HR Tech Website Speed and Core Web Vitals Statistics 2026: 49 Vendor Sites on Real-User Data

Only 9 of the 49 HR tech websites we checked passed all three Core Web Vitals for phone visitors in September 2026, according to Chrome UX Report data read through Treo. That is 18%, against 52.8% of all mobile origins in the HTTP Archive's technology report for the same month. Responsiveness is the weak spot: 27 of the 49 sites missed the Interaction to Next Paint target, and only 7 of 22 HRIS and payroll sites met it.

No public dataset labels a website as HR tech, so this page combines two sources. The first is our own sample of 50 named HRIS, payroll, recruiting, performance and learning vendors, with Chrome's real-user field data for the 49 that had it. The second is the HTTP Archive's data for the job-board and HR widgets those vendors ship onto their customers' websites, which is where an HR product's own code meets the open web.

The thresholds, and what the field data covers

Google judges a page on three field measures from real Chrome users: Largest Contentful Paint (LCP) for loading, Interaction to Next Paint (INP) for responsiveness and Cumulative Layout Shift (CLS) for visual stability. A page passes when 75% of visits meet all three targets, measured separately for mobile and desktop (web.dev).

Google says Core Web Vitals are used by its ranking systems, but that a good report "doesn't guarantee" top rankings (Google Search Central, updated 2026-09-22). Three details of the data matter for an HR tech team.

iPhone and Edge users are missing. The Chrome UX Report counts Chrome on Android, Windows, macOS, ChromeOS and Linux. It excludes Chrome on iOS, Android WebView and other Chromium browsers such as Microsoft Edge (Chrome for Developers, CrUX methodology). A buyer reading on an iPhone or in Edge on a work laptop is not counted.

Small sites may have no score. Origins must be "sufficiently popular", and Google does not disclose the minimum. Search Console falls back to the origin when a group of URLs lacks data, and shows nothing when the origin lacks it too (Google Search Console Help). One vendor in our list, Engagedly, returned no data at all.

A failing metric fails the page. Search Console rates a group of URLs by its worst metric over 28 days, so a site with fast loading and slow interactions is still marked as failing.

How 49 HR tech websites score on real-user data

We listed 50 HR software vendors across three segments and read each origin's September 2026 phone data from the Chrome UX Report through Treo's public site speed pages on 2026-10-28. A metric counts as good when at least 75% of phone visits fell in Google's good band, which is the same as a good 75th percentile. The raw file is saved as research/OB-0237-hrtech-crux-origins-2026-10-28.tsv.

HRIS and payroll sites load fast and respond slowly. Thirteen of 22 had good LCP, but only 7 had good INP. These are the largest vendors in the sample, with pricing configurators, country pickers, product tours and long navigation menus. Paychex's 75th percentile INP was 425 ms, Rippling's and Deel's 325 ms.

Performance and learning sites have a loading problem. Only 4 of 11 had good LCP, and Docebo, Leapsome, PeopleGoal and Betterworks were at 3.5 seconds or slower. Betterworks, Docebo and PeopleGoal each had good server response on fewer than 13% of phone visits, which leaves little time for anything else.

Recruiting sites are lighter. Five of 16 passed. The passing group, Breezy, Manatal, Eightfold, Teamtailor and Recruitee, all held CLS at zero and INP at 175 ms or less.

Layout shift is a quiet cause. Oyster HR, Jobvite, Greenhouse, TriNet and Quantum Workplace had a 75th percentile CLS of 0.2 or more. On HR sites this often comes from cookie banners, announcement bars and logo carousels that load after the hero.

HR tech visitors are mostly on desktop, and that changes the priority

The median site in our sample had 32.4% of its Chrome visits on phones, and 27 of 49 had a third or fewer. At Justworks, Gem, Leapsome, 15Five and SmartRecruiters the phone share was below 16%. Only four sites had more phone than desktop traffic: Deel, UKG, iCIMS and Eightfold, each with a large candidate or contractor audience logging in from a phone.

Because Chrome on iOS is excluded, the real phone share is higher than these figures. Even so, the pattern fits a buyer researching software at work. Two consequences follow.

