Survey of Usage of the .BIZ TLD
Evaluation of TLDs

Jonathan Zittrain* and Benjamin Edelman**
Berkman Center for Internet & Society
Harvard Law School

[ Overview - Data & Analysis - Conclusions ]

Abstract: The authors examined a variety of data about registrations in .BIZ. Analysis suggests that at least 74% of currently registered .BIZ domains provide no web content or provide only error messages or placeholders. Using WHOIS data, we conclude that approximately 25% of .BIZ registrations are registered to the same organization that registered the corresponding .COM, and that such domains are about one-third less likely to provide substantive web content than domains registered by someone without a corresponding .COM registration. We document 4886 distinct .BIZ domains seemingly inconsistent with .BIZ registry anti-warehousing policies, suggesting that such policies are not being effectively enforced. Finally, we compare .BIZ registrar market shares with corresponding shares in .COM, .NET, and .ORG, and we summarize notable differences in registrar market share across TLDs.

Overview: The first six months of .BIZ registrations

One of ICANN's primary tasks since its 1998 inception has been to break a de facto deadlock over the introduction of new top level domains for the Internet's default domain name system -- the DNS. ICANN has launched a New TLD Program, and in November 2000 its board cleared the way for the introduction of seven new top level domains into the root over which ICANN has partial policy authority. Among these new TLDs was .BIZ, a namespace proposed by the Neulevel registry for use by "companies large and small around the world."

After a 3-month period in which those asserting rights to second level names (Coca-Cola seeking COKE.BIZ, for example) could provide notice and a reservation to that effect -- subject to challenge by others -- accredited registrars were permitted to allow the general public to register names first-come, first-served in .BIZ in real-time on November 7, 2001. Six months later, we have sought to review registrations to date so as to methodically evaluate the scope and nature of usage of the .BIZ TLD.

Our work begins with a survey of registrations in .BIZ. In particular, we were curious about the following:

• Overlap between holders of new .biz name and corresponding names from existing registries. Using a random sample of .biz names drawn from the 681,150 registered entries, we have sought to track how many names in .BIZ have also been registered in the longstanding .COM, .NET, and .ORG generic registries, and whether corresponding .COM registrations are held by the entities holding the corresponding names in .BIZ.
  
  The answer to these questions may shed light on the extent to which new TLDs like .BIZ (1) provide "new" territory for those now unable to find a short, mnemonic name in the crowded .COM space, or (2) whether they instead are largely places for the existing .COM registrants to stake corresponding claims. To the extent that (2) is true, we have estimated whether such registrations are truly "defensive" -- i.e. the names are reserved solely for the purpose of preventing someone else from reserving them, and are not themselves deployed by the registrants as a means of reaching the registrants' Internet sites -- or whether they're used to diversify access to a registrant's online presence. For example, if the latter, acme.biz might point to investor information about the Acme company, while acme.com could be the Acme company's consumer commerce site; if the former, the holder of acme.biz may simply not have assigned it to any particular web address -- the mere registration was enough for the registrant's purposes. To do this, we have counted how many live web sites can be found within a sample of claimed .BIZ names, as well as those specifically likely to be names already held by the same registrant in .COM.

• Level of enforcement of registry-specific restrictions. As new TLDs seek to differentiate themselves in part through restrictions on domain registration and use, we further sought to investigate the effectiveness of .BIZ's registration restrictions, many of which are not present in .COM, .NET, and .ORG.

Data & Analysis

Our analysis uses the .BIZ zone file of June 6, 2002, which includes a total of 681,150 distinct registered .BIZ domains. The indicated portions (†) use a simple random sample of 68,192 distinct .BIZ domains drawn from the entire zone, selecting any given .BIZ domain with equal probability.

• Overlap between strings registered in .BIZ and existing registrations in .COM, .NET, and .ORG†
  • Availability of web content on .BIZ domains†
    
    
• Level of enforcement of .BIZ registry restrictions
  • Registration patterns of top .BIZ registrants
  • Registrations that do not conform to .BIZ registration restrictions
    
    
• Additional characteristics of .BIZ including international registrations and registrar market share

Overlap between strings registered in .BIZ and existing registrations in .COM, .NET, and .ORG

In this section, we document the extent of overlap between registrations in .BIZ and .COM, .NET, and .ORG, and second to estimate what proportion of .BIZ names are held by the same entity that registered the corresponding .COM.

