Category: domain names

More shenanigans have now been detected in the ICANN RPM PDP, namely attempts to manipulate the outcome of the policy work through duplicative and/or coordinated comment submissions.

The ICANN RPM PDP Phase 1 comment period ended on May 4, 2020, more than 2 months ago, yet ICANN had repeatedly refused to allow the raw public comments to be downloaded for easy independent analysis. I discussed that in a prior blog post.

After much energy expended in a back and forth with ICANN’s Complaints department, some (but not all) of which has been posted under Complaint #00014905 on ICANN’s website, they finally did the right thing and made the raw comments downloadable yesterday (just go to the File menu, and select the Download submenu, choose a format, etc.). Making the public comments spreadsheet downloadable was something that literally took ICANN staff at most 5 minutes to do, yet they spent months refusing such access. That is a cultural problem at ICANN,  which refuses reasonable requests in the public interest, unless one is an “insider” — had it been Verisign or another “friend” of ICANN, surely they would have simply done the right thing immediately.

As I suspected in the prior blog post, there were seeming irregularities in the public comments:

There may be even more irregularities which would be easier to detect if the data was more easily accessible.

I have now created a new spreadsheet (available in Excel format here, if the link to the Google Sheet is not working, e.g. if you’re in China) that highlights specific comments that were submitted, allowing one to easily compare them and see how similar they are to one another. The “Form Responses 1” tab is the original raw data, in its hideous format which made comparisons and analysis quite difficult (couldn’t resize or move columns/rows,  long comments hard to read, couldn’t copy/paste, etc.). The “Analysis 1” tab takes row 1 and turns it into Column A using the “TRANSPOSE” function (see formula in cell A1). Similarly, the comments of Hermes, UNIFAB, Comite Colbert, Chanel, Moncler and Novartis are taken from rows 7, 25, 28, 30, 32 and 48, and turned into columns B through F, again, using the “TRANSPOSE” function (see formulae in cells B1 through G1). Thus, the cells weren’t copied and pasted — they were pulled using a formula, and thus no duplication errors were introduced by incorrect copying of cells to the wrong place. I then made a few cells bold, added a background colour for column A, and resized the columns B through F, to make it easier to see the similarities.

As you can easily see by simply scrolling down through all the rows (helps to have a large monitor, to see all 6 comments side-by-side without scrolling), these were not 6 independently-generated comments. Instead, they are essentially 1 submission that has been, for the most part duplicated (with minor variations in a few places) 6 times. Remember, the original spreadsheet did not allow downloads or copying and pasting of answers. This implies that those 6 companies likely coordinated their submissions via a separate (and presumably private) master document that they shared amongst each other, which was used to copy/paste submissions. This then allowed them to over-represent their comments, relative to other submissions, in an attempt to manipulate the outcome of this PDP by making their preferred outcomes seem more popular, while simultaneously diminishing the impact of opposing positions.

The “Analysis 2” tab  highlights the comments of Renee Fossen of Forum (National Arbitration Forum) and Richard Hill. As I noted previously, many of Mr. Hill’s comments simply supported Ms. Fossen’s.

Going back to the the Analysis 1 tab, this is not the only time some of those  companies have submitted mostly duplicative coordinated comments to ICANN. Their comments to the WHOIS EPDP comment period can be seen via the spreadsheet here. While not downloadable, one can simply compare submissions in rows 9, 11,  12, 13 and 14 with each other and note the striking similarities.

The goal of these public comment periods is to get a sense of how the community feels about the various proposals. As such, it is important that those submissions be representative. When there is a small number of unrepresentative submissions, policy outcomes can be skewed by those over-represented participants or voices, like those 6 European brand owners.

