jQuery, and the 2,418 sites in this index that still load it
jQuery is a JavaScript library from 2006 that papered over the differences between browsers: selecting elements, handling clicks, animating things, making background requests. Most of what it did has since been absorbed into browsers themselves, but it is still loaded on 2,418 of the 5,294 sites measured here — 45.7 percent, the second most common of the thirteen technologies this index tracks. It sends nothing anywhere on its own. It earns a page because its presence is a reliable marker of when and how a site was assembled.
A technology this common cannot be a signal
When something sits on nearly half the sample, its averages converge on the sample's own averages, and that is roughly what happens. The median jQuery site contacts 5 companies, carries 1 company classified as tracking, and makes 27 requests. Against the other twelve technologies indexed, that is ninth of thirteen on median company count and tenth of thirteen on median requests — the lighter half. Knowing a site loads jQuery tells you almost nothing about what else it loads. That is worth stating plainly, because the opposite assumption is common: old library, therefore neglected site, therefore junk in the page. The measurements do not support it.
Early contact on 2,174 of them
On 2,174 of the 2,418 sites — 89.9 percent — at least one company was contacted before a consent banner could have been answered. That is tenth of thirteen technologies on this measure, so most of the technologies indexed here have a higher share than jQuery does. The median count of companies contacted inside that window is 4, and the median count contacted after it closes is 0. Median time to first third-party contact is 689 ms; the earliest single site managed 55 ms.
Banners, on about a third
A consent banner was detected on 851 of the 2,418 sites. Seen alongside jQuery in that group: OneTrust on 346 sites, an IAB TCF consent tool on 331, Cookiebot on 55, CookieYes on 43, Osano on 32, Usercentrics on 28. These counts are not slices of one pie — a site can carry more than one banner, and some do — so read each as its own count rather than a share of the 851.
The high-ranked end of the sample is Microsoft
Order the jQuery sites by popularity rank and the top of the list is striking: microsoft.com, office.com, live.com, azure.com, sharepoint.com, outlook.com, cloud.microsoft and windows.com all appear, alongside amazonaws.com, roblox.com, cloudflare.net and baidu.com. These are not neglected pages. Their numbers vary widely: azure.com contacted 9 companies with 5 of them inside the window and a first contact at 1,275 ms, while windows.com contacted 7 companies with none inside the window and a first contact at 2,598 ms. baidu.com is similar — 2 companies, none early, 2,397 ms. Whatever those sites are doing with their third parties, the library they use to bind a click handler is not the variable.
Reading a single page
None of the figures above transfer to a specific site. A median is a description of a group, and the group here is wide enough that individual sites inside it look nothing like each other. If the question is what one page loads and when, that has to be measured on that page. MurmStack does this in the browser for the page you have open.
| Site | Companies | Before consent | First contact |
|---|---|---|---|
| microsoft.com | 4 | 3 | 816 ms |
| amazonaws.com | 6 | 4 | 1021 ms |
| office.com | 8 | 4 | 515 ms |
| live.com | 2 | 2 | 1016 ms |
| azure.com | 9 | 5 | 1275 ms |
| sharepoint.com | 6 | 2 | 1759 ms |
| outlook.com | 3 | 3 | 474 ms |
| cloud.microsoft | 5 | 4 | 1606 ms |
| roblox.com | 3 | 1 | 685 ms |
| cloudflare.net | 4 | 2 | 343 ms |
| windows.com | 7 | 0 | 2598 ms |
| baidu.com | 2 | 0 | 2397 ms |
| Tool | Sites |
|---|---|
| OneTrust | 346 |
| IAB TCF consent tool | 331 |
| Cookiebot | 55 |
| CookieYes | 43 |
| Osano | 32 |
| Usercentrics | 28 |
The six most common are listed. The remaining 16 of 851 sites with a banner carry something outside that six. One tool is recorded per site, so these counts never overlap.
Questions
Would removing jQuery change any of these numbers?
Nothing in this index connects the library to third-party contact. The medians describe the sites that happen to load it, and those sites make their own separate decisions about analytics, advertising and embeds.
How is a site counted as running jQuery?
By detection during the crawl of the measured page. A site is counted once regardless of how many jQuery files it loads or which version it uses.
Why do so many sites still load it?
This index cannot answer why; it only records that 2,418 of 5,294 measured sites did. What it can show is that the group includes large, actively maintained properties, not just old pages.
Is 689 ms fast or slow?
It is the sixth-highest median first-contact figure of the thirteen technologies indexed, so it sits in the slower half of that set. Within jQuery's own sample the range is wide: 55 ms at the earliest, and several of the highest-ranked sites past two seconds.