What it is.
Session replay captures performance timings alongside the recording. You get First Contentful Paint, Dom Interactive, and Page Loaded for the pages a live visitor loaded, on their own connection and device.
That pairing is what a synthetic speed test cannot give you. A lab score says the page is slow, and a replay shows the person who waited and then closed the tab.
The tutorial includes a cheat sheet grouping the usual causes into JavaScript, images, and server or network latency, so a slow page turns into a shortlist of things to change.
What you get.
- Performance timings recorded per session, taken from live traffic.
- First Contentful Paint, Dom Interactive, and Page Loaded for live visits.
- The replay of the session behind each slow load.
- A cheat sheet grouping causes into JavaScript, images, and latency.
- A way to prioritize by which slow pages people abandon.
How to set it up.
Enable session replay and confirm performance capture is on.
Wait for traffic across devices, since your own laptop on office wifi is the least useful data point.
Sort sessions by the slowest Page Loaded and open a handful of replays.
Watch what the user does while waiting, because abandonment is the number that matters.
Use the cheat sheet to classify the cause into JavaScript, images, or latency.
Make one change, then compare the same timing on the same page a week later.
Use cases
Tie load time to lost conversions
Read the timings per session on the pages that convert.
Measure on live traffic
Use what live visits recorded instead of a lab run.
Pick the page to work on
Sort by the slowest timings among pages that carry revenue.
Best for
Landing pages behind paid spend
You are paying for the click, so a slow load is money leaving on arrival.
When a lab score disagrees with reality
Field timings from live sessions settle the argument with engineering.
Mobile conversion gaps
Replays on the visitor's own device show what a desktop test never will.