*** Disclosure: The following is the output resulting from transcribing an audio file into a word/text document. Although the transcription is largely accurate, in some cases may be incomplete or inaccurate due to inaudible passages and grammatical corrections. It is posted as an aid to the original audio file, but should not be treated as an authoritative record.*** Workshop on Metrics for Competition, consumer Trust and Consumer Choice Wednesday, 22 June 2011 ICANN Meeting - Singapore >>MARGIE MILAM: Can we get started, please? We ran over time on the last session. Our moderator is Rosemary Sinclair. And the topic for this session is workshop on metric, competition, consumer trust, and consumer choice, and I'm going to hand it over to Rosemary. >>ROSEMARY SINCLAIR: Thanks, Margie. Who's going to click the -- thanks, everybody. My name is Rosemary Sinclair, and I'm a counselor for the non-commercial stakeholder group on the GNSO. Thanks for coming to this workshop. We've got from now until -- our time is pretty tight, because this is going to use the GNSO Council meeting, open meeting after this workshop. Our topic is a workshop on metrics for competition, consumer trust, and consumer choice. And I thought I'll run through a few slides just to give us a background as to why we're all here. I note that there are some people who have been part of this conversation for a while who are sitting in the audience rather than up here on the podium. And they are very welcome to come up here. This is one of the kind of least conducive workshop environments that we've been in. But it's because the room is to be used for the council meeting afterwards. So anybody who feels they can't be part of the discussion properly unless they're up here with us should absolutely feel free to join us. Now, just to get back on to the workshop and the goals for today's workshop, I thought they would be best understood if I go through the big pieces that have been really linking our thinking so far. And, first of all, the Affirmation of Commitments is the document I want to look at now. And you'll see the third point here talks about a commitment to promoting competition, consumer trust, and consumer choice in the DNS marketplace. So that's our hook in the Affirmation of Commitments in general terms. We go from that general statement to quite a specific point in 9.3 about promoting competition, consumer trust, and consumer choice. And, in particular, dot point 2, if and when new gTLDs have been in operation for one year, ICANN will organize a review that will examine the extent to which the introduction or expansion of gTLDs has promoted competition, consumer trust, and consumer choice. As well as effectiveness of, A, the evaluation and protection process and, B, safeguards put in place to mitigate issues involved in the introduction or expansion. The Affirmation of Commitments document then goes on to talk about the review, and I won't go through that. But it's in this slide pack for sake of completeness. Second major document is ICANN's own strategic plan. And we find that in that strategic plan now one of the four focus areas is an area described as competition, consumer trust, and consumer choice. So the work that we do in understanding these terms for the one purpose, the AoC purpose, needs to be coherent and consistent with the work perhaps for another purpose. And that is further work on the strategic plan in coming years. In the strategic plan, the focus area competition, consumer trust and consumer choice is expanded, strategic objectives, strategic projects, community work, and staff work. And I included this because I thought at some point in the future it may guide our work on measuring metrics, targets, and the like. Our earlier discussions were two workshops, one in Brussels in June of 2010 where we started to talk about developing a consumer agenda for ICANN. And we looked at experience outside of ICANN -- OECD, ISOC, and other organizations -- plus started to discuss ICANN's agenda. And we followed that up with a workshop in Cartagena in December 2010 where we started to look more specifically at ICANN's addressing implementation of its consumer commitments. So then in January -- in December, rather -- the board made a resolution. And I see -- did you want to say something now, Steve? >>STEVE delBIANCO: Rosemary, if you don't mind. >>ROSEMARY SINCLAIR: Sure. >>STEVE delBIANCO: Just wanted to make an observation. Steve delBianco with NetChoice and a member of the business constituency. And I was fortunate enough to join the workshop you spoke of in Cartagena. And it was a fascinating dual-track workshop. And there was universal agreement that the need to honor these commitments to consumers, but there was a genuine parallel track between creating an institution, like a consumer constituency, and institutionalizing consumer concerns in all of ICANN. And they're not mutually exclusive. But it was clear that when a board member, Bruce Tonkin, was in the meeting, whatever merits there are to creating and launching the consumer constituency, Bruce was tightly focused on the need, as you're about to read, the need to parameterize the affirmation reviews and give real meaning and definition to those terms. Because I remember him saying -- I wish Bruce was in the room. I could quote what he said during a workshop. He said under the affirmation, we are required to do a review of competition, consumer trust, and consumer choice. And this is Bruce's quote, "My fear is that the way we're doing the reviews is, basically, collecting a bunch of opinions. Because you we don't have any measures or even any strategy in the area, so the opinions are going to be very diverse." And he closed by saying that, "Unless we've got measures to review ourselves against, I can't see how the reviews are going to be useful other than as a bunch of opinions." And that led, two days later, to the so-called Zuck -- Jonathan Zuck resolution who has been harping on metrics for years. And I think that's where you're about to go next. >>ROSEMARY SINCLAIR: Thanks for that background. I should have invited anyone who wants to make a comment, even during this background piece, to do that. So then, as Steve said, we get to this board resolution, which talks about the area of consumer choice, competition, and innovation at that time being one of the strategic areas for ICANN in the 2010- 13 strategy plan. That's now been changed to consumer choice, competition, and consumer trust in the 2011-14 strategy plan. "Whereas, ICANN has committed to promoting competition, consumer trust, and consumer choice in the AoC. And the third "whereas" talks about the introduction of the new gTLDs. And then we go to the resolution, which says that the ICANN board requests advice from the ALAC, the GAC, GNSO, and ccNSO on establishing the definition, measures, and three-year targets for those measures for competition, consumer trust, and consumer choice in the context of the domain name system. Such advice -- and I think we were being very optimistic at this time -- to be provided at the San Francisco meeting in March. That just turned out to be not possible given the workload of the GNSO in particular at that time. But, to prepare for this review, this workshop is meeting today to start the discussion on definitions, measures, and maybe a process for targets for the three areas -- competition, consumer trust, and consumer choice -- absolutely in the context of the domain name system. It's very easy to get into all sorts of other contexts when you're discussing these particular topics. So we need to be very mindful that our context is the DNS. So, with that background, we've got a panel, which is a loosely formed panel because we've actually got one less member and a couple more members. And that's terrific. Everybody should feel free to start the discussion. I know the business consistency has already done some work on trying to develop definitions for these terms. So I thought, with everybody's indulgence, we might let Steve start the