Domain market analytics

01.Where do these domain statistics come from?

From the zone files registries publish through ICANN’s Centralized Zone Data Service. A zone file lists every name in an extension that has nameservers, so it is the registry’s own record of what is registered and live. Each one is downloaded daily and compared with the day before.

example.com.      NS  ns1.example-dns.net.example.com.      NS  ns2.example-dns.net.example.com.      DS  370 13 2 BE74359954660069D5C6……                     one line per nameserver, for every delegated name

Most domain statistics are surveys, samples or projections. Counting the zone file itself means a small extension is measured as exactly as .com, and a figure here is a count you could reproduce from the same file, not a model of it.

Every extension has its own public page

Open daily headline figures and monthly reports for each observed zone.

02.What do delegated, newly delegated and churn mean?

Delegated is every name in the zone today. Newly delegated is what appeared since yesterday, left is what disappeared, net growth is the difference as a share of the extension, and churn is all movement in both directions. None of them is a count of purchases.

03.Does a domain that leaves the zone become available?

Usually not yet, and sometimes never. Most names leave as they enter the 30 day redemption period after expiring, and their owners can still restore them. A .com that nobody restores is released about 35 days after it left the zone.

That makes the daily number of names leaving an extension the most misread figure in the market. For what actually becomes available, read how many domains drop every day. Waiting on one particular name? Domain Hub reads its registry status and counts down to the likely release date.

04.How are DNSSEC adoption and DNS provider concentration measured?

DNSSEC adoption is the share of names publishing a DS record in the zone, which means they are signed. Provider concentration uses the Herfindahl Hirschman Index, the standard measure of market concentration, over which DNS provider answers for each name.

A high index means a few providers answer for most names, so one outage reaches a large share of the extension. The thresholds are the ones the US antitrust agencies publish for the Herfindahl Hirschman Index. For what a DS record does, see DNSSEC explained.

05.How are parked domains and first-year retention estimated?

Both are heuristics and are labelled as such. Parked share counts names delegated to known parking and marketplace nameservers. First-year retention is the share of names first seen a year ago that are still delegated today, the closest public signal to a first renewal.

Parked share is read from nameservers.

At the scale of a whole extension that is a dependable pattern, but not a verdict on any single domain: a parking page served from ordinary hosting is not counted.

Retention measures delegation, not billing.

A name can be renewed and still have no nameservers, or be in grace and still be delegated. Treat it as a strong proxy, not a registry’s renewal figure.

06.What do the registry reports add?

Exact counts the zone file cannot give: which registrar manages each name, how many names were renewed and how many were deleted. Every generic registry reports these to ICANN each month, and ICANN publishes them three months after the month ends.

The reports are defined in Specification 3 of ICANN’s base registry agreement: one row per registrar, with the names it manages, new registrations by term, renewals and deletions. The registry delivers them within 20 days of each month’s end, and ICANN withholds them from publication for three months, so the latest month shown is always about a quarter old.

Renewal rate.

Renewals divided by renewals plus deletions outside the add grace period (the first days after registration, usually five), over the latest three reported months. An extension needs at least 100 such names for a rate to be shown.

Renewal waves.

Names first registered for one year come up for renewal twelve months later. Each month ahead shows those names, and how many of them the extension’s recent renewal rate suggests will not be renewed. First renewals are usually below the average, so treat it as a lower bound.

Registrar share.

Names under management per registrar, as the registry counts them. The daily panels that infer registrars from default nameservers are labelled as estimates.

07.Which figures are forecasts or estimates?

Everything the zone files only imply rather than state. Each one is labelled on the dashboard, says how it was made, and is shown as a range or an upper bound where that is the honest reading.

Drop waves.

Names that leave a zone usually come free about 35 days later, after the 30 day redemption period and 5 days of pending delete. Owners restore some, so each day of the forecast is an upper bound.

Portfolios.

Blocks of at least 500 names on one set of nameservers that share a shape: the same length, character pattern and signing. Likely one owner, but described only by infrastructure, never tied to a person.

Names nobody is serving.

A weekly random sample of about 400 names per extension, each asked for its SOA record at up to two of its nameservers without recursion. Names whose nameservers do not resolve, do not answer, or answer without authority for the name are counted, with a 95% margin. The older term “lame delegation” is avoided because the IETF’s current DNS terminology calls it historic.

Search demand.

Each extension’s share of checks on axath against its share of new registrations. Counted per extension only, never per name or person, with a floor of 10 checks.

08.Why are some days flagged as unusual?

A registrar can suspend tens of thousands of names in an afternoon, or a registry can run a bulk clean-up. When a day moves more than four times the recent baseline it is flagged as unusual, so one event does not read as a shift in the market.

09.What does each plan include?

Everyone sees which extensions are covered and when the data was last updated. Pro adds the daily dashboard with registry reports, forecasts, portfolios and alerts for your own extensions. Scale adds the layer investors and registrars look for: provider market share and flows, parking, naming signals and retention.

Compare the plans on the pricing page.

10.What can zone file statistics not tell you?

Who owns a name, when it was bought, or what it costs. And nothing about country-code extensions, whose zone files are not distributed through ICANN.

Country-code extensions are not included.

Zone files for .de, .io, .co, .ai, .uk and every other country code are not distributed through CZDS, so they are left out entirely rather than estimated.

Delegated is not registered.

A registered name without nameservers, or on hold, is not in the zone. Totals here are live names, a little below total registrations.

Aggregate figures only.

The page publishes counts and rates, never lists of domains. The one exception is your own Hub: its lookalike watch shows you new names that resemble the domains you own.

11.Domain analytics questions

Short answers to what people ask most about domain counts and zone files.