Check desktop as well as phones. Google judges the two device types separately, and a site with 80% of its visits on desktop earns most of its evidence there. Phone scores are still the harder test, and they are the ones most HR tech sites fail.

Phone visits are often not buyers. On vendors whose customers' employees log in, the phone traffic is likely to be staff checking payslips or candidates applying. Where the product runs on the marketing origin, those flows set part of the score.

Platform and third-party scripts in the sample

From each homepage's initial HTML, nine sites run Next.js, nine WordPress and nine Webflow. None of the Next.js or WordPress sites passed; three of the nine Webflow sites did (Breezy, Manatal and Bonusly). Of the 16 sites where we could not detect a platform, five passed. We build marketing sites in Webflow and applications in Next.js, so weigh that, and treat nine sites per platform as far too few to rank platforms. Our Webflow page speed guide lists the fixes that move field data on that platform, and the broader platform pass rates are in our B2B SaaS website speed statistics.

The sharper split is consent tooling. Seventeen homepages load a consent manager in their first HTML response: OneTrust on seven, Cookiebot on four, Osano on three, and TrustArc, Usercentrics or Ketch on the rest. None passed. Of the 32 others, nine passed. This is correlational, and our fetch hit a bot wall or a thin shell on several sites, which may hide a consent script. A banner that renders late can still become the largest element on a phone screen or push the page down when it appears.

Your product's widget runs on customers' sites

HR software does not only live on the vendor's domain. Job boards, careers widgets and applicant forms are embedded on thousands of customer sites, and the HTTP Archive detects them. These figures cover every origin carrying each tool in September 2026, not the vendor's own website.

Most embeds sit close to the average. Nine of the 13 tools are within nine points of the 52.8% baseline either way. That suggests the embed itself is rarely the main cost on a customer's site.

The outliers are about who uses them. Hireology serves franchise and automotive retail businesses, and its customers' sites carry a median 5,131 KB of JavaScript; only 23.3% had good CLS. WP Job Openings sites had good server response on just 10.5% of origins, which points at slow hosting rather than the plugin. Personio's 70.5% may say as much about its customers' sites as about its code.

For an HR vendor, the useful question is what your careers widget adds on a customer's slowest page. A light, server-rendered job list helps your customers' scores; a script that injects the list after load can hurt their CLS.

What the wider web says about speed and leads

The HTTP Archive's 2025 Web Almanac, built on July 2025 data, found 48% of sites passing on mobile, up from 44% in 2024, and 56% on desktop. Two findings fit HR tech sites closely.

Busy sites struggle with INP. Good INP on mobile ran at 63% for the top 1,000 sites, 56% for the top 10,000 and top 100,000 and 77% across all sites. Home pages did better than inner pages, 80% against 69%. Pricing calculators and product tours sit on inner pages.

Most mobile LCP elements are images. An image was the LCP element on 76.0% of mobile pages, and about 17% of pages lazy-loaded it, delaying the element the metric measures. HR tech heroes are often large product screenshots.

Google's INP guidance splits each interaction into input delay, processing duration and presentation delay, and recommends breaking long tasks up and yielding to the main thread (web.dev, updated 2025-09-02).

There is no published controlled study of speed on an HR software site that we could verify. The nearest proxy is the "Milliseconds Make Millions" study by Deloitte and 55 for Google, which followed 37 brand sites and over 30 million sessions at the end of 2019. On lead-generation sites, a 0.1 second gain moved progression from the first form step to submission by 21.6%. A demo request form is a lead-generation funnel, so read that as direction, not a forecast. Our website speed statistics roundup collects the other load-time studies.

What this means for an HR tech website: a five-page test order

Test in the order that buyers move through the site, on desktop and phones, with field data where it exists and a lab run where it does not.

Before the first test, list every third-party script with an owner and a reason, and keep the logged-in app on its own subdomain so payslip checks and job applications do not set the marketing score. On a four-week marketing-site build we would write an LCP of 2.5 seconds and an INP of 200 milliseconds for these five pages into the brief, and check field data monthly after launch. Our marketing websites service page explains how we scope an HR tech site.