M MurmTools

Hotjar on 118 sites: an average page, in average company

Hotjar is session replay and heatmaps. It records how visitors move, scroll, click and give up on a page, plays those recordings back for the site's own team, and stacks thousands of them into a heat map showing where attention went. Product and design teams install it to find out why a page is not working. This index found it on 118 of the 5,294 sites measured, the 35th-widest reach of the 88 companies broken out here. Its own numbers are ordinary in almost every direction — which makes the company it keeps, and the banner situation on those sites, the parts worth reading.

Every median lands in the middle of the index

The median site carrying Hotjar contacted 10 companies over 54 requests, 3 of them classified as tracking, with a median first third-party contact at 686 ms and 6 companies arriving inside the pre-consent window. The fastest first contact recorded in the group was 75 ms.

Against the 88 companies in this index those figures rank 36th, 27th, 37th, 38th and 39th. Five separate measures, all of them within a few places of the middle. That is unusual in itself: most entities here are extreme on something. A page carrying Hotjar is, by this corpus's standards, a completely typical instrumented page.

Google on 114 of the 118

Google was present on 96.6% of Hotjar's sites. Then LinkedIn on 26 (22%), jsDelivr on 23 (19.5%), Meta on 23 (19.5%), Amazon CloudFront on 22 (18.6%), the Meta Pixel on 20 (16.9%), amazonaws.com on 18 (15.3%), Cloudflare on 15 (12.7%), hs-banner.com on 14 (11.9%) and hs-analytics.net on 13 (11%).

The last two are HubSpot hostnames, and with LinkedIn near the top the shape of the list is recognisable: this is a business-to-business marketing stack. Ads on a professional network, a CRM watching form fills, a delivery network for the scripts, and session replay to see why the demo form is not being submitted. Hotjar is the diagnostic tool at the end of that chain rather than the reason any of it is there.

Two-thirds of these sites had no banner to answer

Consent software was detected on 40 of the 118. The IAB TCF framework was seen alongside on 17 of them, OneTrust on 15, CookieYes on 3, Usercentrics on 2, Cookiebot on 2 and Iubenda on 1 — sightings, not slices, since a site can carry more than one.

That leaves 78 sites with no consent interface the crawl could find. It is the single most important number for reading the next section, because on most of these pages there was never a question on screen in the first place.

105 sites where something arrived early

On 105 of the 118, at least one company was contacted before a banner could have been answered — 89%, ranking 70th of the 88 companies here. On the median site, 6 companies arrived in that window and 2 more appeared after it closed.

The corpus counts per site and does not name the arrivals, so it cannot say whether Hotjar itself was among the early ones on any given page. What it can say is that Hotjar was present, and that on nine sites in ten the gap was too. For a product whose job is recording what a visitor does, where the recording starts relative to that window is the question people usually want answered, and this measurement is deliberately not the thing that answers it.

The named sites

tailscale.com contacted 11 companies, all 11 inside the window, first contact at 407 ms, no banner detected. rackspace.com: 14 with 12. bidmachine.io: 22 with 13, the heaviest of the twelve. marca.com: 17 with 12, behind an IAB TCF banner. psu.edu: 5 and 5. jusbrasil.com.br: 7 with 6. pexels.com: 6 and 6 at 293 ms, behind OneTrust. spankbang.com: 13 with 7 at 267 ms.

The slower ones look different. fwmrm.net contacted 14 companies with 5 inside the window at 1,088 ms behind OneTrust; mediafire.com 6 with 3 at 1,173 ms; superhosting.bg 6 with 3 behind an Iubenda banner. mercadolivre.com.br is the only example with nothing at all in the window: 3 companies, first contact at 2,261 ms.

A university, a Brazilian legal database, a stock photo library, a VPN vendor, an ad exchange and a tube site. Session replay is not a category of website; it is a tool that any of them can buy.

Highest-ranked sites where this was seen
SiteCompaniesBefore consentFirst contact
mercadolivre.com.br302261 ms
mediafire.com631173 ms
pexels.com66293 ms
fwmrm.net1451088 ms
tailscale.com1111407 ms
bidmachine.io2213715 ms
marca.com1712590 ms
psu.edu55962 ms
spankbang.com137267 ms
rackspace.com1412597 ms
jusbrasil.com.br76922 ms
superhosting.bg63752 ms
Consent tools seen on these sites
ToolSites
IAB TCF consent tool17
OneTrust15
CookieYes3
Usercentrics2
Cookiebot2
Iubenda1

Seen on the same pages

Companies most often contacted by the same sites. Co-presence only — the crawl records what arrived, not who called whom.

Questions

What does Hotjar record?

It captures interaction with the page — pointer movement, scrolling, clicks, form behaviour — and reconstructs it as a replayable session for the site's own team, plus aggregate heatmaps. This index classifies it under session replay and heatmaps.

Did Hotjar start recording before consent?

This corpus cannot tell you. Early contact is a per-site count of companies arriving in the window, with no names attached, so the supportable statement is that Hotjar was present on 105 sites where such contact happened.

Why did only 40 of 118 sites show a consent banner?

Because a banner is a choice the site makes, usually driven by where its visitors are. The other 78 presented nothing the crawl could detect on a single cold load from one location.

Is a page carrying Hotjar heavier than average?

Not by these ranks. Ten companies and 54 requests on the median site place it 36th and 27th of the 88 companies in this index — squarely mid-table on both.

Check any page yourself MurmStack is the tool that produced these figures. Free, two permissions, no host access.
Add to Chrome — free