This post outlines some faulty decisions by new gTLD registries. The purpose is to guide future expansions and, hopefully, adjust some of the decisions that current registries have adopted, including demand prediction, pricing, marketing, doing good versus not doing good, and launch date.

Demand Prediction

Demand prediction helps not only in determining the economic viability of a gTLD but also in determining the registrar’s required capacity investments. Unfortunately, the people making the predictions have not relied on rigorous quantitative methods, as discussed below.

1. There is no indication that they used any quantitative demand forecasting techniques, as some gTLDs have experienced very low weekly registrations. They could have used prediction markets as suggested in my CircleID post. They could also have used .com-based statistical pricing models for the selected gTLDs, which would have provided better results than going by gut feeling.

2. Some registries seem to have ignored the fact that every business needs future growth opportunities to survive. No matter the size of the initial market, growth stops when new territory stops opening up. Factor in nonrenewals and a “trapped” business would end up with negative growth and mounting operating expenses. No business can survive that way.

3. Some registries seem to believe that supply creates its own demand. For users to choose a new gTLD, or a new anything, they need to have reasonable confidence of a big leap in benefits. Incremental benefits don’t lead to switching, and new registries have to look hard at what they’re offering customers. Are these benefits that change minds?

Pricing

There seems to be a lack of effort to investigate the best pricing mechanisms. Some registries have used English auctions (ascending, à la eBay), others have used Dutch auctions (descending), and some adopted preset fixed prices. But no consensus has emerged, and there may be some level of unspoken enthusiasm for being different rather than choosing the best pricing option for the registry.

Marketing

1. Diluted message

1. A number of registries go wide with the messages that promote their gTLDs. They assume that each gTLD has the same audience and that a broad, unspecific message will do the job. Mistake. They must sharpen the target audience by conducting focused surveys and using analytical tools that also take data from other registrations. They should have started a long time ago, but it’s not too late! Nebulous, diluted messages are most often found among registries for generic words, but some location gTLDs have also fallen into this trap. Consider this sales pitch: “.city offers uptown web addresses to businesses looking for exposure, as well as area bars, event calendars, new outlets and more. It’s also great for any blogger who writes about city housing, transportation or apartment life.” Is there any city entity that would not be jumping to register names under the gTLD, thus eliminating any differentiation within a city and across different cities?
2. The “elevator pitch” Whatever happened to it? Some of the registries’ product descriptions go on and on.
3. Unenticing display of price. Here’s an example of the clever thinking at work. One registry displayed its registration price as $12,524.99. Wow! I would pay one whole penny under $12,525. Yay!

2. Performance guarantees

With so many registrations doing less well than expected, should such registries offer performance guarantees? A registry, for example, might promise a business that the adoption of its gTLDs would increase traffic by x%.

Having a number of different gTLDs, registries are in a position to offset
performance by domain names that don’t meet expectations. They can offer guaranteed returns, which will reduce adoption risk to businesses and increase the likelihood of registrations. The registries can diversify further by playing up Web hosting as a sideline. If they don’t mind taking on the need to monitor traffic volumes, registries will find that Web hosting shares a number of capabilities with their registry operation. And having two products to offer acts as a hedge to a registry’s business model risk.

Doing Good vs. Not Doing Good

Doing good can create competitive advantage, as noted in my CircleID post. Below are some areas of concern and suggestions.

1. Control registrations so that only gTLD-relevant domains are allowed to be registered. For example .Vegas “domain names are available and accessible to all who wish to become part of the Vegas family” without having to prove business association with Las Vegas. That can end up confusing users who want to go to find information about the city itself.

2. Realize that the distribution of free domain names by some registries casts doubt about the volume of registrations, which can negatively affect potential users’ decisions to register any of the new gTLDs.

3. To make the process easier, registries should integrate registration requests with a trademark database such as the Trademark Clearinghouse (TMCH), as suggested in an earlier post. Currently, a gTLD owner has to apply for the inclusion of their trademark in the database independently of any domain name registration.

1. The cost to enroll is $150 per trademark, and proof of ownership (such as a trademark registration) is required.

1. Some registrars such as Donuts, which applied for more gTLDs than any other entity, are offering additional protections to trademark owners, such as blocking services.

Registries should exert serious effort to prevent mass parking pages that have dominated use of new gTLDs.

Launch Date

GTLDs that have no competition do not need to race to launch. They have the luxury of waiting to better assess prices and postpone investments in capacity. However, this does not mean that they should not start immediate marketing of their gTLDs.