News, views and what I choose to dos

What ICANN needs to learn from the EOI collapse

Category : Domain names, ICANN, Internet governance · by Mar 18th, 2010

At a recent meeting in Nairobi, the ICANN Board decided not to press ahead with a process to allow people to pre-apply for new Internet extensions.

This process was called “Expressions of Interest”, which was sadly but inevitably shortened to EOI within hours of being out for discussion.

Ultimately the Board decided that there wasn’t sufficient consensus in the community about allowing people to pre-apply before the full process for decided new “top-level domains” is finalized: an EOI process would use up valuable time and resources, better spent finishing up the remaining issues; and it was not certain what the process would achieve. So, prodded by many to make a decision, the Board did so – and said No.

Most people think this was the smart move – it didn’t create a new process, a clear decision was made, we can move forward.

I, on the other hand, think that the decision is the reflection of a much bigger problem: ICANN doesn’t learn its own lessons. And if the organization as a whole doesn’t start reflecting on what it has learnt over 10 years of activity – and start applying what it learns – then it is going to continue to waste time and energy and resources, to the detriment of itself and the Internet.

What do I mean that ICANN doesn’t learn lessons?

In response to comments on my blog post about the EOI process, I put down what I believe are three major lessons that have resulted from the unique multi-stakeholder decision-making process that ICANN represents.

Each time something has successfully moved forward in ICANN, one or more of these threads have been present. And, conversely, every time something has faltered or failed, it’s because one or more of these threads had been broken or ignored:

  • Lesson One: Relative deadlines are more effective than absolute deadlines.
  • Lesson Two: Wherever possible allow voluntary over compulsory contribution.
  • Lesson Three: The results of ICANN processes, no matter how broadly structured, represent a minority view.

In this case, a huge amount of time, energy and resources was wasted on the EOI because of a failure to follow lesson two: voluntary over compulsory. I’ll stick examples of where each lesson has appeared at the bottom of this post.

Down the wrong path

As soon as I read the staff analysis of public comments [pdf] to the EOI process, I knew the Board was doomed to take the most conservative route – which in this case was to say No and throw the process away. How come? Because the voluntary aspect was dismissed out of hand and without any real analysis.

One of the main conclusions in the document (p15):

“Of three options: a ‘mandatory’ EOI is preferred, not conducting the EOI is the second choice and a ‘voluntary’ EOI is a distant third. That is because a vouluntary [sic] EOI will incur time and expense (more than a mandatory EOI) and not achieve any of the program objectives.”

Of course if ICANN learnt its own lessons, it would realize that any program’s “objectives” are secondary to the larger realities of the multi-stakeholder model. The prime objective of an EOI process must surely have been that it won community approval. But that was never going to happen while the process was mandatory.

In that respect, the staff made a bad miscalculation by recommending that the Board go ahead with the EOI process – something that the Board felt compelled to strike down because there was not community consensus. This is a bad dynamic – staff making recommendations that the Board does not feel it can approve.

If on the other hand, ICANN’s own lessons has been taken into account and the start-point for the EOI process had been that it would be voluntary, then you would have seen much more fruitful discussion as people tried to figure out whether and how they could make it work, rather than be forced to come down on one side or the other.

By trying to achieve everything by making the EOI process compulsory; ICANN instead achieved nothing.

This is why ICANN – the staff, the Board, and the community – need to recognise the broader lessons and the patterns in their complex interactions. Everyone needs to take a higher-level reflection on the work that has been done over 10 years, rather than constantly fight in the trenches and so lose sight of the larger picture.

If these lessons are then applied, we should see much more effective and efficient work planning and discussions.

Real-life examples of where these lessons or threads have come into play.

If you can think of more, add them as a comment below. I’m sure there are more lessons that can be drawn out if people put their minds to it. A clear one to my mind is: openness – make everything available.

Lesson One: Relative deadlines are more effective than absolute deadlines.

Successes: Comment periods that account for other external factors.

Failures: New gTLD process timelines that constantly fell over; GNSO PDP deadlines in the bylaws that have to be fudged; comment periods that keep being extended.

Lesson Two: Wherever possible allow voluntary over compulsory contribution

Successes: GAC membership; ccNSO accountability framework/exchange of letters; RAA adoption (prior to end of contract); comment periods.

Failures: EOI process; efforts to force ccNSO managers to sign contracts with ICANN.

Lesson Three: The results of ICANN processes, no matter how broadly structured, represent a minority view

Successes: ???

Failures: IRT process (trademark rules created by closed group of trademark lawyers); Whois; dot-xxx.