For a random sample of distinct .BIZ domains, the authors determined whether the corresponding .COM, .NET, and .ORG domains was registered -- whether or not by the same registrant -- as of June 22, 2002. The results are as follows:

These results show that more than half of .BIZ registrations remain available for registration in at least one of .COM, .NET, and .ORG; since such .BIZ domains could have been registered in at least one of .COM, .NET, and .ORG, this statistic suggests that .BIZ has in a certain sense expanded the namespace less than a simple accounting of the total number of .BIZ registrations would suggest. We also note the many .BIZ domains registered when the corresponding .COM is taken but even when .NET and/or .ORG remain available; we take this to suggest that, for at least some registrants, .NET and .ORG are less desirable TLDs than .BIZ itself.

Among the sampled .BIZ domains also registered in .COM, we have further attempted to assess whether the registrant of the .BIZ domain is the same individual or organization that registered the corresponding .COM domain. Of the sampled .BIZ domains also registered in .COM, we drew a further a random sample of 823 .BIZ domains with retrievable .BIZ and .COM WHOIS data; we then manually compared the sets of records, using information about registration postal code, nameserver designation, and admin email address to determine matches in registrant identity. These several comparisons are necessary since there is no standardized format for WHOIS data across (and even within) registries, and because there is no unique marker -- like a tax ID number -- to readily identify a given registrant. For example, a name held by IBM might say "IBM" in one record, "International Business Machines" in another, or the name of a subsidiary in a third.

Postal codes were considered to be the same if they matched exactly; US ZIP codes were checked at the five-digit level even if nine-digit ZIP codes were available. Name servers were compared at the second-level domain in order to treat pairs like ns1.domainname.com and ns2.domainname.com as a match; where indicated, email addresses were also compared in this way, but email domains were compared at the third-level domain in case of ccTLDs with non-delegated second-level domains, such as *.ac.uk.

Comparing the respective WHOIS data yielded the following summary statistics:

This data suggests that roughly 20-30% of .BIZ domains are openly registered to the same entity that holds the corresponding .COM. (We attribute the relatively lower rate of matching of nameservers to non-retroactive changes in registrar choice over time by the same registrant and to the use of registrar-provided name-parking services.) Since the accuracy of WHOIS data is entirely in the hands of the entity registering a domain, there may be a number of other names held across registries by the same entity, but for which WHOIS analysis discerns no similarities.

At the conclusion of the subsequent section, Availability of web content on .BIZ domains, we analyze the usage of .BIZ domains likely registered to the same registrant as the .COM. We find that when a .BIZ domain is registered by the registrant who also holds the corresponding .COM, the .BIZ domain is approximately one third less likely to provide substantive web content.

Availability of web content on .BIZ domains

To speak to the usage of .BIZ, the authors attempted on June 21-22, 2002 to retrieve the default web page from the www. host of each of a random sample of 68,192 distinct .BIZ domains drawn from the complete zone file. The results of these attempts are listed below. Corresponding results for a random sample of 212,563 .COM domains are listed for comparison purposes.

The rate of valid HTTP responses from .BIZ domains is statistically significantly smaller than from .COM domains. (The estimated 95% confidence interval for the .BIZ proportion of valid HTTP responses is [0.661, 0.669], while the corresponding estimated confidence interval for .COM is [0.763,0.767].) We cannot say whether this effect is properly attributed to the greater age of the .COM registry and a typical registration within it; to some identifiable characteristic of the .BIZ startup period or registration system; or to another cause.

For the sampled .BIZ domains providing a valid HTTP response, their HTML titles were as indicated below. The left column gives results as a proportion of .BIZ domains providing a valid HTTP response, while the right column gives results as a proportion of all .BIZ domains. Classifications used both automated and manual techniques.

In total, we find that of .BIZ domains with valid HTTP responses, at least 60.7% provide responses that are blank, errors, or placeholders for possible future content. This is likely to be an underestimation of the proportion of .BIZ names that have no unique functioning web site behind them; anecdotal examination of many of the remaining web pages shows that they also fail to offer web content, while others offer the same content available at existing sites in other TLDs. Of .BIZ domains sampled, then, at most 39.3% of those with valid HTTP responses provide actual web content. Combining this result with the proportion of .BIZ domains found to provide valid HTTP responses, we estimate that at most 26.1% of .BIZ domains offer actual substantive web content. (A 95% confidence interval for this quantity covers [0.258,0.264].)

These results are summarized graphically in a pie chart.