discussion by taking us through the work that the B.C. has done. Thank you. >>STEVE delBIANCO: Thank you, Rosemary. And the B.C. took a stab at this draft, because it's something we deeply care about. B.C. is often abbreviated as B.C., but its official name is the commercial and business user constituency where business users are both the registrants and the users of the business system. So we felt that consumer interests were always at the core of what the B.C. was all about. And we said let's get out in front of this, if we can, and put a strawman out here. We'll take criticism and we'll take suggestions, and maybe we need to throw it away. But, Rosemary, the resolution that Bruce Tonkin and the board adopted, the Zuck resolution, asked to establish these resolutions and metrics. And the review, 9.3, that is going to be conducted is going to be a full year after the new TLDs are out. And the Affirmation of Commitments indicates that the review has to do two things. It has to look at the application and evaluation process -- and this is what you had up on an earlier slide -- and also the safeguards that are put in place to mitigate issues involved in the expansion. So, with that in mind, the B.C. tried to propose three definitions. And, if you don't mind, I'll walk through those quickly. And then we can debate whether they're the right words. And then we can talk about whether we can attack and hang metrics and measures on these. But the first is competition. We suggested it is the availability of multiple suitable TLDs and multiple registrars where a registrant could seek their desired name at a reasonable price and terms. So, again, the focus is on the registrant here. Because registrants and users are what we think of when we think of consumers, consumers of the DNS management that ICANN is responsible for. And the availability of multiple suitable TLDs -- let me just try to use an example. Suppose we have a Cairo hotel. Today they could be in dot EG or might be in dot com or might be in dot travel. Is that multiple suitable TLDs for a Cairo-based hotel. Probably. The competition would be the introduction of additional suitable TLDs offered by multiple registrars for that Cairo hotel. So they might find dot hotel in new gTLDs both in IDNs and in Latin. There might be a dot Cairo ccTLD that would be another suitable place. Also dot Cairo might be both in IDN and Latin scripts. And that would be adding the additional multiple suitable TLDs. And registrars would offer those out. So the idea, then, is to evaluate whether we have reasonable prices and terms. And, if you were going to hang metrics on something like this definition, assuming the definition survives even 10 minutes of the scrutiny of this panel, the metrics would be things like look at the supply and the prices. You could measure the number of TLDs and the average prices. >>Maybe if I stand up, I'll be able to see. >>STEVE delBIANCO: You could measure the number of TLDs and the prices in every script and language both before and after, before and after the introduction of TLDs. And you could look at barriers to entry for new registries that could offer that. So, Rosemary, do you want to stop there and take a look at a genuine debate on a proposed definition? Is that how you'd like to do it? Take them one at a time? >>ROSEMARY SINCLAIR: That's very difficult for me to see who wants to comment. So perhaps, if we do stop at the end of each point and take the comments as we go along. Jonathan, Tim, and then Werner. Sorry. I can't even see the right hands. Wendy. >>WENDY SELTZER: Thank you. Wendy Seltzer here, non-commercial. I apologize for interrupting. I was just thanking Olivier for acting as relay of the hands from the stage back to the other end of the table. Yes, I wanted to -- I think it's helpful to go through these point by point. Because several of the terms in here are confusing. For example, what's a suitable TLD? I would strike that. >>STEVE delBIANCO: May I reply? Suitable would be the notion that some of the new TLDs that we expect would be targeted to a particular community. A dot hotel is suitable for hotels and the travel industry. A dot bank is suitable for banks. Another indication of suitability is to be in the scripts and languages for the targeted community. So community-based TLDs are, by definition, designed to be suitable for a certain class of registrants and users. That's what I mean by "suitable." Is that a definition that makes more sense to you, Wendy? >>WENDY SELTZER: It's not. I think we're not going to get through this panel by engaging in prolonged debate over each of the subjects. But from my perspective, because I don't agree on sort of purposes and semantics specifications for TLDs, my definition would tend to strike many of those qualifiers. >>JONATHAN ROBINSON: Rosemary, a couple of quick comments, if I may. First of all, to compliment Steve and I want to get the full commercial and business user's constituency for their good work, as you said, in getting out ahead and creating a strawman. And I think we've actually circulated this. So, thinking about the registries constituency we've circulated this in this constituency and suggested we use this as a strawman and come back on try to develop some of these ideas. For this workshop, the registries constituency is slightly behind on that, so thinking in that way, as a registries constituency rep. That said, I don't see that's any reason why in this workshop we can't make progress. And I'm not suggesting we shouldn't. And a key point of progress for me, ultimately, will be reflecting back on your comments -- or was it Steve? I think it was Steve, actually, on what Bruce had said. And for me, critical to this, will be measurability, quantification, and, you know, effective measures both now and in the future will be key to all of this. So sorry to interrupt you, Steve. But that's just a couple of key context points mostly in response to what you've said so far. >>ROSEMARY SINCLAIR: Okay. I'm developing quite a long queue. So I think I'll go the floor, the table, the floor, the table. Werner, I think you were the first next on the floor. >>WERNER STAUB: Thank you. Just the first question. Competition. Do you include diversity in competition, or is that separate? >>STEVE delBIANCO: I think the choice -- again, the affirmation had the word "competition" and the word "choice" in it. So perhaps diversity is something you suggest in the "choice" category. And I don't want to jump ahead, but maybe that's where you talk about the availability down there under choice for the purpose and integrity. So it could be competition, or it could be choice. >>WERNER STAUB: The thing is if it's choice, I have a problem with it also being seen as perspective of consumer. And then the question is who is that consumer. So it will not be restricted from a consumer perspective diversity. But, anyway, if you look at competition, I feel a little bit the same way as what Wendy said. Is it the TLDs? Is it the registrars? What about the resellers and the resellers of the resellers? And do we just count units and somewhere? Or in the -- do we just look at what's the DNS as an industry providing if there's competition inside of that industry? Or do we look at the fact that DNS itself has been and has the vocation to be a mechanism to leverage competition. Without domain names there would be a lack of competition in so many activities, because it would be -- there will be a bottleneck with respect to naming. We still have that in many ways. So what about the competition between potential users of domain name holders. If you don't have access to reasonable domain name, the other ones having been taken, you're not going to be part of that market. So there's very much the competition that happens at another level that we probably are going to be concerned with. >>ROSEMARY SINCLAIR: Thank you. Tim? >>TIM RUIZ: Yeah. My comments or suggestions about the competition definition I guess would be around the -- I guess there's like the missing piece or something