.com: the third-party load on 2,831 measured sites
A top-level domain is the last part of a web address, and .com is the one most sites in this index use: 2,831 of the 5,294 measured, or 53.5 percent. That makes this page less a profile of a category than a look at the baseline. Whatever counts as ordinary on the open web shows up here in bulk, so the medians below are close to the middle of the whole index rather than a departure from it.
It is the largest of the nine top-level domains here
Of the nine top-level domains broken out in this index, .com ranks first by number of sites measured, and it covers more than half of everything measured, 53.5 percent. Size changes how the rest of this page should be read. A median taken across 2,831 sites is steady; one taken across a few dozen moves if a handful of sites change. So when a figure below looks unremarkable, that is the finding: .com is where the index's centre of gravity sits, and other groups are best read as departures from it.
The median .com site: five companies, 31 requests
Half the .com sites measured reached five or fewer separate companies during the page load, and half reached more. The median count of requests was 31, and the median count of hosts classified as tracking was one. Against the other eight top-level domains here, five companies ranks fourth and 31 requests ranks third by size. That places .com in the middle on company count and towards the busier side on raw request volume, which is roughly what a mixed population of shops, publishers, corporate sites and tools produces once averaged together.
Contact starts at 652 milliseconds, and mostly before a choice is possible
On 2,618 of the 2,831 .com sites, at least one company was contacted before the point where a consent banner could have been answered. That is 92.5 percent; 213 sites had no such early contact. The median time from navigation to the first third-party request was 652 milliseconds, and the quickest observed was 55 milliseconds. The median .com site had reached four separate companies inside that early window, which ranks second of the nine top-level domains here, level with .ru and .edu.
After the window, the median site adds nobody
The median number of companies first seen after the consent window closes is zero on .com. Read alongside the four already reached inside the window, that says something simple about how these pages are built: the third-party set is essentially settled by the time the question of consent arises, not assembled afterwards in response to an answer. Individual sites vary widely, and this is a median, not a rule. But it is the reason a screenshot of a page after clicking accept tells you very little about what has already happened.
Fewer than half carry a detected consent tool
A recognisable consent tool was detected on 1,277 of the 2,831 .com sites, 45.1 percent. OneTrust was seen on 644 of them and an IAB TCF consent tool on 429, with Cookiebot on 66, CookieYes on 46, Osano on 45 and Usercentrics on 35. These are not slices of the 1,277: a site can carry more than one banner, so the counts are best read as how often each tool was seen alongside a .com site, not as shares that divide up the total.
Twelve .com sites, one at a time
Medians hide the spread, so here are individual results. On netflix.com six companies were reached and all six arrived inside the early window, on a page carrying a OneTrust banner. On azure.com nine companies were reached, five of them early. office.com reached eight, four early, with first contact at 515 milliseconds. Small sets exist too: github.com reached one company at 393 milliseconds, twitter.com one, and linkedin.com one with none inside the early window and first contact only at 2,536 milliseconds. cloudflare.com reached two, microsoft.com four, amazon.com three at 462 milliseconds, live.com two, bing.com two, and amazonaws.com six.
| Site | Companies | Before consent | First contact |
|---|---|---|---|
| cloudflare.com | 2 | 2 | 1242 ms |
| microsoft.com | 4 | 3 | 816 ms |
| amazonaws.com | 6 | 4 | 1021 ms |
| twitter.com | 1 | 1 | 814 ms |
| linkedin.com | 1 | 0 | 2536 ms |
| office.com | 8 | 4 | 515 ms |
| live.com | 2 | 2 | 1016 ms |
| amazon.com | 3 | 3 | 462 ms |
| azure.com | 9 | 5 | 1275 ms |
| bing.com | 2 | 1 | 1540 ms |
| github.com | 1 | 1 | 393 ms |
| netflix.com | 6 | 6 | 1003 ms |
| Tool | Sites |
|---|---|
| OneTrust | 644 |
| IAB TCF consent tool | 429 |
| Cookiebot | 66 |
| CookieYes | 46 |
| Osano | 45 |
| Usercentrics | 35 |
The six most common are listed. The remaining 12 of 1277 sites with a banner carry something outside that six. One tool is recorded per site, so these counts never overlap.
Questions
How many .com sites are in this index?
2,831, out of 5,294 sites measured in total. That is 53.5 percent, and it makes .com the largest of the nine top-level domains broken out here.
Do .com sites load more third parties than sites on other domains?
Not by this measure. The median .com site reached five separate companies, which ranks fourth of the nine top-level domains in this index. On median request count, 31, it ranks third.
How fast does the first third-party request happen?
The median .com site made its first third-party request 652 milliseconds after navigation. The quickest measured was 55 milliseconds, which is before a typical page has finished painting.
Which consent tools turn up on .com sites?
OneTrust heads the list at 644 sites, followed by an IAB TCF consent tool at 429. Cookiebot, CookieYes, Osano and Usercentrics each appeared on fewer than 70. Some sites carry more than one, so these counts overlap.
Can I check a single .com site rather than read medians?
Yes. MurmStack runs the same measurement on whatever page you have open, and reports the companies contacted and how early each one appeared.