Fewer Than Half of University Websites Pass Core Web Vitals

Fewer than half of America's universities pass Google's Core Web Vitals on mobile. Across every .edu site Google has real-user data for, only 48.3% clear the bar, which puts higher education near the web-wide average and well behind the flagship-only picture you get from ranking a handful of famous schools. The institutions that teach computer science are, on the whole, running middle-of-the-pack websites.

That matters because the university website is now the front door to enrollment. Prospective students research programs, compare campuses, and start applications on their phones, and a page that takes eight seconds to load loses them before the pitch begins. Slow pages depress conversions, frustrate current students trying to reach a course portal, and cost search visibility, since Google treats loading, responsiveness, and stability as ranking signals. A campus can have a brilliant program and still lose the applicant at the loading screen.

These figures come from Google's Chrome User Experience Report, the real-user field data drawn from the Chrome browsers of actual visitors, the same signals Google's ranking systems use. We measured every university website Google has that data for, all 34,321 of them, roughly 700 times the sample of a ranked-50-schools study, so this is the real state of higher-education web performance.

Every origin on a .edu domain with field data qualifies, 34,321 in total, measured on mobile at the 75th percentile against Google's Core Web Vitals thresholds. The .edu top-level domain is used almost entirely by accredited higher-education institutions, so this is a near-complete population rather than a curated list. Passing means clearing all three metrics at once: good loading (Largest Contentful Paint under 2.5 seconds), good responsiveness (Interaction to Next Paint under 200 milliseconds), and good visual stability (Cumulative Layout Shift under 0.1), for the three-quarters of real visitors who define the typical experience.

The shape of higher education's performance is distinctive. Universities are excellent on the two metrics that are cheap to get right on a content-first site: 93.2% have good layout stability, the highest of any sector, and 89.3% have good responsiveness. Academic pages are text-heavy and stable, without the client-side application code that hurts responsiveness on retail and software sites, so they hold still and react instantly.

The weak point is loading. Only 74.7% of university sites have good LCP, and because a site must clear all three metrics to pass, that single number is what keeps the sector under 50%. The deeper reason is scale. A university is not one website, it is a federation of hundreds: the admissions site, the course catalog, individual department pages, faculty bios, library subsites, research-lab tools. Fully 26% of university origins have too little traffic to even report a responsiveness score, the highest share of any sector, a direct measure of how many small, lightly-maintained subsites hang off every campus domain. A fast homepage does not mean the university passes.

At the fast end, purpose-built platforms do well. Western Governors University, a large online school, loads in 1.4 seconds and passes, because its site was engineered as a single product rather than assembled from decades of departmental templates. At the slow end sits a recognizable outlier: the University of Colorado's PhET interactive physics simulations site takes 8.1 seconds, a heavy interactive tool loaded straight onto the homepage and roughly six times slower than the fast leaders.

Academia.edu fails at 2.8 seconds and Boston University's main site fails at 2.6. The failures are a mix of heavy interactive academic tools and older, template-driven campus sites that were never tuned for mobile loading. The common thread is that loading, not interaction, is what trips them, which is exactly the pattern the metric breakdown predicts.

Fast: Western Governors University (1.4s, pass). Slow: phet.colorado.edu (8.1s), Academia.edu (2.8s), Boston University (2.6s), all fail.

For a university web team the target is loading and nothing else. Layout stability is the highest of any sector at 93.2% good CLS and responsiveness sits at 89.3% good INP, so those are already handled. Only 74.7% of sites have good LCP, and that single metric is what holds the sector at 48.3%.

Scale is the real problem. A campus is a federation of hundreds of sites, and 26% of .edu origins carry too little traffic to even report a responsiveness score, the highest share of any sector. You cannot tune all of them, so raise the pages the most students actually hit: the homepage, admissions, the course catalog, and the learning-management entry points. Compress and lazy-load imagery, defer heavy interactive widgets so they do not block the first paint, and put the busy pages behind a CDN.

The named examples mark the range. Western Governors University loads in 1.4 seconds and passes because it was engineered as one product, while the University of Colorado's PhET physics tool takes 8.1 seconds loaded straight onto its homepage, and Boston University fails at 2.6. The payoff is measured in applications completed, because the applicant comparing schools on a phone rewards the pages that appear first.