(16) comments

4 years ago ·

While this article is well written and you certainly put in a lot of thought and experience, it fails to see the bigger picture. You yourself, Kieren, are trapped in the trenches, of the ICANN battles in this case.

Who the heck feels the need for new gTLDs in the first place? The internet to many people out there means dotcom. Full stop. Unless the community starts questioning this underlying assumption, nothing is going to come out of this anyway. With or without EOI.

fred krueger
4 years ago ·

Kieren, this post is severely misguided.

You state that what was wrong with the EOI process was the “mandatory” nature of it. That is absolutely not what went wrong with the EOI, or in my opinion, why it ultimately was voted down.

A voluntary EOI what have been something like a voluntary jail sentence. No one would have “participated” (to use an ICANN term).

I hereby declare a free, mandatory EOI process to all. Please send your EOIs via email to, so that they can be used to assess root scaling. No strings will be published so as not to hurt applicants “business models”. A four month “communication period” is now in effect, so the EOIs must be turned in between July 19th and September 19th.

After the EOIs are turned in, ICANN will proceed with an eight week “analysis” of the strings, followed by a GAC “categorization” to better assess “track differentiation”, and special treatment for strings of “public interest”, specifically dictionary words of less than 6 letters in the major 127 languages and scripts.

This new “voluntary accountability mechanism” will help keep ICANN beholden, as you say, to “relative deadlines” – such as how to delay making a decision to the next ICANN meeting, wherever that might be.


4 years ago ·

Hi Eberhard, well, wrt the whole issue of new gTLDs, I think there is another high-level lesson that we have learnt when it comes to the Internet: more options and more competition is a good thing.

None of us grasp what expanding the top-level domain space will do, in the same way that no-one grasp what the Internet itself would do. I recall Vint Cerf talking about the fight he had in the early days to persuade people to build the network so it could expand to one million nodes. No one thought it was necessary.

All I know is that we are the dotcom generation, and that our children will find it quaint that there was one big Internet extension for so many years. The Internet is already much bigger than the top level extensions we already have.

4 years ago ·

@Fred: Well, unsurprisingly, I don’t agree. And I think that’s because you come from a very specific viewpoint – you want new extensions and you want them now because you have spent a huge amount of time and money preparing for them.

I am looking at it from the organizational perspective – trying to figure out a way to progress through the multitude of different groups within ICANN.

The fact is that the process failed. And the fact is that alot of people saw value in the idea. So something went wrong – and that was that the community as a whole became divided. And as soon as that happens, you lose consensus and ICANN will either drop the plans or spent years working on it in the hope of finding a compromise.

The solution to progress in ICANN is to keep finding things for people to work collaboratively on. As soon as you say something must be done, one group or other will recoil and walk away.

I note you have written off a voluntary EOI process with the same haste that the staff wrote it off. But it is this voluntary approach that has worked time and again within ICANN. Why? Because ICANN is about coordination and not control. Control does not work.

If for example the Nairobi meetings had focussed on the question: so what do we need to do to make as many new gTLD applicants as possible apply through the EOI process? Well, then I think there would have been some interesting discussions. And at the end it would have been a case of whether the community was willing to accept the price of the incentives to participate. Whether the benefits to getting hold of new gTLD data outweighed the costs of providing incentives.

Even if the process then failed, you would have had a conversation that was useful – all parties would have understood better each others’ perspectives when it comes to new Internet extensions. It would have put people fears and concerns on the table and made progress on dealing with them.

Instead what you had was people fighting for and against a series of questions for a process that was ultimately discarded. It was a waste of time and resources unfortunately.

My whole point to this post is: there are bigger lessons and principles out there that you should be used as guidance in creating new processes. After 10 years, ICANN is now in the position where it doesn’t have to reinvent the wheel every time. We know what works and what doesn’t.

Just as a last point – my analysis is just one person’s view. Really there should be someone talking to a broad group of people about why the EOI process failed and what lessons can be learnt.

But here’s the problem: no one does that. There is no self-analysis within ICANN and as a result the organization is doomed to repeat the same mistakes over and over again. It’s time for ICANN as a mature organization to review its performance over time rather than get stuck in the day-to-day.

The Affirmation reviews are a manifestation of that need but, I suspect, will be too short-term and rife with politics to do a useful job this time around.

Khanan B.
4 years ago ·

If there’s a two-word term that best encapsulates what ICANN is all about, it must be “relative deadlines.”