in there that multiple suitable TLDs, but that those TLDs are across multiple registry operators, perhaps even multiple registry back-end operators. Because I think, when you get into -- right, the diversity of suppliers. Great way to put it, Steve. Because I think that is really where -- or at least I think that's an important aspect of really determining whether there's competition in the marketplace or not. >>ROSEMARY SINCLAIR: Okay. Philip? >>PHILIP SHEPPARD: Thanks. Philip Sheppard, B.C. and also one of the contributors to the B.C. paper. It's always difficult, when you launch straight into an end point of a definition, rather than seeing all the intelligent script that goes before that. And I suspect that perhaps the entire panel hasn't read all of the B.C. paper. I would recommend it to you. It's not terribly long. The way we got to the word "suitable" was really a paraphrase for what in competition law terms is known as relevant market. And we started off in the paper talking about two of the three key principles of competition law, which we think are most relevant here, and want to start off with considerations of dominance. In order to define dominance, that leads you on to wanting to define the relevant market in which dominance may be taking place. And relevance is what we are seeing here paraphrased as suitable. So it doesn't come out of the blue. It comes out of the established principles of competition law. It is difficult for us to decide. If you want a lawyer's definition, you can have one. If we want a more easy-to-digest definition, that may be something more suitable for the board or community here. But shall I say every word in this definition doesn't come just out of the blue. So we need to think about it in the context for which we -- how we got there. Thank you. >>ROSEMARY SINCLAIR: Evan. And then Carlos. >>EVAN LEIBOVITCH: Hi there. For the record, I'm Evan Leibovitch. I'm vice chair of ALAC but speaking in my own personal capacity. I've had a real difficulty getting my head around these concepts, in part because they dance back and forth between consumers as end users and consumers as registrants. And we dance back and forth between them to suit matters of convenience. And I think this is something that is a real difficulty, even in the issue of competition. The competition in the ICANN case has very little to do with end users because the competition for end users is that of the vendors themselves, those that are supplying, goods, services, and information to end users, not supplying domain names to end users, per se. And, in fact, if there's a consumer trust issue, it's not necessarily because there's 100 TLDs, 1,000 TLDs or a million or a billion second-level domains, but whether or not there's enough confidence in the domain system that people know they can use a domain name to find what they want as opposed to just bypassing the entire mechanism and using a search engine to find what they want. I really think that we're missing the end user driven issues of consumer needs, consumer choice, by limiting this to the ICANN component to the food chain which ends at the registrant. >>ROSEMARY SINCLAIR: Carlos. >>CARLOS DIONISIO AGUIRRE: Thank you, Rosemary. For the record, Carlos Dionisio Aguirre, NomCom appointee, GNSO. But for the last four years, I was ALAC member, so I attempt to talk in my personal capacity as Evan, in my personal capacity as an end user. In this order, I ratify all that Steven said, because it's very important. I'm worried also in the definition of "competition" how we are capable to make reality the measuring of this competition and who decides if the competition is properly or is needed more or less competition. Who is the measure, the exact measure, the proper measure, about competition? Also, I want to make mention about what Werner said, in relation with the different communities, linguistic and regional communities, because the competition is a very complex point to define. So this is my worries, and I think it's needed to complete the definition of "competition." There are here three different concepts and points very, very important, because make -- or are very important to reach a healthy Internet governance ecosystem, and when we analyze the three concepts, we can see they are not separate concepts but I believe in order to accommodate our -- our participation or make better our work to consider one-by-one. First, competition and then the others. Thank you. >>MARY SINCLAIR: We'll go to the floor, and then Beau, and then I think we'll close the queue on this one so we can have a go at getting to the next point. Thank you. >>JONATHON ZUCK: : It might be difficult to say something new but I think there's two questions, and one is that -- Phil raised, is the relevant market for this and what is the market that ICANN can reasonably be expected to try and influence or manage. Is it the market for domains or is it the market for hotels in Cairo? And I think that's going to be a very difficult imposition to put on ICANN that they are somehow in the job of trying to promote competition between the hotels in Cairo. But I think in either definition -- and so that's something that has to simply be resolved. That decision has to be made or this conversation won't go anywhere. But I think for real competition and all the promise that's been made about this explosion of new TLDs, it's going to have to really focus on behavior, and the mere availability of TLDs, I think, is not going to represent competition. There was talk about innovation. You know, is dot shop going to do something sufficiently innovative in the way that they manage their domain as opposed to simply providing a bunch of new names that people could use instead, so that there's a reason to use that domain instead of an existing one. What percentage, for example, of the new registrations are new Web sites instead of simply redirects to old ones? So if the hotel in Cairo simply means that these new TLDs simply represent an opportunity for somebody to tread on their brand and so they just end up buying-Erik them defensively and redirecting them to their site, that's something that's measurable. If you'll these new registrations are redirects -- right? -- then we haven't really created competition even in the TLD market. So I think behavior and how these TLDs are used and how that uptake takes place in the market is how we can really create some measures for competition, and I think the mere availability of new name TLDs is really going to be irrelevant. If all this program results in, I think it will be a failure. >>MARY SINCLAIR: (Speaker is off microphone). >>JONATHON ZUCK: I'm Jonathon Zuck. Sorry. The one the resolution was named after. >>MARY SINCLAIR: Beau. >>BEAU BRENDLER: Yeah, Beau Brendler of North America. Forgive me for coming in late from an ALAC meeting. I mean, is each panelist going to say something, in which case I'll just delay what I'm going to say until, you know -- >>MARY SINCLAIR: You've got the floor now, so why don't you just say what you're going to say. >>BEAU BRENDLER: Well, it just had to do with competition, and it, in essence, harmonizes with what both Evan and Jonathon said, in the sense that, you know, there are -- there are competitive metrics that I think in some particular circumstances we're just not seeing because it all goes back to the very confusing difficulty of defining the word "consumer," and until a decision is basically made on what that definition is going to be and it's embraced throughout the organization, it will be very difficult to move forward. But, you know, for instance, there is -- ICANN has a list of -- or has a stable of ICANN-accredited registrars. That's a metric in and of itself. There's also an analysis that could be done of which ICANN accredited registrars are actually, you know, saying that they are but in fact don't meet the specifications. That's a metric. The kinds of metrics that consumers, in my definition of consumers, are looking for are already here. You know, DNS abuse is a