Practical guide
What Is Session Replay Used for on a Website?
Session replay reconstructs a visitor journey: pages viewed, scrolls, clicks, backtracking and successive interactions.
The point is not to watch random visitors. It is to understand how a blocker or hesitation builds up over time.
What it really is
Session replay reconstructs interactions, it does not film your visitor
A session replay tool rebuilds what happened in the browser from the pages displayed and the recorded interactions: clicks, scrolls, page changes, events or errors.
You observe the journey in the interface. You do not watch a video recorded on the user's device.
What is replayed
- page displayed
- clicks
- scroll
- navigation
- events
- errors if the tool captures them
What it is not
- a webcam
- a physical recording of the person
- proof of their intention
The sequence
What a replay shows that aggregated numbers do not
Moment 1
Hesitation
Does the visitor know what to do next?
Backtracking, repeated movements between two areas or a long pause can signal that a step needs more attention.
Moment 2
Blocker
At what point does the journey break?
A field that causes repeated attempts, a button with no reaction or a visible error give a much more precise lead than a simple drop-off rate.
Moment 3
Sequence
In what order do the actions occur?
Replay shows the sequence: reading, scroll, click, back, form, drop-off. This order can explain why an aggregated metric seems unusual.
The problem is sometimes only visible when you watch the complete sequence
In this example, the visitor reaches the form, goes back to proof points, tries again, then leaves the page after an error.
- 00:08
Arrives on the service page
- 00:31
Scrolls down to references
- 00:54
Scrolls back up to the offer
- 01:22
Opens the form
- 01:36
Corrects the same field twice
- 02:14
Leaves without submitting
Analytics can record a visit without a conversion. Replay here shows the sequence that deserves examination.
Example · No real data
Limits
An individual journey is not a general conclusion
Limit
One replay ≠ a trend
A session can be interesting without being representative. You need to look for behaviours that repeat before drawing a general conclusion.
Limit
One behaviour ≠ a certain intention
A visible hesitation does not tell you exactly what the person was thinking. Replay shows a sequence, not its motivation.
Limit
One spectacular case ≠ the priority
A very chaotic session can be rare. Analytics, funnels or segments are used to verify how many visitors actually encounter the same problem.
Levels of reading
Three levels of reading
Analytics, heatmap and session replay answer different questions. They complement each other, they do not compete.
Level
Analytics
How many?
Aggregated measurement: visits, exits, events, conversions.
Level
Session replay
How?
Individual sequence of interactions over time.
Filter before watching
Do not watch fifty replays at random
Start with sessions where a behaviour is already suspect or a metric is unusual.
- 1
an expected conversion did not occur
- 2
a technical error was raised
- 3
multiple repeated clicks appear
- 4
the visitor returns several times to the same step
- 5
a form is started then dropped
- 6
a behaviour differs strongly between mobile and desktop
- 7
a traffic source or campaign deserves examination
Replay becomes much more useful when you already know which segment or which problem you are trying to understand.
Plan before installation
A replay tool must be configured before collection starts
Forms, personal data and other sensitive information should not become visible simply because a replay tool is installed.
Tool selection, field masking, collection rules, consent where required and retention period must be examined in the actual site configuration.
This page explains the principle of session replay; it does not replace a legal or compliance analysis adapted to the project.
From observation to fix
Seeing a blocker is not enough: you need to verify what causes it
01
Identify
Identify a sequence that deserves examination.
02
Verify
Check whether the same behaviour appears in other sessions or in aggregated data.
03
Correct
Fix the likely cause: interface, content, field, validation, navigation or bug.
04
Measure
Check whether the behaviour and the relevant metric change after the modification.
Technical intervention
You already have replays, but nothing changes?
A marketing team may already have identified friction: a form that blocks, a step that is too long, a recurring error, a mobile interaction that is poorly understood.
Atomic Design can intervene on the technical side: instrumentation, journey correction, interface changes, event integration or development required to test the hypothesis.
You keep the strategy and marketing analysis. Atomic intervenes on what needs to be built, corrected or connected.
Related pages
Pages liées
Quelques pages utiles pour prolonger la lecture selon votre besoin actuel.
What Is a Heatmap Used for on a Website?
See where visitor behaviours concentrate on a page.
Open page →What Is an A/B Test Used for on a Website?
Verify whether a change actually improves the measured outcome.
Open page →How to Track Conversions Correctly with GA4 and GTM
Understand what is measured as events to interpret a replay correctly.
Open page →Practical intervention
You see where the journey blocks, but not yet what to correct?
We can compare replays against available data, identify the technical point to verify and intervene on the relevant page or journey.