Session Replay
Session replay captures anonymized recordings of real user behavior — the qualitative layer that explains the "why" behind your GA4 numbers.
Session Replay
Anonymized recordings of real user sessions — clicks, scrolls, hesitations, rage clicks — used to diagnose UX and conversion issues.
Session replay is a behavioral analytics technique that captures a visitor's interactions with your site and plays them back as a video-like timeline. You see cursor movement, taps, scroll depth, field focus, page transitions and the exact sequence of events that led to a conversion — or a bounce.
Unlike heatmaps, which aggregate behavior across many sessions, replay shows you one human at a time. That makes it the qualitative half of behavioral analytics: GA4 tells you 38% of users drop on the shipping step; replay tells you they're tapping a greyed-out country selector that doesn't open.
Replay tools work by recording the DOM and user events client-side, then reconstructing the page in playback. Nothing is filmed — the player rebuilds the page using the same HTML and CSS your visitor saw, with PII automatically masked so passwords, card numbers and email fields never leave the browser.
On Shopify, WooCommerce and Magento stores, replay is most valuable on three surfaces: product detail pages (where hesitation predicts non-purchase), the cart-to-checkout transition (where unexpected costs cause exits), and post-purchase upsells (where mistapped CTAs kill AOV). Pair it with funnel analytics and you stop guessing why a step leaks.
Watch Budget (min/week) = Sessions Sampled × Avg Replay Length × Review Multiplier
Sessions Sampled
Sessions sampled per week
Number of replays you actually intend to watch — not record. Sampling is the whole game.
Avg Replay Length
Average replay length (minutes)
Typical session duration on your store. DTC averages 2-4 minutes; checkout-only filters are shorter.
Review Multiplier
Review speed multiplier
Most teams watch at 2x with skip-inactivity on, giving ~0.6x real time. Pure 1x watching = 1.0.
A Shopify apparel store wants a weekly replay review on cart abandoners.
Sessions Sampled: 40
Avg Replay Length (min): 3.5
Review Multiplier: 0.6
→ 84 minutes / week
Roughly 1.5 hours per week to review 40 high-signal abandoners — a sustainable cadence for one CRO owner. If the result exceeds 3 hours, tighten your filter (e.g. only sessions that reached checkout step 2).
Recording everything is a trap. The cost isn't storage — it's reviewer attention. Smart teams sample by intent signal: rage clicks, dead clicks, U-turns, sessions that hit checkout but didn't convert, or anyone who triggered an error event. That turns replay from a curiosity feed into a hypothesis engine.
Typical session replay signal rates by store type
| Store type | Rage-click rate | Dead-click rate | Avg session length | % sessions worth reviewing |
|---|---|---|---|---|
| Shopify apparel (€1-5M) | 2.1% | 4.8% | 2:40 | 3-5% |
| Shopify beauty (€1-5M) | 1.6% | 3.9% | 3:10 | 2-4% |
| WooCommerce electronics | 3.4% | 6.2% | 4:20 | 5-7% |
| Magento home & garden | 2.8% | 5.5% | 3:50 | 4-6% |
| DTC subscription (any platform) | 1.9% | 4.1% | 2:55 | 3-5% |
Privacy is non-negotiable. Modern replay tools mask input fields by default and let you flag custom selectors (loyalty IDs, gift messages) as sensitive. Confirm your vendor stores recordings in an EU region if you sell into the EU, and that masking happens client-side before data leaves the browser — not server-side after the fact.
Session Replay FAQ
Heatmaps aggregate behavior across hundreds or thousands of sessions into a single visual — useful for seeing patterns. Session replay shows you one user at a time, in sequence, so you can diagnose specific failures. They're complementary: heatmaps tell you where to look, replay tells you what happened.
A modern replay snippet adds 15-40KB gzipped and runs async, so the impact on Largest Contentful Paint is usually under 100ms. Older tools that record every mouse movement at high frequency can add noticeable load; check that your vendor throttles event sampling and lazy-loads on idle.
It can be, if input fields are masked client-side, recordings are stored in an EU region, and the user has consented under your cookie banner. Replay falls under behavioral analytics consent, not strictly necessary cookies, so you must gate it behind opt-in for EU visitors.
Recording is cheap; watching is expensive. Most CRO teams record 100% of eligible sessions but sample 20-50 per week for actual review, filtered by high-signal events like rage clicks, checkout exits, or errors. A weekly batch of 30-40 well-filtered replays beats 500 random ones.
A rage click is three or more rapid clicks on the same element, indicating the user expected something to happen and it didn't. It's the single highest-signal frustration event in replay — a 2%+ rage-click rate on a button or link usually points to a broken interaction or unclear affordance.
If your behavioral analytics platform already includes replay, yes — consolidating onto one snippet reduces page weight and cross-tool tagging headaches. Stacking GA4, Hotjar, and a separate testing tool tends to add 200-400ms to TTFB on Shopify stores.
Watch for hesitation on price, repeated taps on non-interactive elements, scroll thrashing (up-down-up suggests the user is searching), form re-entry, and the exit moment. The last 10 seconds before a drop-off carry most of the diagnostic weight.
Standard retention is 30-90 days, which matches typical CRO sprint cycles. Some vendors offer 12 months on higher tiers, but anything older than a quarter is rarely actionable — your site has changed, your traffic mix has shifted, and the insight is stale.
Yes — modern tools record taps, pinches, scrolls, and orientation changes on iOS and Android browsers. Mobile replays are often more revealing than desktop because fat-finger taps, viewport issues, and slow network conditions surface clearly in playback.
Replay is where you generate hypotheses; A/B testing is where you prove them. Watch 20-30 replays of users dropping at a step, spot a recurring friction pattern, ship a variant that fixes it, then test. Skipping the replay step is why most A/B tests come back inconclusive — the hypothesis was a guess.
Get an AI expert review of your site
Paste your URL — Metricuno's AI runs the same heuristic checks a senior CRO consultant would, scoring your page and prioritising the fixes that'll move conversion fastest.