metric. You can look at DNS abuse over time in whatever particular axes you want, and that's a metric by which to measure consumer trust or success for ICANN addressing consumer issues. >>MARY SINCLAIR: Thanks, Beau. And now I'll ask Jo Lim. Jo, you just might just introduce yourself. I know who you are and thank you so much for coming to be part of our panel. >>JO LIM: Sure, Rosemary. Yes, for the record my name is Jo Lim and I represent dot au. I work for the dot au domain administrator and Rosemary asked me to come along today just to bring the dot au experience to the table. I must say I'm new to this workshop and I'm new to this particular part of ICANN's constituency group, so I apologize for lack of background there and I hope my question isn't too stupid, given that. But in terms of competition and what level of competition we're assessing here, particularly in relation to the relevant market, is there consideration of competition for applicants for new gTLDs in terms of the registry operators that they are able to select? So competition between registry operators for providing service for new - - to new gTLD applicants? Is that part of this consideration or am I way off base? >>STEVE DelBIANCO: The Affirmation of Commitments paragraph says that a year after launch of the gTLDs, some other review team is going to get appointed by the GAC chair and the board chair, and I doubt any of us will be involved at that point, but they'll look at two things. They'll look at the application evaluation process, which is what you just asked about, and they'll look at safeguards that were put in place to mitigate issues involved in the expansion. So the first review under 9.3 will look at the application process, and then every three years after that, subsequent reviews may or may not have any application processes to look at. >>JO LIM: So is it looking at the application process from the point of view of the number of registry operators that have been involved in setting up new gTLDs? So, I mean, for example, if all new TLD applicants are all choosing one particular registry operator, there's obviously not a lot of competition happening at that level. And similarly, is there -- are they also considering whether the new gTLD process has actually allowed new entrants into the registry market, so that it's not just existing registry operators who are being selected by applicants? >>STEVE DelBIANCO: And when you say "they" select, the parties who negotiated the affirmation -- Commerce Department and ICANN board -- and they signed it. Other governments may sign the affirmation, and I hope they do someday, but that's it for them and "they." From then on, it's all about us. As you know, all of the review teams end up being composed of community members who then conduct the review. And Bruce Tonkin's endeavor was to define metrics before we begin the new gTLD evaluation process so that we can manage to do well at meeting those metrics, knowing that later on the evaluation's going to occur. >>JO LIM: Yes. So sorry, so just to clarify what I'm saying is: Has that been considered as a possible metric -- >>MARY SINCLAIR: We're just really at the early stages, so the contribution is great, Jo, because it expands -- >>JO LIM: (Speaker is off microphone). >>MARY SINCLAIR: Yeah. No, no, no, no, no. It expands our thinking. Jonathon, 10 seconds and then Philip 10 seconds. >>JONATHON ROBINSON: Yeah. Real brief, I'm Jonathon Robinson and I'd just like to endorse what Jonathon Zuck said. I think in the end the litmus test is out in the market, and Jonathon expressed it very well, so I just wanted to give some support to his point. >>MARY SINCLAIR: Philip, last one. >>PHILIP SHEPPARD: Thanks. What I love about trying to structure a conversation, a panel conversation, is that people always leap ahead which is, of course, what we're doing here. And some of those questions are actually answered in our suggested metrics here, in terms of the measures, because for this particular thing on competition, we suggest three things, one of which is value- added, which goes back to Jonathon's point in terms of innovation. Is it really differentiating? The other is barriers and that's all to do with a plural supply base. And the third, of course, is dominance, is opening up something to dominance there. So these ideas, in fact, are all captured. So the answer to most questions so far is yes. It's just we haven't got to that bit of the dialogue yet. >>MARY SINCLAIR: Okay. Let's have a go at consumer trust. >>STEVE DelBIANCO: If it's all right, we'll skip to choice, because trust is a little bit of a different animal, but choice and competition are so very similar, and we didn't write the Affirmation of Commitments, we're in charge of fulfilling it. And it made a distinction between competition and consumer choice. So we have to figure out the supplier side is where the competition comes in, and it mostly addresses registrants, but consumer choice seems to also affect registrants but also end users. So if choice -- we'll know choice when we see it, but it would look like an increased availability of TLDs that have competing propositions, differentiation, Philip Sheppard just said, competing propositions as to the purpose and integrity of the domain name registrants. And again, the example of something like a dot bank or a dot food, dot hotel, things that have special purpose. Some may not have special purpose at all. A truly generic gTLD like dot Web isn't going to have a special purpose and integrity, but it does bring some choice to the table because it would allow a whole new body of second- level domains to be registered. But the idea that I think it was Jonathon Zuck had said that you have to measure this not just by the numbers, but by the behavior, but if the behavior follows the new gTLD program, well, then the numbers would show that there is a greater availability of TLDs that have competing propositions. They're not just dot com clones, but, in fact, competing propositions that serve special linguistic communities, for instance. They might serve cultural groups. What does Bertrand de la Chapelle call them? Categories. They might serve multiple categories who don't get a lot of satisfaction from the current TLD space. So that definition of choice there closely matches competition, but it serves not just registrants but users, because when a user gets into a search engine and looks for hotels in Cairo or looks for banks in Singapore -- and I go into Google and do a search -- the results that come up will have URLs next to them. He'll have domain names next to them. And I don't know, Evan, you were suggesting -- maybe I misunderstood you -- that the use of a search, in general, bypasses the DNS, but not at all. The Google search engine is going to display for me a whole list of results and next to the results I'll say that ooooh, here's a bank that ends in dot com, here's a bank that ends in dot ru, and here's a bank that ends in dot bank. And if dot bank has done its job and built the consumer trust, then I now have a choice of the results that I pick and the banks that I use and registrants have a greater choice to decide to go to a dot bank versus a dot com, dot net, dot ru. So with that, let's attack this definition. >>MARY SINCLAIR: Thanks, Steve. I've got Evan first up and then Tim and Beau. Okay. And Wendy. Good. Evan? >>EVAN LEIBOVITCH: I guess I'll have to agree to disagree with that particular analysis. If you're going in and doing a Google search, the domain name is the real small print at the bottom and it's the name of the organization at the top that people are actually selecting, not the domain name. So it's still a matter of bypassing the DNS if people are going to organizations based on the name of the company as opposed to the name of the domain. But the other thing -- and this goes to the subtlety of the distinction between consumer choice and competition -- competition is a supplier-eye view of the marketplace, and is driven by suppliers and can be measured through what the suppliers