Khanan B.
4 years ago ·

Kieren, let me add that your belief that every single individual and group in the ICANN community can be made to agree on every issue of relevance, if only the issues are discussed at enough length, to be almost touching in its naivete. Some people want new TLDs; others don’t. You will never persuade either side to abandon their views on that issue. If “consensus” between those two groups is required to move forward with new TLDs, ICANN might as well scrap this charade of a process.

4 years ago ·

@Khanan: You’re using a ridiculous form of argument here – I never said anything close to every single individual and group being made to agree – yet you state it as if I did, and then criticise it.

The problem is that too often ICANN listens to those that want to win or ruin, mostly because they make the most noise. You can’t see how you can “win” in the new gTLD debate and so you want to ruin it by “scraping this charade of a process”.

Meanwhile, the people doing the real work are constantly trying to find suitable ways forward and compromises. It’s what ICANN is – a multi-stakeholder model. It can be slow and it can be noisy. The aim of my post was to point out what I think could be an improvement in the effectiveness of that model.


Khanan B.
4 years ago ·

@Kieren: Do you, or do you not believe that there will ever exist a “consensus” between pro and anti-TLD camps? As we saw in the EOI process, if one of those constituencies makes enough noise, ICANN will opt for inaction. As such, I don’t see how ICANN ever makes a final decision on this program or sets a final application date.

As an unrelated aside, re: “Relative deadlines are more effective than absolute deadlines.” – the three examples you use are instances in which ICANN couldn’t get their act together to actually make a decision to meet a deadline! So the “lesson” to be learned by ICANN is to never set deadlines because they are consistently incapable of meeting them! As Fred said in his tweet in response to this post, this is truly straight out of Kafka.

Antony Van Couvering
4 years ago ·

The EOI process failed for a very simple reason, one I pointed out in my letter to ICANN’s CEO and its Chairman a month ago – it had become a proxy for the gTLD program itself. It has nothing to do with relative deadlines or voluntary participation or anything of the sort.

One of the effects of the EOI would have been to expose the plans of some powerful players who have been very content to keep them quiet. These include brands, who say they want nothing to do with the gTLD process, even as they make plans to apply, and incumbent registries, who claim benevolent disinterest even as they seek to bleed new participants so that they are severely weakened by the time new gTLDs finally get started. This part of the “community” was against EOIs (duh) and they were joined by several from the non-commercial group, who thought that EOIs would hurt small registries financially — an absurdity that can only be ascribed to their being effectively innumerate when it comes to financial matters (see quote below).

Here’s what Citi analysts, in making a “buy” recommendation for VeriSign, said about the Nairobi decision:

“Last week in Nairobi, ICANN essentially upped the financial threshold for those wishing to submit an application for a new gTLD (eliminating the “Expression of Interest” process). The Expression of Interest (EOI) was a smaller ($55K) fee that a party could pay to essentially “reserve” a new gTLD. This could be paid as essentially an “ante” for a party to test the waters before having to pay the full $185K application fee. Without the EOI, parties now enter the process “blind,” having to commit full financial and other resources to competing for a new gTLD, without having any knowledge of others that might be bidding. As this raises the stakes to apply for a new gTLD, it will likely cut down on the number of applications.

“…a more orderly new gTLD process and fewer applicants create less noise in the industry, which is positive for VeriSign…. we do believe more noise around gTLDs could create investor uncertainty. Thus fewer new gTLD proposals would be positive for VeriSign.”

So, a win for incumbent registries and brand owners. What a shock! Was it not of interest and note that VeriSign kept its employees out of Nairobi because of security concerns — except its two lobbyists, who were very much in evidence?

This is the real lesson of the EOI program – that a proposal created by the community, supported in the comment period by a 2-1 margin, supported by just about all potential applicants, was voted down by the Board because it was afraid of (get this)…. delays. How can that be a credible justification from a body that ripped up the timeline for gTLDs in Seoul without even an attempt at justifying their action?

Virtually the only good thing about the EOI vote is that it did not die a lingering death. A voluntary, non-deadlined process would have had exactly the same result, except it would have been more painful and taken longer.
A few counter-examples to your rules.

1. The registrar test bed. Had hard, non-relative deadline, participation voluntary but restricted to those who applied (just like EOIs). Worked, thriving registrar market, consumer prices lowered considerably.

2. Nominating Committee – Hard deadlines, participation voluntary but restricted to those who applied. Has produced better, more intelligent, harder-working Board members (though admittedly the bar was low).

3. Recent GNSO trademark group (STI). Hard deadline, participation voluntary but restricted to those who applied. Produced a very good report on trademark issues that really showed what could be agreed upon and what remained to be settled. Biggest issue there is that they *weren’t* asked to look at the Post Delegation Dispute Policy.