The 438 distinct domains found to contain "for sale" or similar content are included in "Registrations that do not conform to .BIZ registration restrictions," infra.

For the .BIZ domains previously found to be likely to be registered to the same entity that registered the corresponding .COM name (via a match of at least two of registrant email domain, name server, and postal code), we examined web page titles and contents in search of notable difference from .BIZ registrations generally. Results were as follows:

These results suggest significant differences between the use of .BIZ domains registered to the same entity that registered the corresponding .COM domain and the use of all .BIZ domains. In particular, .BIZ domains registered to the same entity that registered the corresponding .COM are more likely to provide an invalid HTTP response or an under construction page; in other words, such .BIZ domains are less likely than other .BIZ domains to make active substantive web use of the registered .BIZ domain. However, their somewhat higher rate of blank response pages likely reflects a higher rate of redirects to other web content.

Level of enforcement of .BIZ registry restrictions

Registration patterns of top .BIZ registrants

A few .BIZ registrants account for a number of .BIZ registrations; the top 10 registrants represent 3.0% of all registrations, and the top 50 registrations represent 6.3%. As a result, reviewing the registrations of top registrants can provide significant insight into the nature of .BIZ registrations and into the usage of .BIZ. The listings below report .BIZ registrations by registrants with 20 or more .BIZ registrations. Distinct registrants are identified by unique combinations of email address, registrant name, and registrant organization.

.BIZ Registrations by Top Registrants
.BIZ Registrations by Top Registrants, with domain listings
    Page 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14

Inspecting the registrations of top registrants shows several notable categories of .BIZ domains. Some domains, including 3,812 registrations by Novartis and 2,188 by Johnson & Johnson, seem to be reserved for future use or branding; when these domains provide web content, they generally point only to "under construction" or similar pages. Others, like 2,645 registrations by by RareNames and 193 by a registrant of name "shoval - DOMAINS FOR SALE," make offers for domain purchase either on default web pages or in WHOIS contact data. (Registrations that make offers of sale are included in the subsequent section entitled Registrations that do not conform to .BIZ registration restrictions.) Finally, many other registrants register large numbers of generic .BIZ names; a contact listed as Alfonso Jimenez Nunez registered achievement.biz, anonymous.biz, audiovideo.biz, and 2,291 other names.

Registrations that do not conform to .BIZ registration restrictions

The .BIZ TLD has as its primary registration restriction certain constraints on the warehousing and reselling of domain names. In particular, .BIZ Registration Restrictions define .BIZ to be used only for "bona fide business or commercial purposes" and define such purposes to exclude "selling, trading or leasing the domain name for compensation" and "the unsolicited offering to sell, trade or lease the domain name for compensation." Each registrant must affirm eligibility under these limits at the time of domain registration, and, as we discuss in Conclusions, a dispute resolution procedure may be invoked, at some expense, to challenge a registrant's usage of a given domain.

Using automated methods, we identified certain .BIZ restrictions that do not appear to conform to .BIZ registration restrictions. In particular, we identified a total of 4,886 distinct .BIZ domains that listed offers for sale of one or more registered .BIZ domains. We list apparently non-conforming domains alphabetically by domain name, by registrant, and by registrar.

Alphabetical listing of non-conforming domains
     A B C D E F G H I J K L M N O P Q R S T U V W X Y Z numbers
Non-conforming domains by registrant
Non-conforming domains by registrant, with listing of non-conforming domains
Non-conforming domains by registrar

These registrations seem to be clustered with a few large registrants, including Rarenames (2,645 registrations), Domains4days.com (518), "shoval - domains for sale" (193), and Excellentdomains.info (147). Apart from members of the public obtaining such registrations, certain registrars were associated with disproportionately many such registrations relative to their overall .BIZ market shares; such registrars include Domaindiscover (2,657 non-conforming registrations) and Enom (929). Among large .BIZ registrars, registrars with notably few such registrations (relative to their .BIZ market shares) include Network Solutions, Register.com, Tucows, and Melbourne IT. This may reflect simple registrar preference by those seeking such registrations, actual or perceived differential registrar enforcement of registry restrictions, varying registration costs, or some other factor.

Additional .BIZ summary statistics - registrant countries, registrar market share, distinct registrants

Available data allows straightforward tabulation of additional summary statistics about .BIZ registrations that may be of interest.

