Quick answer: The site speed report in Google Analytics lives under Behavior > Site Speed. It shows how long pages take to load for your real visitors, broken down by page, browser, country and device, with timings for redirects, DNS lookup, server connection, server response and page download. By default Analytics only measures a 1 percent sample of pageviews, so on smaller sites the numbers can be thin or misleading. Raise the sample rate in your tracking code, look at pages with enough samples, and use the reports to find the slowest pages and the stage where the time is lost.
Why use Google Analytics for speed
Speed test tools load a page once from a test server. Google Analytics measures what happened in real visitors’ browsers, on their devices and connections. That makes it a useful complement to lab tools. It shows:
- Which pages are slow for real people, not just in a test
- Whether slowness is concentrated in a country, browser or device type
- Whether speed changed after a release, a new plugin or a hosting move
- How speed relates to bounce rate, conversions and revenue
If you have not set up Analytics properly yet, start with our guide to Google Analytics setup.
Where to find the reports
In Google Analytics, go to Behavior > Site Speed. There are four reports:
- Overview
- Page Timings
- Speed Suggestions
- User Timings
Overview
The Overview shows averages for the whole site over your chosen date range:
- Avg. Page Load Time (sec): from the start of navigation to when the page finished loading
- Avg. Redirection Time (sec): time spent on redirects before the page itself starts loading
- Avg. Domain Lookup Time (sec): DNS lookup time
- Avg. Server Connection Time (sec): time to set up the connection, including TLS for HTTPS
- Avg. Server Response Time (sec): time for your server to respond after the request, similar to time to first byte
- Avg. Page Download Time (sec): time to download the page’s HTML
You can switch the bottom table between browser, country and page.
Page Timings
This is the most useful report. It lists pages with their average load time and a comparison against the site average. Switch the view to Comparison to see bars showing which pages are faster or slower than average.
Tabs at the top let you look at:
- Explorer: the core metrics per page
- Technical: network and server timings per page
- DOM Timings: Avg. Document Interactive Time and Avg. Document Content Loaded Time, which describe when the browser could start working with the page
Add a secondary dimension such as Device Category or Browser to see where problems are concentrated.
Speed Suggestions
This report lists pages with their pageviews, average load time and a PageSpeed Insights score, plus a link to run PageSpeed Insights for each page. It is a handy bridge between real visitor data and the lab tool that suggests fixes. PageSpeed Insights then lists specific fixes, such as resizing images or reducing render blocking scripts.
User Timings
User Timings shows custom timings you send yourself, for example how long a key script, a product image gallery or a search request took. It is empty unless a developer adds the code. For most small sites it stays empty, and that is fine.
The sample rate problem
By default, Google Analytics collects timing data for only 1 percent of pageviews. On a site with 3,000 pageviews a month, that is about 30 samples. One slow visit on a poor mobile connection can drag the average up dramatically.
You can raise the sample rate in your tracking code.
With gtag.js:
gtag('config', 'UA-XXXXXX-1', {
'site_speed_sample_rate': 100
});
With analytics.js:
ga('create', 'UA-XXXXXX-1', 'auto', {'siteSpeedSampleRate': 100});
With Google Tag Manager, set the siteSpeedSampleRate field in your Google Analytics settings variable or pageview tag, under Fields to Set.
A value of 100 asks Analytics to measure up to all pageviews. Google still limits how much timing data it processes on high traffic properties, but smaller sites will see far more samples than before.
Reading the data correctly
Watch the sample count
Build a simple custom report that includes the metric Page Load Sample next to each page. Ignore pages with only a handful of samples. Averages from three visits are noise.
Averages hide outliers
An average load time of 6 seconds might mean most visitors waited 3 seconds and a few waited 40. Build a custom report with the Page Load Time Bucket (sec) dimension to see how timings spread across visits.
Segment by device
Desktop and mobile behave very differently. Apply the built in Mobile Traffic and Tablet Traffic segments, or add Device Category as a secondary dimension. Our article on mobile page speed covers the usual mobile problems.
Compare time periods
Use date comparison to check whether a change made things better or worse. Compare full weeks, so weekday and weekend patterns do not skew the result.
What the timings tell you
Each timing points to a different kind of fix:
- High redirection time: chains of redirects, such as http to https to www. Link directly to the final URL and fix redirect chains.
- High domain lookup time: slow DNS hosting. A faster DNS provider can help.
- High server connection time: network distance and TLS setup. A CDN can bring servers closer to visitors.
- High server response time: slow hosting, missing page caching or heavy database queries. See our guide to server response time.
- High page download time: very large HTML documents
- High page load time but normal server timings: heavy images, scripts, fonts and third party tags on the page itself
Building a speed dashboard
To keep an eye on speed without digging through reports each week, create a small custom dashboard with widgets for:
- Average page load time by device category
- Slowest pages with at least a minimum number of samples
- Server response time trend over time
- Load time for your key landing and checkout pages
Our article on Google Analytics dashboards covers building one. You can also set up custom alerts that email you when average server response time rises sharply.
Connecting speed to results
Speed matters because it affects behaviour. In Analytics you can look at speed alongside:
- Bounce rate for landing pages with slow load times
- Goal conversion rate by page load time, using a custom report
- Ecommerce conversion rate for faster and slower sessions
Treat these as signals rather than proof. Slow pages may also be different in other ways, such as attracting different visitors.
FAQ
Why does my site speed report show no data?
Usually because traffic is low and the default 1 percent sample produced few or no measurements. Raise the sample rate and wait a few days.
Why are Google Analytics load times higher than speed test tools?
Analytics measures real visitors on real devices and connections, including slow mobile networks and distant countries. Test tools usually run on fast servers in fixed locations.
Does a higher sample rate slow down my site?
No. The timing data is collected by the browser and sent as a small hit, which has no noticeable effect on load time.
Which pages should I fix first?
Start with pages that have high traffic, enough samples and slow load times, especially landing pages and checkout or contact pages.
Know where the seconds go
The Site Speed reports show how fast your site really is for the people using it, and which pages and stages are worth fixing. With a higher sample rate and a little care reading averages, they become a reliable early warning. If you want reporting that shows what is working and what is holding you back, see our Know what’s working service.