are doing. Consumer choice is a bottom-up approach, and in order to have -- apply the metrics on whether or not there's sufficient choice, you can't ask the suppliers whether or not the consumers perceive there to be enough choice. That's a question that has to be posed from consumers, and that, in fact, requires different questions but also asking it of very, very different bodies. One is a supplier perspective; the other one is a consumer perspective. >>MARY SINCLAIR: Thanks, Evan. Tim? >>TIM RUIZ: Actually, just a question, and I just might be confused by the wording of the definition, but it says, you know, "Availability of TLDs that offer competing propositions as to the purpose and integrity of their domain name registrants," so is that meant to imply that the -- you know, that the registrant's Web sites are involved in this particular measure, as far as whether or not, within that TLD, the registrants are offering choice? I guess I'm a little confused by that. >>STEVE DelBIANCO: If a particular gTLD applicant proposes a purpose, it will be baked into their registry contract, and part of compliance is whether or not they honor that contract at screening the actual registrants who are allow to buy in the domain. For instance, dot bank would have in its value proposition the use of second-level domain names that satisfy whatever criteria they put out and the degree to which a registry operator fulfills the commitments they made in those contracts will determine whether those are effective choices or not. If -- if a domain gets -- starts and has a tough time -- if a top- level domain has a tough time making it financially and seeks to change its promise and open up to registrations from anyone, well, then we no longer have a choice of competing propositions. It's another clone for a completely generic domain name. >>MARY SINCLAIR: Next in the queue was Beau. >>BEAU BRENDLER: I was looking at the definition there of -- I thought we were on consumer trust. We seem to have gone to consumer choice but I'll go back to consumer trust because it just happens to be second on the list. I like the definition proposed by the BC, in a sense, because I think it gets at the heart of the matter, the heart of the controversy, because it's really going to be a situation of trademark issues, you know. But my concern with it is not that it was intended for the general public to read, but language like, "The perceived integrity of domain name registrations such that Internet users have confidence that a domain name is held to the advertised purpose." I mean, that's just simply not going to be in the realm of anything that a regular human being is going to pay attention to. So the biggest issues related to consumer trust, when even thinking about domain names, assuming consumers even think in terms of what a domain name is, is going to be when I am on -- to put it in the realm of new gTLDs -- when I am on dot bank, am I on, you know, peoples.bank? Am I on peoplesbank.com? Is my information going to be stolen? I mean, you know, we've got to have more of a discussion in the ICANN community that's at that -- I'm not saying we have to have that at that left of simplicity, but that's -- when we talk about consumers, that's our mind, in my opinion, needs to be. >>MARY SINCLAIR: Sorry, Beau. Just for clarity for the record, that was a comment around consumer trust? >>BEAU BRENDLER: Yes. >>MARY SINCLAIR: Thank you. Okay. Next in the queue was Wendy. >>WENDY SELTZER: Thank you, Rosemary. Now, when -- in early discussions in this workshop, you had circulated some definitions from the strategic plan, and I thought it would be helpful to add that to the record here, their just a definition of consumer choice. It says, "Consumer choice includes, but is not limited to, the concept that users can access unique identifiers in their own languages and language scripts." And I think that has a couple of key points here. First, we're focusing on unique identifiers as the focus of our definitions and our program here. And then we're not trying to plan what it is that people may be trying to choose. When we're doing bottom-up view, it really is up to the end user to make the determination what he or she wants to choose, and so a minimal definition that asks the end user to make those choices for him- or herself seems, to me, the optimal. >>MARY SINCLAIR: Thanks, Wendy. Now, our colleague from the floor. >>TOBIAS MAHLER: : Hello. This is Tobias Mahler from the Norwegian Research Center for Computers and Law. I find these definitions rather interesting and I would encourage you to look at particularly the first and the last definition in conjunction, because they are really very similar, and I think to a certain degree, you're defining one through the other, because if you read your definition, you say that competition is the availability of choice, so basically you're saying that the first is the last and consumer choice is competing propositions. So that would be competition. So I would encourage you to separate these two concepts a bit further. So in my view, much of the existing text competition could actually go better with the consumer choice, so these could be some of the metrics that could be used in the context of consumer choice. What seems to be lacking from the aspect of competition is a bit more of rivalry. As you have it right now, read through the definition and think of a system where there's no competition at all. Think of a socialist system. You have a shop that offers three types of bread, and that's it, and the state mandates that each bread costs 10 whatever, and that would be -- that would fit within your definition. It's the availability of multiple breads at reasonable prices. And what is reasonable is determined by the state. So that can't really be all that is to competition. So I would rather put elements of that into -- into the last one. And then regarding the consumer choice, for me the key aspects are to define what you mean by "consumer" exactly, and maybe there are different types of consumers you have to look at. And that will then also help you to define further what the consumer might choose amongst. So different groups of consumers might have to have different choices, and that can then be further examined. Thank you. >>MARY SINCLAIR: Thanks, Tobias. I've got Jonathon, Jo, Steve, and Philip in the queue and I think we might close it off after Philip and then get on to consumer trust. So Jonathon? >>JONATHON ROBINSON: Thanks, Rosemary. I hope I can frame this point effectively, because it's more some thoughts I've got in and around this. One of the things that struck me -- and I realize I might be even stepping on interesting territory here -- is that from a domain name registrant's point of view, domain names today are probably dominated by business or business-oriented users rather than end users. And one metric by which you might judge that is that domain names can be measured in the hundreds of millions, the numbers registered, whereas, you know, Internet users can be measured in the billions. So I guess when we measure this in the future, we've got to be careful we don't sit within a very small paradigm and say, "Right, well, actually we've got lots more choice," but we've also got to be open to some kind of fundamental market shift that may and could take place -- and in some people's view, might be highly desirable to take place -- whereby consumers in their billions -- in other words, real end user individuals -- take up domain names for one or more reason, but of course just to continue that train of thought one more step, and then I'll -- I'll close, is that that could lead to problems of its own if it were so successful that end user consumers were, in effect, in a -- in one walled garden. So, you know, we could say, "This is fantastic. We've got now billions. We must have more choice." But actually everyone's gone into -- so it's -- there's some challenging things and that's why I say it's not as well framed as I might like, but just to give you a bit