Session recording

Also known as: session replay

On this page 7

Session recording, also called session replay, is a behavioral analytics technique that captures how individual visitors interact with a website, including page changes, mouse movement, scrolling, taps, clicks and form interactions, and reconstructs each visit as a watchable playback. In E-commerce conversion work, it shows the qualitative reasons behind the numbers in a conversion funnel: why a shopper hesitated on a product page, failed to apply a coupon or abandoned the checkout.

Session recording does not explain behavior on its own, but it turns an abstract drop-off rate into observable user friction that designers, developers and conversion specialists can diagnose and fix.

How session recording reconstructs a visit

Modern tools do not capture video of the screen. A JavaScript snippet in the page records an initial snapshot of the Document Object Model (DOM) and then logs every subsequent change and user event with a timestamp. The replay player rebuilds the page from these records, which makes the data far smaller than video and allows search and filtering by events.

Because the replay is rebuilt from page structure, visual fidelity depends on whether stylesheets, fonts and images remain available. Content rendered inside canvas elements, cross-origin iframes or third-party payment widgets is often missing or shown as a blank area. Native mobile apps use separate SDKs that capture view hierarchies or periodic screenshots.

The table below lists the main data types a recording usually contains.

Data typeWhat is loggedDiagnostic value
DOM mutationsElements added, removed or changedShows what the visitor actually saw
Pointer and touch eventsMovement, clicks, tapsReveals hesitation and misclicks
Scroll depthVertical position over timeIndicates whether content was reached
Input eventsFocus, typing, field changes, usually maskedExposes form friction
Console and network errorsJavaScript exceptions, failed requestsLinks behavior to technical bugs
Session metadataDevice, browser, viewport, referrerAllows segmentation
Data captured in a typical recording

Behavioral signals worth filtering for

A busy store can generate tens of thousands of recordings per day, so nobody watches them at random. Analysts filter by funnel step, segment or automatically detected frustration signals, then watch a focused sample until patterns repeat.

SignalTypical definitionWhat it often indicates
Rage clickSeveral rapid clicks on one spotUnresponsive button or element that looks clickable
Dead clickClick with no page responseMisleading design or broken link
Error clickClick followed by a JavaScript errorFront-end defect
U-turnLeaving a page and quickly returningMissing information or confusing navigation
Thrashed cursorErratic, fast pointer movementConfusion or slow loading
Form field refillRepeated edits to the same fieldUnclear validation or format rules
Common frustration signals

Exact thresholds differ between tools, so the same visit can be tagged differently. Signals are prompts for investigation, not proof of a problem; a burst of clicks may simply be a user quickly selecting product variants.

Session recording compared with other conversion research methods

Session recording sits between quantitative analytics and moderated usability testing. It shows real behavior at scale but cannot reveal intent directly, so it works best in combination with neighboring methods in conversion rate optimization.

MethodAnswers the questionMain limitation
Web analyticsHow many visitors drop off and whereNo visibility into why
Session recordingWhat individual visitors did on the pageIntent must be inferred
HeatmapsWhere visitors click and scroll in aggregateLoses sequence and context
Usability testingWhy participants act as they doSmall, artificial samples
A/B testingWhich variant converts betterRequires a hypothesis and traffic
On-site surveysWhat visitors say about their experienceSelf-reported and biased by response rate
How session recording relates to other methods

The strongest workflow uses analytics to locate a leaking funnel step, recordings to form a specific, testable hypothesis, and an A/B test to confirm that the fix actually improves conversion.

Session recording processes personal data and behavioral data, so it falls under privacy laws such as the GDPR in Europe and the CCPA in California, as well as rules on storing information on user devices. Requirements differ by jurisdiction, and legal interpretation of replay tools is still developing.

  • Input masking by default: names, addresses, emails and passwords should never be stored in readable form.
  • Payment data must never be recorded; card fields fall under PCI DSS obligations and usually sit in provider iframes.
  • Consent management: in many regions recordings start only after the visitor accepts analytics cookies.
  • Retention limits: recordings should be deleted after a defined period, often 30 to 90 days.
  • Disclosure in the privacy notice, including the vendor that processes the data.

Teams also need element-level exclusion for account pages, order histories and support chats, because personal details often appear outside form fields as ordinary page text.

Practical setup and common mistakes

Good programs start with a narrow question, such as why mobile shoppers abandon the shipping step, rather than recording everything and hoping insights appear. Sampling a share of sessions reduces cost and performance impact while still showing repeated patterns.

MistakeConsequenceBetter practice
Watching random sessionsHours spent with no conclusionsFilter by funnel step and signal
Generalizing from one replayFixes based on outliersConfirm patterns across many sessions
No masking reviewSensitive data storedAudit masking after every site release
Ignoring performance costSlower pages and lower conversionLoad the script asynchronously and sample
Skipping validationUnproven changes shippedTest fixes with A/B experiments
Frequent mistakes and how to avoid them

The idea in brief

Picture a shop owner who could quietly watch, afterward, how each customer walked through the store: which shelves they stopped at, where they looked lost and where they put the basket down and left.

Session recording gives an online store that view. It replays a visit like a short film, built from the clicks, taps and scrolling of a real visitor. Personal details are hidden. The value lies in seeing problems that numbers alone cannot show, such as a button that does not work on some phones.

A practical scenario

An online furniture retailer notices that checkout completion on mobile is about 15 percentage points lower than on desktop, while analytics only shows that most mobile visitors leave on the delivery details step.

The conversion team filters recordings to mobile sessions that reached that step and exited without ordering, then watches about 60 of them. A pattern repeats: the postcode field rejects entries with a space, the error message appears below the visible area, and visitors tap the continue button several times, producing rage clicks, before leaving.

Developers make the field accept both formats and move the message next to the input. A two-week A/B test confirms the change, with mobile checkout completion rising by several percentage points in the variant. The team then adds a saved filter for rage clicks on checkout pages to catch similar defects after future releases.