Every published figure for “how big is .com” counts registrations — how many names exist. That is a number about a database. We wanted the other one: of the websites people actually open, what share carries a .com?
The Chrome UX Report publishes the origins that receive enough real Chrome traffic to be measured, ranked into magnitude tiers. The list holds 17,628,280 origins across 13,415,066 registrable domains. It is the closest thing to a census of the visited web. We took every origin in it and counted the extension.
The long tail is longer than the ranking suggests. Second place is .org at 3.4% — one seventeenth of .com. Two of the next four are country extensions, and .net, .com’s oldest sibling, is sixth.
The twenty most-loaded extensions
Share of all origins in the Chrome UX Report. One bar is one extension; length is share.
Source: Chrome UX Report top-origins list (17,628,280 origins). Rank tiers are CrUX magnitude buckets, not integer ranks.
The share barely moves
The story everyone tells about .com is a story about the top of the web: serious sites buy .com, and the cheap extensions fill the tail. It is a good story. The data does not support it.
Split the list by rank tier and .com’s share is close to flat. It is highest in the top 1,000 at 49.9%, drops into the low forties almost immediately, bottoms out at 40.8% in the 5M tier, and then climbs back to 43.7% in the very tail. Across four orders of magnitude of popularity, the whole range is nine points.
.com share by rank tier
Each tier holds the origins ranked up to that magnitude. If .com were a top-of-the-web phenomenon, this line would fall from left to right. It does not.
- .com
- .com + .net + .org
Source: Chrome UX Report top-origins list (17,628,280 origins). Rank tiers are CrUX magnitude buckets, not integer ranks.
The three original generic extensions together hold between 46.4% and 57.4% of every tier. Whatever the new extensions have taken since 2013, they have taken it evenly.
What does change down the ranks
.com is flat, so the movement is elsewhere. To find it we indexed each extension’s share of the top 100,000 origins against its share of the whole list. An index of 1.0 means an extension is spread exactly like the average; 2.0 means it is twice as concentrated among popular sites.
Concentrated at the top, and concentrated in the tail
Share of the top 100,000 origins ÷ share of all origins. Extensions with at least 20,000 origins.
- Over-represented among popular sites
- Under-represented among popular sites
Source: Chrome UX Report top-origins list (17,628,280 origins). Rank tiers are CrUX magnitude buckets, not integer ranks.
.in leads at 2.74 — India’s popular sites are far more visible than its long tail. .gov at 2.51 is the shape you would predict: a small set of extremely well-used sites and almost no tail at all, because you cannot register one on a whim.
The other end is more interesting. .de, .ch, .au and .nz all sit near 0.4. These are mature national extensions with enormous registered bases and very few globally popular sites — a national web rather than a global one. And .pm at 0.03 is not a national web at all: two addresses, mp3.pm and booth.pm, carry 27,419 of its 27,590 origins, which is a story of its own.
How this was counted
- The corpus. The CrUX top-origins dump, all ten magnitude tiers — 17,628,280 origins. Not the 1M-capped public dashboard.
- The unit. An extension here is the extension IANA delegates, so a .co.uk address counts under .uk. Of the 1,438 delegated extensions, 1,015 appear at least once.
- The 11 we dropped. Every origin except eleven resolved to a delegated extension. The eleven are .onion addresses, which IANA reserves and never delegates.
- Rank tiers, not ranks. CrUX publishes a magnitude bucket rather than a position, so “top 5k” means an origin somewhere in ranks 1,001 to 5,000. Every tier figure here is a tier, never an ordering inside one.