of a train of thought in that respect. >>MARY SINCLAIR: Jo? >>JO LIM: Thanks, Rosemary. I just wanted to pick up on comments about the integrity part of that definition and what was said earlier about commitments made by registry operators in the way that they run their space. And also the comment made earlier about what type of consumers are we talking about. And I was sitting in the UDRP session earlier this morning in this room, and it strikes me that a type of consumer is a complainant or someone with a grievance about a domain name registration. And if you look at it, you know, in terms of choice, what choice do they have in terms of the process that they follow to pursue their grievance or pursue their complaint, and just as an example from the dot au space, we have a modified version of UDRP, the AUDRP, which people can use. Because we are a very policy-heavy space and we have strict eligibility criteria, someone can also complain to the regulator about a breach of policy and that may result in the domain name being deleted. So that's another process they can use. Or they can, of course, choose to go to -- through legal processes, so go to court, litigation. So I wonder if that's sort of another metric in terms of choice available to a particular type of consumer with respect to new TLDs. >>MARY SINCLAIR: At a different point along the kind of supply chain, uh-huh. I had you, Steve, next and then Philip. >>STEVE DelBIANCO: Yeah. Before we jump to trust, we probably ought to start thinking about the kind of metrics and measures that would be appropriate to something like choice. Again, the availability or numbers that one can measure both before and after the new gTLD program. So a year after launch, you measure how many you have in Arabic script, how many TLDs in Arabic and if the number is much greater -- and we certainly hope it is -- that would be an indicator that we created more choice. We might look at industries. How many industries got their own TLDs? That would be another indicator choice and you'd expect those numbers to go up. But if the new TLDs that I've just spoken of are mostly dominated by defensive registrations, that would be choice where the behavior didn't actually generate as much benefit as the promise implied. So there might be measures of the percentage of registrations that are defensive, assuming we could even define what we know to be a defensive registration that doesn't seek to get additional traffic but just simply defend a name. So can we talk a little bit about metrics? Does anybody have any ideas on that? >>MARY SINCLAIR: Could we come back to that in just a minute? Philip? >>PHILIP SHEPPARD: Thanks. I'd just like to respond to the observations we heard from Tobias, which I thought were excellent, because he hit on I think a slight disconnect between how a competition law might approach the subject and how the diversity of the ICANN board has approached the subject. They're looking at these three topics, these three concepts -- competition, trust and choice -- as separable, definable things. Now, within the context of competition law, when we look to this in the background to this paper, of course, competition law typically does not try to define choice as an objective or a policy. It's merely the -- as Tobias was pointing out to us, it's merely the positive consequence of a rivalry between competitors. And then we ask ourselves, okay, so what are the fundamental determinants of choice, and you got two things, typically. You got the availability of competing differentiated offers and you've got the freedom to exercise the choice and those then sort of fall down. But once you go through that logic, it leads you inevitably to the circular definition that you see here, which plays back to the other two. So it's not circular for a bad reason, it's circular for, we hope, the right reason because it plays very much going back into the other two and that's how we got there. >>MARY SINCLAIR: Thanks, Philip. I think that we might go to the consumer trust point because we then need to talk a little about metrics and then targets, if we can, trying to finish, I think, just a little before 2:00 so that the council meeting can be set up. So back to you, Steve. >>STEVE DelBIANCO: Thanks. Consumer trust, the middle definition, is much more challenging to measure, since it's based on perceptions. We've tried to suggest that it's the perceived integrity of the domain name registrations such that users -- focusing more here on users than on registrants -- have confidence that a domain name is held to the advertised purpose and standards of the TLD operator, of ICANN, and of relevant law. So if you have TLDs that are completely general and generic, like a dot Web, it may not have an advertised purpose but it certainly has standards that are upheld by ICANN as well as relevant law, and the metric for that could be something like measuring the number of violations of the registry contract and the names they've sold over time or the number of violations of applicable law, the number of UDRP and URS proceedings and measuring those incidents as a percentage of the domain names in the new space and comparing it to what it is in the current space would give you some before and after metrics to evaluate whether we've done the gTLD program right in fulfillment of the Affirmation of Commitments review. And turning to surveys, the perception of consumers is important, and the current WHOIS review team just received funding from ICANN to do a survey of consumers specifically to look at, quote, the level that they trust the Internet and specifically ICANN's role in establishing that trust. So I'd be hopeful that if their survey goes well, we would log the results -- it will be later this year -- and perhaps repeat the survey one year out, after the new launch, during the affirmation review, so you'd have two statistics on trust. The percentage of trust in 2011 versus the trust in 2013 as a result of the new program. >>ROSEMARY SINCLAIR: Any comments? Perhaps if I start with Debbie and then Beau and then Wendy and Werner. >>DEBRA HUGHES: Thank you, Rosemary. One of the things I would like to add or consider is security and stability of these new TLDs that are going to be added hopefully as a result of competition and choice; that these new TLDs would instill in consumers the assurances that the information and activities that are being transacted are safe, secure; and that consumers, when they visit these new registrations that are resulting from these new registries are confident that the transaction is what it is. The other thing I would add, too, is that we might want to consider -- I'm looking at the definition. I don't know how you would add this in, but making sure measuring from how the ability of these TLDs and their success -- I mean, did any of these TLDs within the first year or so of their operation fail? Or were there significant breaches or things like that that we could try to measure that would be quantifiable to define consumer trust? >>ROSEMARY SINCLAIR: Thanks, Debbie. Beau? >>BEAU BRENDLER: Yeah, this comment is really winds up being sort of in partial agreement to things both Debbie and Steve have said. I think you could put consumer trust and consumer choice in a fairly simple grid and have metrics on either axis. Basically, a consumer, again, defined as a regular human being who may or may not register domain names, when they think about choice, probably what they are going to think about is which registrar should I use to choose -- to register a domain with if I have never done one before. And if there were guidance material of, say, the top 25 registrars broken down by all of the different qualifications or whatever that we've just heard, you know: Are they an ICANN-accredited registrar? Are they really an ICANN-accredited registrar or just say they are? How many -- you know, the different types of other points of data that have just been mentioned could be put in there. Consumer trust and choice, if defined very specifically at