – or the laptop and sells it at auction. And whether there's – because the way the domain industry works right now, it's not as if it's just a registration to a third party. You sometimes see registrars registering names, auctioning names off and then is it possible that they have more access to information because they are a registry and registrar versus someone who's not affiliated?

So there's this concern that because they're playing in different roles in the distribution chain whether there should be some extra precaution, or restrictions, or contract provisions to deal with the situation that you're finding right now in the dot com and existing TLDs that Alan knows about from the post (inaudible) this year. Or from the post expiration debate, because there's the secondary market for domain names and the concern is that if there's cross ownership that the registrar may have better access to the good names and make money off of the (audio gap).

Evan: And at a certain level, my response to that is so what? If there's an auction for highly desired names, there's an auction –

Alan Greenburg: Okay, let's not try to make it a debate. I've got a speaker (audio gap).

There is one other issue and you're right. Registrars – registries may go out of business. The other aspect however is if this is a particularly successful, valuable TLD, such as dot mail, or dot sport, or something they may have the market by the somethings and at that point they have full control. However, Adam is next, I believe.

Adam: (Off mic) Yes, because you have a Dell model, but you also have a telecommunications (inaudible) where we're looking to separate them vertically (off mic). So you can take (inaudible) or take whatever. But for this particular issue, I suppose we should be looking at things like significant market power, which would be something (off mic) the telecommunications world. I think there was an ALAC statement (inaudible) we may not make a statement but I think Vitorio certainly did that for very small even cultural TLDs, the registry may wish to be the registrar to defend that particular small issue market may not want to (inaudible) by the registrar.

Alan Greenburg: I should have pointed out; all parties agree that in that case, no separation is probably okay. That's not debate – that's not the subject.

Adam: Yes, but that's kind of the opposite of (inaudible) market power in a way. But even looking at the economics of businesses who do study significant markets now may be helpful.

Margie Milam: Sure, and to answer now, if our Richard Tindal, who was making these arguments yesterday, his comment was that the rules as they exist today were developed based on the model from dot com and that dot com at the time we negotiated the new contracts and set up the rules was the monopoly player at that time. And is it appropriate to apply those same principles to these smaller TLDs that might have special community interests, might not have – may never have market power? And so that's the debate is that the rules that are applicable right now may not necessarily make sense going forward because they were designed specifically for dot com.

Alan Greenburg: Sebastien?

Sebastien Bachollet: Thank you. I am not talking on behalf of ICANN. I am – I think really we need to be careful when we speak about ICANN, ICANN, ICANN. Every one of us is ICANN or all of us is ICANN. I have trouble when it's – when we say ICANN say, ICANN say. We can say everything we want, but we have to be care on that.

My point is that historically, I am sure that Richard Tindal is right, but if I recall well, the negotiation between VeriSign and ICANN was not to split between the registry and the registrar, but in the exchange VeriSign accepted to put to the market dot org. To re-delegate to dot org and to re-put on the market the dot net. And in this one they were about to make another bet – I don't know. They can – they were able to be (inaudible).

The – then they decide to split between registry and registrar. But it was after this first long discussion a few years ago. And if we don't have this history, and I was a little bit disappointed yesterday that we don't have the full landscape of this discussion. It's a little bit complicated.

The other point is that I am sorry, but domain name, it's not an apple, either fruit or a computer. It's not an (inaudible) either a fruit or a mobile phone. It's something else. And if we are here today, it's because it's something different. I don't see any industry – somebody told me yesterday when I said that that yes, cartel of OPEC for the (inaudible) may be something. But I hope that we are not doing the same it's something different.

We are not in the market – everybody wants to say that we are of market, market, but where is the market? We create monopoly each time that we create a new TLD. Then we can't imagine that there will be less that come, but the proof was made since nine years that is was the case. Then why we need new TLD? It's not to compete with dot com; it's to allow new community, new usage, new services to come alive for the good of the users. It's not for market reasons. I am really sorry about that. Thank you very much.

Alan Greenburg: James?

