Develop models for predicting domain name renewal rates

Co-authored by Rickard Vikström, Founder and Danny AertsSenior Advisor at DomainCrawler.

One of the biggest problems in the domain industry is how to accurately predict, year over year, the number of customers who will renew their domain names. It’s quite common for a registry in the first year to have a renewal rate of around 60%, and after the first year that rate increases to around 80-85% on the remaining domain names. But how can we better understand why this is so, how can it be more accurately predicted, and what can be done to help maximize these opportunities?

The important first year

The important thing for a registry is to survive the first year renewal. From the first year, the chances of the next renewal increase significantly, and after that, the domain name will most likely remain for a long time. For domains under management with a renewal rate of around 80-85%, which is at least typical of older registries, this is a good benchmark.

To accurately predict renewals, it is necessary to focus on months 6-11 of the first renewal period, as this is where the major churns are when a business decides if they want to keep the domain name. or not. It depends, of course, on the renewal period (i.e. a year or more), but it is the next renewal time that is important, and for a year, it is 6 months- 11. Interestingly, a total renewal rate above 85% may not be a good sign. There could be a problem because if the renewal rate is too high, it could mean that the business is not growing. Renewal rates which are generally high mean that there are only older domain names in the portfolio and not enough new sales are generated. Even if new sales lead to lower renewal rates in the first year, a balance must still be maintained for domains under management to grow. One of the most important things is to not stop selling, to remember to make sure that new accounts come into the zone and survive at least the first year.

Predict the probability of renewals

When trying to analyze the likelihood of renewal, a fairly accurate number can be obtained using historical registrar data, for example, analyzing a registrar’s typical performance during certain campaigns. If the campaigns and registrars are correctly monitored, it is possible to know exactly the number of sales made according to the campaign and to anticipate what percentage will be renewed in the future. It is also possible to calculate the lifetime value of a domain name that passes through certain registrars, which makes it possible to adapt the discounts accordingly and thus increase the probability of renewal.

For registrars that normally operate at lower levels, it is possible to set performance targets and incentives and measure this effect and adjust accordingly. Predicting future activity based on historical registrar data is fairly easy to do and can give a pretty clear picture of what is likely to happen. Moreover, depending on the amount of information obtained about the customers, it is also possible to see the services connected and the type of use of the domain, which can further improve the modeling process.

Domain data collection and analysis

To further analyze the activity of a domain, it is necessary to collect more data and then follow up on a daily basis. Some domain-based data is readily available, but domain-connected DNS, web, and binding data requires further investigation. To fully understand a domain, it is necessary to devote time to investigations and then to follow this activity on a daily basis.

A domain could be analyzed using a point-based scoring system, for example, based on some of the most important content and registration metrics. For example, what is the content of the site, how is the site used, is it e-commerce, are there other sites connected, links, is there an email address connected? All of these factors increase the likelihood that a site will stay in business for the longer term.

For registrations, assume a customer has a portfolio of domain names; it is a very important indicator of longevity. If there is a neighboring domain match, for example in Sweden, there may be a registry for .se and .nu domains. The latter is very strong in this country, and if both domain names exist, then it is very likely that a defensive registration has been made, probably because a company is already planning to keep the domain names long-term. The same with .ch and .at for .de for that region, or even if registrations have been made in neighboring countries with a similar language and culture, which could indicate the future direction of the business.

It is then possible to observe indicators or signals related to the domain name in order to perform a scoring analysis. For instance,

Low value indicators:

DNS server, A record, MX record – there are no dots for these because if you have a mail server connected you will have a DNS server.

High-value indicators indicate that there is a high probability of renewal of the domain name:

  • Content: is there a website linked to the domain name?
  • Usage: Are any services, like e-commerce or email, connected?
  • Encryption: is there a TSL/SSL certificate attached and if so, what type?

The full version of the scoring system for predicting domain name renewal rates is available on the DomainCrawler website via the link: Scoring system to predict domain renewals, DomainCrawler

Conclusion

By analyzing customer data and the use of a domain using such a system of indicators, as well as a complete picture obtained from all surveys, it is possible to predict rates quite accurately. renewal in the future. This information can then be used in business decision-making or deciding where to focus on building relationships and more wisely allocating resources and funds to increase sales and ultimately increase profitability. .