Your example of the ccNSO letters exchange program is particularly inapt. Under Twomey at least, arm-twisting and favoritism were the order of the day (can you say “zone file updates”?) Some of these letters were signed “voluntarily” in the sense way that KGB confessions were “voluntary.” And don’t even get me started on some of the re-delegation decisions.

So I don’t agree with your analysis of why EOIs failed, and I don’t agree with your prescription for how to fix ICANN. What ICANN needs more than anything else is a top-to-bottom restructuring that does away with interest silos — such as stakeholder groups and constituencies — that concentrate and harden self-interested positions and prevent rapid policy development.

4 years ago ·

Hi Antony,

As ever, some interesting points and analysis. I’m not going for it though because of two simple reasons:

1. You are putting the failure of the EOI process down to a conspiratorial effort of others against your position.

I have spent years combing through all the ICANN conspiracies both as a journalist and as a staffer. The truth is that talking about conspiracy does not create conspiracy.

There is a simpler truth: efforts to impose one group’s perspective on another are resisted. The more they are pushed, the more they are resisted. Unfortunately what this means is that we keep ending up with the status quo because people continue to believe – wrongly – that they can force through change.

2. Your solution is to tear up the ICANN model and start again. This isn’t a solution – it is a sign of frustration.

I have lost track of the countless times that, rather than than seek to engage or understand other groups, when one group doesn’t get want they want, they point to conspiracy and then call for banishment of vested interests.

This argument is anti-consensus and anti-community. As such it goes against the model itself, and so is disregarded. You just have to face the realities of the decision-making model that ICANN represents.

To my mind, the best way to improve the functioning and effectiveness of ICANN’s processes at this stage is to improve the level of analysis of the results. That and find ways to allow more people to effectively participate by removing the brick wall of ICANN’s “steep learning curve”.

4 years ago ·

@Khanan: So this is the thing – the argument about *whether* to have new gTLDs is done. It was over the day ICANN was formed.

There’s certainly nothing stopping people from railing about it – and occasionally the efforts of those people to prove new gTLDs are a bad idea actually help firm up plans to introduce new extensions – they are a useful check.

As to noise levels – it’s true, ICANN is like an animal with a shell. When it hears alot of commotion going on around it, it stops and pulls its head in. And it will keep doing that until people learn that it’s not in their own interests to yell at one another.

Khanan B.
4 years ago ·

“You are putting the failure of the EOI process down to a conspiratorial effort of others against your position.”

Kieren, do you dispute that trademark interests, incumbent registries, and certain brands are doing their utmost to delay the new gTLD round based on whatever reasons they can latch on to, and out of pure self-interest? If so, you clearly have not been paying attention.

Antony Van Couvering
4 years ago ·


As a former staffer, not long parted, perhaps it is you who are viewing things from a particular point of view. Your tenure may have vested you with a desire to make something work that doesn’t really work. Your cures for ICANN’s ills depend on a particularly benign view of human nature that most assuredly has not been on display for much of human history. Your ideas about voluntary participation and relative deadlines will do nothing to fix the pitched battlefield that ICANN has become.

Asserting that incumbent registries and IP interests will do anything legal to advance their cause is not a conspiracy theory — an employee who protested against that would likely be fired, and if not the shareholders might fire the management that didn’t fire her. That’s what companies do, that’s what they tell anyone who will listen that they will do.

Conspiracies as common as domain names, but successful conspiracies are a result of planning and luck and come along only infrequently. ICANN meeting hallways are filled with little else than people planning conspiracies, but we don’t see many results because through lack of interest, perseverance, resources, and resolve, most of them never come to fruit. Read any constituency list: they are either announcements or plans to fight or co-operate with this or the other constituency. If you prefer to call this co-operation instead of conspiracy, then you have called wormwood a rose, but it is no less poison.

Large companies with hundreds of employees are simply better provisioned and more adept at haunting the corridors of power and waylaying their prey with sugared arguments — probably through hard experience. I do not impute to them any motives than those which they are paid to have.

I’m not saying that anyone conspired with anyone. That explanation is at once trite and incomplete and calls into mind images of dark meetings and blood oaths. Why bother with conspiracies when you can do it all right in the open, according to ICANN procedures?

Rather say that as EOIs became a proxy for the “overarching” battle to slow down and entangle new gTLDs with ingenious new costs and procedures, the interests who would benefit from such as result sprang into action — as they will, as they must.