James: Sorry, Sebastien. I hadn't – I take a slightly different opinion of the history than you do. But we arrive to a similar conclusion, surprisingly. Actually if you look at the history of ICANN, one of the purposes of ICANN was specifically to look at network solutions back in 1999 when it had absolute market power over dot com. And the registry/registrar agreement that was completed at that time was the result of that action in 2001. Subsequently with a (inaudible) and then how VeriSign was acquired, (inaudible) solution and eventually the network solution was being acquired back to Revising and that caused a problem. And then you eventually reach the history that you talk about.

So there was a little bit of history and the original intention of the registry/registrar separation was rooted on the fact that network solution had absolute power over dot com in 1999. It was the case; there was a lot of debate in 1996. People went to FCC. FCC did not know what to do with it. ICANN was formed or part of the reason ICANN information was to look at that market.

Let me finish. So given that understanding of the difference in history, I do agree with Adam in the sense that when we are looking at this group, market power does matter a lot. Because there is a reason we set a very strict rule on dot com back then and insisting on registry/registrar separation. Because network solution was (inaudible) priced. Prices were $35. The (inaudible) is $35. $35, there's no debate. There's no one else to debate about it.

With the separation, the price went down. Now today on average, you're looking at $10. That's proof ICANN has done its job. But in these new cases when we're looking and the new batch of gTLDs, we need to look at the fact that are these little monopolies? Some (inaudible) yes, they are monopolies. They set their own rules. But at the same time, do they have market powers? And my answer is that they may never have market power. Look at dot names. Look at dot museum or dot cat. They serve a small community; they have a small base. They don't (inaudible) – they charge (inaudible) but people are happy because they serve a very special need. We do not need to apply the same rules that we have on the monopolies on a small player.

Today's com regulation works the same way. We always regulate the market player. We do not regulate the others. That's why the (inaudible) is more heavy regulated than ALAC.

Alan Greenburg: Okay, I had put myself in the list for a very short comment, then we'll go on to Gareth. The issue is not whether the – in my mind is not whether these new ones are going to be monopolies. We don't know which of them will become monopolies and which won't. Some of these top-level domains are going to be auctioned off for a huge amount of price since the people paying that believe they're going to have huge market power. Gareth?

Gareth: I'm going to go out on a limb here and make a leap that I probably shouldn't make. But strikes me that somewhere in the – if you back off from this and look at the broader picture, what we're talking about here when we're – when the issue is being raised about the concentration of power in one – by one group, and the possible elimination of the – of the benefit from the other group, we're – maybe we're talking in this both the reason why the United States and Canada are falling down the list of broadband leading countries the way they are these days because there's too much concentration by the telcos and the cablecos in North America and there's not the ability to allow other players to get involved. Now maybe that's a wrong direction to be going, but that's what struck me when I was listening to the debate the other day. Thank you.

Alan Greenburg: Would you care to finish the sentence though and tell us which is the right answer because of that analysis? I think I know what you're saying, but I'm not 100% sure.

Gareth: As far as I'm concerned, the right answer when you're looking at that particular part of it, if you – the right answer in this particular debate is to not allow the one group to meddle in the other person's playing field. Is that clear now?

Alan Greenburg: So you're advocating maintaining the separation?

Gareth: Yes, maintain the separation.

Alan Greenburg: Okay, I thought that's what you were saying, but I wasn't 100% sure.

Gareth: Yes.

Alan Greenburg: Okay, Adam is next.

Adam: Which part of the contracted user's house are they going to sit in? Which part of the contracted user's house will these joint registry/registrars sit in? Because there's a registrar bit and there's a registry bit. And that's going to be fun, isn't it?

Margie Milam: Because it depends on their charter.

Alan Greenburg: Take your pick. The old bylaws, which expire at 8:00 tomorrow morning specifically said there is nothing to stop you from participating in more than one constituency. The new bylaws I believe do not say that. So I think you have to take –

Margie Milam: They're silent on that part.

Alan Greenburg: I think you have to pick which one you like better and hope your friends and your new constituents, your new safe holder group will not reject you because of your other persona I guess.

Margie Milam: Can I follow up on a couple of the points raised? Some of the presenters asked ICANN to evaluate the market to see what some of the questions that, Adam, you asked regarding whether is there a monopoly in each TLD? There's questions whether there has been sufficient analysis on the domain market. So that some of the observations from the meetings I've seen on this topic is that they're thinking that ICANN should take a deeper look at the market to understand the players and whether there is monopoly power beyond the largest provider.