that level of just a regular human being, doesn't really seem like that insurmountable of a definition or an operation to put together. >>ROSEMARY SINCLAIR: Thanks, Beau. Wendy? >>WENDY SELTZER: Thanks. Again, I want to return to the definition provided in the strategic plan. There it is consumer trust includes but is not limited to the concept that unique identifiers work all the time and deliver consistent results when used. And I would tend to limit the notion of trust to that consistency and stability of the unique identifier and its resolution. I've heard in other discussions around ICANN broader notions of trust in the Internet or trust in the entities using the Internet. And I think all ICANN can do is provide a framework for establishing that trust. So, for example, DNSSEC, the signing of DNS entries enables the entities to establish a trusted pathway by which the end user can verify the identity of the registrant or the Web site that he's visiting. But beyond that, I think there isn't meaning to a general trust in the Internet as a whole. I think we need to focus trust on particular functions. >>ROSEMARY SINCLAIR: Thanks. Werner was next. >>WERNER STAUB: Just a question, first. When you say that in the proposed revision we have perceived in there, is that the proposal or was that part of the task that we have? Does it have to be limited to "perceive"? >>STEVE DelBIANCO: What I'm putting here is the word "perceived integrity" is indicating from the standpoint of consumers we may have to measure trust by what their perceptions are. That opens the notion that we may have to survey. What Wendy talked about was being entirely quantitative. This would open the door to say that the metrics we would design include people's perceptions about what it means to have a trustworthy DNS. >>WERNER STAUB: Okay. But it would not exclude that we actually measure the level of trust that actually deserves to be given? >>STEVE DelBIANCO: Not deserves to be, that would be normative. It would be descriptive. We would measure it before and then measure it after the launch of the new gTLDs. And I was just explaining to the WHOIS team who is doing a measure of trust in a survey later this year as an example. >>WERNER STAUB: Because one of the things that worries me about it is that, of course, we need to measure what people feel. But they may be misled by what they see, and that's something that worries me quite a bit. I would actually like people to learn to distrust, to be careful. How can we teach them that it is not the same whether you are in a specifically controlled TLD that you can trust and one that you should actually be careful? So to know which part of town you are in at what time of day or night. >>ROSEMARY SINCLAIR: Thanks, Werner. Jonathan? >> JONATHON ZUCK: Thanks, Rosemary. >>ROSEMARY SINCLAIR: Sorry, the floor, Jonathon. >> JONATHON ZUCK: Either way. I think this is the perennial ICANN problem, which is that the road to consumer trust is not paved with good intentions. Over and over again we use our intentions in the fact we launched a new program as a proxy for results. This is why I've become such a harpy about metrics for so long. I think we really have to separate any kind of measure of consumer trust to try and build their trust. That's, in fact, why we are measuring it, just to see if those activities were, in fact, effective in building that trust. So if the goal as set out in this document is to build trust in the Internet, then we need to do things to build trust which may just involve education, as Werner mentioned. But then there have to be some measure of that trust to see if whether those of the activities were worth having done in the first place. Just doing them is such a perennial problem, if we don't measure the success of things we do, then we are just wasting our time. The survey that Steve mentioned, I don't know if the survey that's actually going in the field is going to be sufficient. We need to look at it in some detail. It is the delta that ends up being interesting. I think in some measure what we will be looking for in this rapid expansion of TLDs is whether there is a loss of trust as a result and trying to prevent that. That may end up being a big part of what we accomplish in the new gTLD program, is actually maintaining the current level of consumer trust rather than expanding it. I think one of the opportunities to expand consumer trust which speaks to your -- Evan's issue about search engines, et cetera, is if some of these domains are able to really establish enough of a brand and enough of a differentiation that people's gateway into the Internet is through, you know, some sort of a structured mechanism within that domain, then those domains may, in effect, build trust in themselves. And that would be very interesting to measure as well. But I think in both cases, we have to measure the trust and do things to try and build it and stop using our intentions as a proxy for results. >>ROSEMARY SINCLAIR: Thank you. Jo? >>JO LIM: Thanks, Rosemary. I just wanted to add the dot au experience. Trying to measure trust in the dot au space is something as we as the administrator take very seriously. And we've done market research, public surveys twice in the last -- over the last ten years and we're about to do another one. And what we always find difficult is the very, very lower level of awareness among even registrants, not just end users but even registrants about the DNS and what a domain is anyway. The average end user just goes to a Web site and doesn't necessarily think about the domain name behind that. So if you are talking about trust in a domain name space or a particular TLD, I think you're going to find that most end users, consumers won't even understand what you are talking about to even ask them that first question in the first place. So it is quite difficult to measure. Having said that, I can't see any other way that you can do it. You do have to go out to consumers directly and understand what their perceptions are. And I do have another point, but I've completely forgotten what it is now. Oh, no, I do. Sorry, Rosemary. So I got sidetracked because my first point was actually in terms of trust, the concepts that we at dot au think about when we try to measure what trust there is, is that we hope that there is trust in the policy rules around who can get a dot au domain name and how those rules are enforced. We also look at the way the industry is being regulated and the actions that the regulator is prepared to take in order to ensure that the industry is clean and is not misbehaving. And we have -- we've done that in a number of ways. We've actually litigated and different things to make sure that we can say that our industry is well behaved and well run. And then the last one would be the dispute resolution processes and how well they are working and are people satisfied with those. All of those things together, we feel, adds to the trust in our space. >>ROSEMARY SINCLAIR: Thanks, Jo. I have got Tim, Evan, Jonathan. And this Jonathon wants to make a quick response, is that right? Do that now and then I have got Philip. And then I think we will cut the queue there and see if we can get five minutes just for a general discussion on metrics and then a few minutes for general discussion on targets. Jonathon? >> JONATHON ZUCK: Just briefly, I really want to encourage everyone involved in this process not to make the perfect enemy of the good. Metrics are such a foreign concept to this organization. If we just do a little bit of them and decide a year later we chose the wrong ones, it will still be worth having done. We really need to start measuring the things that what we do. I think another broader issue that's really going to be a part of this debate is the interplay of search engines versus the domain system. I think that's something we will find ourselves discussing a lot because a lot of consumer mistrust is because they do a search and get a link to something that is not actually the URL of the bank they are searching