ICANN would work much better if these defined silos of interest were knocked down and everyone had to rough-and-tumble with everyone else. Imagine the U.S. Congress with pre-defined voting blocks for states, or manufacturing sectors. Caucuses of like-minded legislators do form, and meet, but they are not woven into the fabric of the institution, as ICANN constituencies are.

As potential applicants, we have no power base, no voting rights to affect PDPs, no way to participate in many of these wieldings of influence. That’s not a conspiracy, but it is a flaw that points to a bigger problem, which is the nature of representation within ICANN. Voluntary participation and relative deadlines will do nothing to heal that larger wound.

4 years ago ·

Anthony, you know that I have a lot of personal sympathy for your position – the new gTLD process has gone on far too long. It is convulted, it’s messy, it’s inconsistent and it risks undermining the very competition it was designed to introduce.

But slipping into the mistake that so many frustrated ICANNites have followed in the past – waving at vague conspiracies and declaring the whole model broken – will do nothing to aid your cause. ICANN as an organization works most effectively and moves forward most quickly when people are working positively and collaboratively.

There is no doubt that representatives of all sides put forward arguments that most benefit them. And in the case of incumbent gTLD owners, and also trademark holders, it’s not hard to see why the better option is currently to go slowly, step-by-step.

But to confuse that perfectly rational behaviour with a malicious conspiratorial intent only raises the temperature and – ironically – increases the likelihood that people will aim to work around others.

ICANN is a particularly political and occasionally unpleasant environment but it has taken great strides in the past few years, mostly because people have started working with one another more, having recognised their co-dependency in the model.

Incidentally if you had recognised ICANN realities and pushed for a voluntary EOI – which you would probably have achieved – then you would be in a strong position to apply to have applicants form a new GNSO constituency. And then you would have had direct voting rights. It’s all about how you work within the system.


Antony Van Couvering
4 years ago ·

@Kieren – please re-read what I said. I specifically eschewed any notion of conspiracy. As I said, why bother, when that is how ICANN works anyway.

Spare me please sage words about working within the “system.” I helped form the system, I’ve seen it rooted up and re-cast more than once, each time improving the initial set-up. I’ve seen the DNSO become the GNSO, the ccTLD constituency become the ccNSO, I’ve seen the constituencies within the GNSO become the contract and non-contracted houses. The “system” you graciously advise me to work within (as if I haven’t been) is nothing other than what all of us have decided at any given time to use as a framework.

ICANN’s structure was not proclaimed by a prophet holding stone tablets. There is nothing permanent about how ICANN works, and certainly there *should* be nothing permanent about it.

If you’re telling me that the path to representation within ICANN depends on proposing a “voluntary” procedure that would have been gamed like crazy and would have produced no useful results, but would have given important stakeholders some vague and dubious standing to apply for representative rights just because they used the word “voluntary,” then you’re talking about a system in serious need of reformation.

This isn’t about “can’t we all get along” and recognizing that we all depend on each other, etc. etc., within a closed and as-yet imperfect structure. This is about a continuing effort to construct a system that is representative of Internet users and achieves a consistent, transparent and fair method of developing policy, with rules that preserve a balance of power even when people act badly.

Certainly things work better when people recognize their common interests, but building a system that depends them doing it it is not a foundation for a durable institution — it’s a simple failure of imagination.

Khaled Fattal
4 years ago ·

Dear Kieren,

In over ten years ICANN has shown less and less propensity to listen and learn from its mistakes as time went on and the issues became more and more contentious. I agree with your analysis and conclusions. They represent some of operational reasons why EOI was voted down, but not a main strategic reason.

I believe, above all, ICANN board has still not come to terms what their true new role is after the AOC. And they will continue to fail the global Internet community until they acknowledge to themselves and the world that the AOC now makes them a ” Global Public Service Provider”.

In fact ICANN is a “monopoly” Global Public Service Provider, stakeholders go to ICANN because they have to not because they enjoy ICANN’s superior equitable and fair service – typical of a monopoly. I recommend reading my blog on circleID the morning of the Nairobi open forum and my accurate predictions of the board decision on EOI, and the purely legal manipulation on .XXX to defer it to public comment to wiggle out from standing up and being counted one way or another. Primarily, a lack of true leadership.

Finishing on a positive note. Someone like yourself who has worked inside ICANN for a few years does know more about the internal mode of thinking and operation at ICANN. Your opinions and positions that ICANN continues to fail to learn from its mistakes echoes what I have been saying for years. In many instances medical deafness often can be cured. ICANN on the other hand suffers from systematic failures to listen and learn but at no cost to its senior staff or board but to the detriment of the Global Internet stakeholders. I commend you for what you wrote. It speaks volume about ICANN

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>