Unidentified Participant: Just a very quick comment. Or rather, observation. If you do allow – if you do insist on the separation of registry and registrar in this round, then the changes that the people is going to apply for gTLD are mostly from the same gTLD place that we have today. Because the others registrar can also create new gTLDs.

Margie Milam: Can I respond to that? There's – part of the frustration with this issue is that there's new models being proposed. And so even though you're correct, there's a certain limited number of registry operators right now, in the new model they're thinking that some of the players, like Aphelius (ph) and NuStar, and VeriSign will be backend providers. And so they'll provide the – it's almost like they'll outsource their technical backend services and not really be the registry because is the applicant, just say for it's about food or whatever. They're the actual registry, but they're not performing the technical services at all. They don't have the expertise, nor do they expect themselves to get it. They get it by contracting with these backend service providers, which is the reason that Brian (inaudible) yesterday was mentioning that the rules need to change to acknowledge this new model, where there might be outsourcing through various existing registries.

Unidentified Participant: So the (inaudible) the monopoly on a handful of registries, we're concentrating the power – market power to a handful registry operators.

Margie Milam: That's one way to look at it.

Alan Greenburg: It's quite clear there are going to be more backend operators because of the whole bunch of people who were ccTLD backend operators or registrars who have already set up the business. The only question is can they sell their own TLDs? Patrick.

Patrick: Well, on that point, I know that many of those opposing a change in the current situation are inconvenienced with just fees, which happen to also be potential backend providers for new ccTLDs. And providing backend services to ccTLD operators is a very profitable business. And by trying to maintain a (inaudible) in a certain way, they are preventing others.

Another point I want before coming to the conclusion. Another point is that they suggest that very small TLDs could be – could work with an integrated registry and registrar operation. The question remains on what constitutes a small TLD? Is it 50,000 registration? Is it 100, 200, 300? Whatever. But that would make a lot of difference if we – we need to have a clear limit if we accept that we need to (inaudible) a limit on what constitutes a small TLD. Especially in light of the fact that new DAG imposes some on the registry so many constraints that do not exist for incumbent TLDs, among others, regarding trademarks, et cetera, that it's not even reasonable to think that you can run a profitable TLD if you have less than 100,000 domain names.

So if on the one hand, if they say, look 50,000 is a small registry and you don't need separation that means that especially eliminates the possibility to have an integrated approach for small TLDs.

Which leads me to the conclusion that in that case, behind the strict registry/registrar separation, there is also a major (inaudible) issue preventing competition on the backend service side. And I think we should keep that in mind because it's not really addressed anywhere in the dbase we get currently.

Alan Greenburg: I would suggest that on both sides of this debate, there's huge amounts of money involved. I don't think either side has a monopoly on that. The question really is, how – from my point of view, is what's going to be better for the registrant? I know someone's going to make a lot of money and someone's going to lose a lot of money, depending on how the decision is made. I'm not sure I care.

Margie Milam: And I would actually question your assumption that a small registry can't make money. Because in the new model, there is no price cap. And you're thinking – I mean a lot of us are thinking what we see today. Today we see domain names for $10, some nominal amount, but there's nothing to prevent a new registry from charging $1,000, if that suits their model and their expected customer base.

Alan Greenburg: Before you do, it's a quarter past 5:00. We did want to get back to the other thing. We had a relatively long speaker list. We need to decide whether we continue with this one, go back to the other one, or just go to the gala.

Cheryl Langdon-Orr: If I might suggest, I think in the time available, if you complete this speaker list and we move the continuing discussion, which is a very important discussion to part of our later agenda on another day, I think that would be the best way forward. So continue with the current speaker list is my proposal.

Alan Greenburg: Patrick asked to be honored again. That's the only speaker list I believe I've seen. Evan? I'm sorry. I didn't see his hand. Patrick.

Patrick: Just as you said, there's no price cap, but everyone knows that the more expensive the domain name, the less you will sell. So it's as simple as that. So even though small TLDs might find out that they can ask a higher price than mainstream TLDs, in the end they cannot afford to sell domain names for $100 a year. No one would buy it.