for, for example. So the domain name system, if done right, might play a role in restoring that mechanism for entry into the Web and make -- and downplay the role the search engines will play, which is far too generic a way into the Web. So that should have -- be a separate discussion probably. >>ROSEMARY SINCLAIR: Tim? You okay? Evan? >>EVAN LEIBOVITCH: Actually, I wanted to essentially expand on that a little and also on the point that was made before as the complaint as an unit of metrics. One thing that has always been constant in At-Large meetings at ICANN meetings is having compliance people come to our meetings. And we're constantly asking: What are the metrics? How big is the queue of uninvestigated complaints? How many times have you come across times where people have had domains taken away from them and put up for auction? How many times do we have people thinking that they own a domain when, in fact, they don't? There are all sorts of things like this that have to do with complaints. How ICANN responds to them, how transparency -- how transparently they make the investigation, and how much they come forward is also a matter of building confidence. You can't educate confidence. You can't convince people to be confident when they come up with complaints and the complaints don't go investigated and the queue of compliance issues gets longer and longer and longer. There's some metrics that can happen right there and right now with what we have. >>ROSEMARY SINCLAIR: Thanks, Evan. Jonathan? >>JONATHAN ROBINSON: I will pass for now, Rosemary. >>ROSEMARY SINCLAIR: Thanks, Philip. >>PHILIP SHEPPARD: Thanks. I thought even the short debate has been excellent because I think that would encourage us in the BC to make some edits to the paper that we originally submitted because I think there are some great points being raised. One shortcut to us determining -- trying to define a definition of "trust" because people were in the conversation already here throwing in their own ideas and their own measures, maybe something as simple as saying that trust in the context of the DNS can be defined as something like the measures that are crucial to consumers in determining whether to trust or distrust a domain name system. And then all you have to do is have a debate on the list of measures, huh? And if you agree on a simple umbrella definition, it may stop us spending too long on one thing before we were to get to what's going to be crucial, which is the measures and then the even more challenging subject of targets. [ Applause ] >>ROSEMARY SINCLAIR: Okay, thank you for that. Steve. And I wonder, Steve, if you could segue us into the discussion of metrics by referring to your paper. >>STEVE DelBIANCO: And, Rosemary, when we have discussed each of these definitions, we have thrown out metrics and metrics like survey percentages. I think we have covered a lot of it since it really flows from the definition. To suggest a metric in an abstract suggests you probably need to define the thing you are measuring. And, therefore, you are having a definitional conversation. It might be more fruitful to think about where this process goes from here. The board resolution was passed in December. And it was handed directly to four different groups: Government Advisory Committee, ALAC, GNSO and ccNSO. And they were each asked for advice. And so this is one of these hybrid processes where each of these groups may go off on its own and come up with its own definitions and each sends a separate report back into the board in response to a resolution. And I don't exactly know what they will do with it. Alternatives are to attempt to be aspirational here and at least attempt to see if we can get a common set of definitions. Perhaps we go our own way with metrics. After you define metrics, one has to say what are the targets for that metric? It could be the ALAC would have much more aggressive targets on a particular statistic than might the GNSO. That would be an appropriate amount of diversity the board could resolve. The biggest challenge is how do we maintain momentum from this meeting without having three or four dueling parallel processes. I said three or four because I'm highly doubtful the GAC will engage and take advantage of this. They may, and they will do it on their own time frame. But we will do our best to keep them involved to come up with this. >>ROSEMARY SINCLAIR: Just on that, I had a very short conversation with Heather Dryden this morning to let her know that we were having this workshop and that we would let them know how far we got. And it was a short but positive conversation. And, of course, we have ALAC contributing thus far, so that's positive. And, well, I asked Jo really to come and give us her perspective from dot au. But, yes, that's a springboard to then go to ccNSO possibly now that we are a bit clearer on what the right questions might be. So perhaps the most useful thing given that we are very close on time is to just follow this track of Steve's, to talk about what's next. The suggestion on the table is that we think about three stages: Definitions, measures, and targets. And at the moment, the suggestion is that we try to approach the definition's phase of this project as a four-way conversation, engaging ccNSO, ALAC and GAC in that process. So are there any reactions to that proposition as to the way forward? Evan, thank you. >>EVAN LEIBOVITCH: One of the problems to me with the way we have been handling it is the fact that there is almost zero intersessional work going on between the silos. We come to these meetings, and we have these workshops at the meetings, and then we go back and we are basically in our silos dealing with the same issue. Now, at the risk of perhaps needing to put on my asbestos suit before suggesting this, there is a concept that has, in fact, worked called the cross-constituency working group. It has happened once with the Rec6 issue to do with new gTLDs. Some people may love or hate the current progress of the joint applicant support combined working group, but it has made progress. It has achieved attention of the board, the GAC and elsewhere. So these groups have been effective in being able to bring together multiple components of the ICANN community together in doing work and doing important work that doesn't wait for face-to-face at meetings to break down the silos. So while the mechanics and the logistics may be different from the CWGs that have happened before, I would suggest that this is perhaps a way to bring together the parties, not just the GNSO and ALAC as this happened, for instance, on the applicant support issue but to actively bring in ccNSO, GAC and perhaps others into something like this so we can continue to do this work and not wait for the next workshop in Dakar before we take the next step on this. >>ROSEMARY SINCLAIR: Anyone else? Jonathan? >>JONATHAN ROBINSON: Just to respond to the previous comment, I think it is a very constructive suggestion. The only challenge we've got is trying to make sure -- there is a number of parallel tracks of work. I'm aware of the work in the GNSO because that's a group that I'm effectively leading. There is also work in the ccNSO that I'm aware of to try and establish best practice or a way forward in the way in which committee working groups work. But notwithstanding that, I think it is a very good suggestion. We just have to balance that and make sure we get the mechanics of how it works right. >>ROSEMARY SINCLAIR: Okay. I think we're unhappily going to have to close off the discussion there. I wonder if we could have as an action just anyone who's interested in being part of the next stage of this work, if you could perhaps let me know. And then we can constitute ourselves however so that we've got the four-pronged approach and we scope out our work in the next little while. But I would like to really thank everybody on the panel and on the floor for participating in what I think has been a really very significant workshop. Thank you very much, everybody. [ Applause ]