Why does this matter? In the RPM PDP comment period, there were only 55 submissions that were counted (some of the obvious duplications were already eliminated, e.g. the multiple submissions from Hermes and the IPC that I had previously discussed). If one eliminates 5 or 6 of the above comments as duplicative (let’s say 5, to be conservative), that reduces the total number to 50. That might not seem like a big difference, but it’s enormous, given that the superficial analysis performed by the working group members (remember, I’m unfairly banned, and am unable to participate, but have reviewed all the recordings) focused on just the “numbers” as to how many “votes” a proposal got from the community. Removing 10% or more of “votes” from specific proposals would shift the balance considerably, on nearly every issue.

Let’s look at a specific example, URS Individual Proposal #15, a deeply flawed proposal that I’ve discussed in the past and which you can see my own comments in cell C19 of the “Public Comment Review Tool” (in the URS Proposal15 tab). The public comments were reviewed on June 25, 2020, and one can read the transcript of that working group call (starting from page 7). Without adjusting the responses, 45.5% of respondents, i.e. 25 out of 55, did not support the proposal. That level of opposition alone should have been enough to kill the proposal outright. But, on the other side, UNIFAB, Colbert, Chanel, Moncler and Novartis all supported it with a “minor” change and Hermes supported it “as written”. “Support as written” and “support concept with minor change” had a combined 15 out of 55 responses (27%).

But, adjusting the responses to eliminate the over-representation of those European brand owners has 2 effects. First, the opposition becomes 25 out of 50 (i.e. the denominator changes from 55 to 50), and so that becomes 50% (instead of 45.5%). And, for those supporting it, it goes from 15 out of 55 to just 10 out of 50 (i.e. affects both the numerator and the denominator), and thus that becomes only 20% support.

Shockingly, and a sign of the true level of capture of the working group, that proposal continues to live to fight another day (see page 25 of the transcript), rather than being tossed in the scrap heap.

Phil offered us an opportunity to live another day, to take the concept back, to rework it and reintroduce it to the working group. So I think that those of us who wanted the concept to survive should declare victory, retreat, and work on this together and then try again with introducing something to the working group that might be more acceptable to the working group at-large.

Truly disgusting, and I’ll have more to say about the repeated double-standards evident throughout the superficial “review” of the public comments in future blog posts.

Similar swings in relative support levels would happen on other proposals and recommendations, once the proper adjustments are made.

It’s clear that the working group must now go back and redo the analysis of the past 2 months, to remove the effects of those attempting to manipulate the apparent popularity of their positions. That’s for starters. The blame for this falls squarely on ICANN staff and the co-chairs, for (1) not providing a downloadable version of the comments 2 months ago, when multiple people asked for it, (2) designing a purposely-flawed comment period that created enormous burdens for those wishing to make comments, and (3) not doing widespread outreach to get a large representative sample of the views of affected stakeholders.

You’ll recall that ICANN dismissed and denigrated the thousands of comments that were submitted in the dot-org and dot-com contract comment periods. They were dismissed as spam, for example.  Verisign themselves stated:

The Internet Commerce Association (“ICA”) and other allies in the domain name speculation business including registrars like Namecheap and Dynadot, have made a concerted effort to distort this public comment period. They have distorted the facts in a campaign-style effort to flood ICANN with public comments created by form-letter generators and templates created for the sole purpose of protecting their own financial stake in the speculation business.

Verisign went on to write about attempts to “hijack the legitimate public comment period”. Those are strong words. If Verisign is to be consistent, then the distortion caused by those European brand owners, the attempt to “hijack the legitimate comment period” must be addressed.

This failure caused by ICANN staff and the co-chairs can also be addressed by having a second public comment period. This should be by the traditional email interface, rather than the flawed Google Forms that has been repeatedly and roundly criticized by multiple stakeholders.

In conclusion, this new evidence reinforces the conclusion that the RPM PDP is a captured and sham working group, that needs a complete overhaul, if it’s even worth salvaging in any form. Their work product, upon any serious examination, has been complete garbage. While some may “hold their nose” and live with it, the stakes are too high for the public (and domain name registrants in particular) to allow this fiasco to continue.