Unidentified Participant: Can I just put a reality check on that? There is a little ccTLD that's called dot AG. And they sell a domain name serendipity. AG in German means company. So they sell a domain name for a US $1,000. And they make lots of money.

Margie Milam: You see it in the ccTLD arena. Dot TV names, they have premium names that sell for very high. So that's what registry operators could do. They could save the select valuable names and charge a higher price for those.

Alan Greenburg: Muldova's doing a good business in the US, dot MD, selling to medical doctors.

Evan: So you're making the point. And in fact, ICANN has opened the door for this by making this a high verification zone. I forget what the acronym is or whatever. But there's this now optional designation of higher confidence. Well, if a company has jumped through all sorts of hoops to do that, well maybe their domains are a little bit more valuable than those of those who haven't gotten that designation. I'm getting the idea that that's exactly what this was meant to be. Not required, but probably of some value to those TLDs that decide to go through those hoops. So you're already allowing for something where certain kinds of domains may be of more value for something that has nothing to do with what the letters are.

(Off mic)

Right. But I mean not even – I mean has it been considered for these – like the dot IBM, the designation for a company that just we're going to distribute those – we're going to distribute the second level domains internally. We have no intention of making a business out of this. But we have a real business need internally to use this. I mean has there been interests from organizations like dot IBM or anything like that?

Margie Milam: I don't know. You may know that I just joined ICANN recently and in my prior job I was with a brand protection registrar and they certainly had clients that were interested in exactly that. I mean that's – yes, I'm personally aware of companies that are – a lot of them are hesitant, think they don't want it. But there is a select few that think there might be value there.

Alan Greenburg: Again, there's no debate really that assuming we can define them properly, those who do not plan to sell domains at all, but just use them internally. Or in the case of Facebook, give them away. No one seems to be arguing that they should have to use a registrar for their one domain.

Margie Milam: Actually, Alan, I think you're incorrect. I think the registrars are very much saying you have to use a registrar even in that situation.

Alan Greenburg: I take it back then.

Margie Milam: But a lot of the community disagrees with the registrar position on that as you can imagine.

Alan Greenburg: Evan, your thing is on. Anybody else? I think Sebastien wants to talk.

Sebastien Bachollet: I think the rule of your – I know this dream appears in the price meeting to (inaudible) TLDs. If (inaudible) TLDs it's for inside the company, I don't know why they need it because they are second, three-level, and (inaudible). And we need gTLD. That's gTLD or everybody becomes a US government with his own TLD for his own use. Or it's open. And gTLD (inaudible) meets in open at least to one community. But not just for the inside of one company. And I would like very much that we say that if a company decides to have his brand as a TLD, it's to be open in one way or another. It could be just for their providers, it could be for the customer. But it must be open not just for the employees of the company and not just for internal use. It's not a gTLD anymore if it's like that.

Alan Greenburg: It's a position taken by a number of people, but unfortunately, no one ever put that in the policy. So at this point, it's valid unless we change the rules in DAG 4 or 7.

(Off mic)

Unidentified Participant: I have a question. Actually this means for – a lot more than. If one – I will – if ICANN's purpose or general purpose was try to do some (inaudible) regulation on pricing on dot com, and if this (inaudible) gTLD the ICANN to not do any pricing regulations. How do we serve – are we serving the – our end user in the right way? Just a question. I do not know the answer.

Alan Greenburg: Same answer as I gave to the previous one.

Margie Milam: Yes, I don't have any answer to that. Maybe the new TLDs have like a – with this new designation has more value. Maybe more security. Yes. It's an open – it's certainly an open issue.

Unidentified Participant: I'm sure there are people who pay for dot TV and pay that reluctantly.

Alan Greenburg: That doesn't stop us from saying in the application you must specify what your price cap is. And it can vary from gTLD to gTLD. But we didn't do that either.

Margie Milam: No, we didn't do that.

All right. Thank you very much, everybody.

Alan Greenburg: Thank you, Margie.

—
Some of the comments labeled "Unidentified Participant" was actually yours truly: the interjection on .ag, for example.

contributed by carlton.samuels@uwimona.edu.jm on 2009-11-24 14:30:13 GMT