Summary

  • ICANN’s first At-Large election seated five directors after 34,035 people voted, but that number described the voters who completed the process, not a representative sample of Internet users.
  • The Board accepted an independent monitor’s account of the regional results while its later study documented a much larger, uneven registration and activation funnel.

In 2000, ICANN tried to bring individual Internet users into its boardroom by putting five seats to a direct vote. The exercise answered one question unusually clearly: who won each regional contest? It left another unresolved: what could those winners claim to represent?

The distinction matters because election statistics tend to travel. A count can begin as an operational fact—how many people cast ballots—and later acquire a political meaning: how much public authorization the winners possessed. ICANN’s first At-Large vote shows why the two should not be fused.

The plan began with a Board decision in May 2000 to prepare elections for five At-Large directors, one from each geographic region. Separate regional contests were a response to an imbalance already visible in the applicant pool. The Election Committee cited that roughly 80 percent of applicants came from North America and Europe. It proposed that voter eligibility follow residence, while candidate-diversity rules used citizenship. These were design choices to avoid letting one concentrated pool determine all five seats; they were not evidence that the eventual electorate matched the world’s Internet users.

ICANN’s archived election page later reported more than 158,000 applications and over 76,000 people who activated membership by the September 8 deadline. The 2001 At-Large Study Committee used different, later categories: 176,849 total registrations, 143,806 valid registrations, 76,183 validated registrations and 34,035 voters. Those figures should not be collapsed into one funnel without their definitions and dates. The final report says the five directors were selected by a direct global election of 34,035 email-address holders and took their seats in November 2000.

The Board’s October 17, 2000 minutes put turnout at about 45 percent among those who returned to the site and activated their registrations. They also record a voting-site outage of roughly 45 minutes near closing; the deadline was extended by about an hour. The Board then relied on an independent monitor’s finding that the regional tabulations accurately reflected the choices expressed by the ICANN electorate. That is meaningful evidence about the count and the winners. It does not establish that the electorate was a statistical proxy for everyone online.

The later study committee treated the five-seat arrangement as an interim compromise and raised separate questions about voter-list integrity and access. Its report noted that 9,721 PIN letters had been returned by August 19, 2001, mostly from Asia-region destinations, and described the affected group as 6.8 percent of valid registrants unable to vote. This later postal-verification evidence identifies a design and access concern; it is not a finding that the 2000 results were fraudulent or wrong.

The election therefore had two records that can both be true. A monitor could validate the tabulation of ballots cast by a defined electorate. A later institutional review could still ask whether the registration, activation and verification system had made that electorate broad and dependable enough to carry a larger representational claim. Counting accurately is a property of a result. Representation is a claim about the people behind it.

Primary records: ICANN At-Large election archive; At-Large Study Committee Final Report, 5 November 2001; ICANN Board minutes, 17 October 2000; Election Committee recommendations; Board preliminary report, 4 May 2000; ICANN At-Large history.