Registrant countries. Neulevel's web site reports a favorable reception to .BIZ outside the United States. According to a December press release, 50% of .BIZ registrations came from outside the United States; at least as of that date, Neulevel reports that a third of .BIZ registrations are in Europe and 11% are in the Asia/Pacific region. We report here the specific registrant countries as provided via Neulevel's WHOIS.

Count of .BIZ registrations by country

As Neulevel suggests, non-US registrations form a significant part of .BIZ registration counts; according to these results, 51.9% of .BIZ domains are registered outside the United States. Top registering countries include Germany (66075 domains, 9.7%), the United Kingdom (43213, 6.3%), South Korea (31099, 2.9%), Australia (19832, 2.9%), and Canada (19665, 2.9%).

Registrar market share. The authors' previous work has analyzed registrar market share in particular segments of the .COM, .NET, and .ORG registration market. (See Alternative Perspectives on Registrar Market Share.) With available data, it is straightforward to tabulate registrar market share within .BIZ registrations. For comparison purposes, the table below includes .COM, .NET, and .ORG registrar market shares as reported by the Q1 2002 issue of State of the Domain.

Registrar market shares within .BIZ

Comparing .BIZ registrar market shares with registrars' market shares in .COM, .NET, and .ORG shows notable divergences. Certain registrars have significantly larger shares of .BIZ than they do of .COM, .NET, and .ORG; such registrars include Schlund + Partner AG, Ascio Technologies, Yesnic, and Netnames. Others have a notably smaller share of .BIZ than of .COM, .NET, and .ORG; such registrars include Gandi, the Registry at Info Avenue, and Aitdomains.com

Distinct registrants. Analysis of Neulevel WHOIS data suggests that .BIZ domains have been registered by a total of 284,013 distinct registrants, where a distinct registrant is identified by a unique combination of registrant name, registrant organization, and registrant email address. Of these, 205,913 registrants registered exactly one .BIZ name, while 36,777 registered exactly two, 14,186 registered three, and 7,574 registered four. Additional tabulations of registration count by registrant are provided in the following table:

Number of .BIZ domains per registrant

Conclusions, Implications, and Future Work

We have studied .BIZ at an instant still early in its development. We expect to see change over time in the characteristics we have studied, and we expect that the direction and rate of such changes will prove important in evaluating the ultimate success and utility of .BIZ. In particular, we intend to watch the overall rate of growth of .BIZ, the rate of provision of substantive web content, and the rate at which .BIZ domains are registered by registrants of the corresponding .COM domains.

Comparison of registrants of .BIZ and .COM domains has shown that approximately 25% of .BIZ domains are likely registered by the same organization that holds the corresponding .COM domain. Further inspection of these names shows that the overwhelming majority are not being put to active use. Accordingly, we conclude that these domains are, by and large, "defensive" registrations, for which the registrants pay their respective registrars only to prevent other would-be registrants from using the names. If 25% of .BIZ's 680,000+ domains are defensive registrations, and if a .BIZ domain costs on average $20, the registration of defensive .BIZ domains yielded payments of $3.4 million from registrants to registrars and registries, apart from registrants' administrative overhead in performing and recording such registrations.

We have further identified a large number of domain names that appear, from web content and/or WHOIS data, not to conform to .BIZ registration restrictions. Neulevel offers a Restrictions Dispute Resolution Process ("RDRP") by which such registrations can be challenged. However, submission of a RDRP complaint costs $1,150 in arbitration fees (National Arbitration Forum) in addition to costs for preparation of the complaint. While large companies may find such fees practical in obtaining use (or others' non-use) of a desired domain, it seems unlikely that any ordinary Internet user would invoke such a procedure to challenge a domain warehouser's registration of a .BIZ name, especially since that user would himself be required to demonstrate compliance with .BIZ registration restrictions in order to receive the disputed domain at the conclusion of a successful RDRP challenge. In this regard, the .BIZ RDRP is similar to the .NAME ERDRP; as previously documented in .NAME Registrations Not Conforming to .NAME Registration Restrictions; the ERDRP has proven and is likely to continue to prove ineffective in limiting the non-conforming registrations within the .NAME TLD. If no effective restriction enforcement mechanism can be found, ICANN may need to reevaluate the addition of new TLDs intended for large-scale use that also include substantive registration restrictions.

This research is a portion of the authors' Evaluation of TLDs project.

* Jack N. and Lillian R. Berkman Assistant Professor of Entrepreneurial Legal Studies, Harvard Law School.
** J.D. Candidate, Harvard Law School, 2005.