Latest / Software Testing with Fexingo: QA, Automation, and Reliable Software Engineering / How Session Replay Finds the Bugs Users Never Report
Transcript
- Lucas: There is a category of bug that QA never sees, automation never catches, and users almost never report. But session replay catches it every time. Luna: Session replay — so that's the tool that records a user's mouse movements, clicks, scrolls, and what they actually see on screen? Lucas: Exactly. Companies like FullStory and Hotjar offer this. You embed a small JavaScript snippet on your site, and it captures a video-like log of every user session. Not the actual video — it reconstructs the DOM events, so it's more like a playable timeline of what happened. Luna: And how does that find bugs differently from, say, error logs or crash reports? Lucas: Error logs tell you something broke. Crash reports tell you the app died. But session replay shows you the user's entire path — the hesitations, the repeated clicks on a non-responsive button, the field they filled out that suddenly cleared. It's the behavioral context that's missing from most monitoring. Luna: Right. So you catch things like a form submission that silently fails, or a dropdown that closes before the user can select an option. Lucas: Yes. One classic example: an e-commerce checkout flow where clicking 'Place Order' does nothing — no error, no spinner, nothing. The user tries three times, then leaves. No error log because the JavaScript event listener never fired due to a race condition. Session replay shows exactly that moment. Luna: That's huge. I've heard numbers like 1 in 10 users actually report a bug when they encounter one. The rest just leave. Lucas: That's the silent majority. And for the ones who do report, they often describe what they think happened, not what actually happened. Session replay gives you the ground truth. Luna: But there must be privacy concerns. You're recording real user behavior. Lucas: Right, and that's the biggest tension with this tool. FullStory and others allow you to mask sensitive fields — credit card numbers, passwords, personal data — so they're never recorded. You can also exclude pages or entire user segments, like logged-in admins. But critics argue that even anonymized replay can reveal identity through patterns. Luna: So how do companies justify it? Is it worth the risk? Lucas: For many, the improvement in user experience justifies it, as long as they're transparent. A typical consent banner might say 'We record sessions to improve our site — no personal data is captured.' And you can see the trade-off: a session replay tool catches bugs that would otherwise take weeks of manual testing to find. Luna: I want to dig into the technical side. How does session replay actually work without overwhelming the server with data? Lucas: Good question. The snippet captures a sequence of DOM mutations and user events — clicks, scrolls, input changes — and compresses them into a log. It's not streaming video; it's a list of operations that the player reconstructs. A typical session might be a few hundred kilobytes. And it's sent asynchronously, so it doesn't block the page load. Luna: And do you need special infrastructure to store and search all those sessions? Lucas: The tools handle that. FullStory indexes sessions by events — so you can search for all sessions where a user clicked a specific button, or where an error appeared. That makes it powerful for debugging. Instead of guessing, you can pull up exactly the sessions that hit a known issue. Luna: So you combine session replay with your bug tracker. A developer sees a bug report, searches for relevant sessions, watches what the user did, and fixes it with full context. Lucas: That's the ideal workflow. And some teams go further — they use session replay to identify 'rage clicks,' where a user clicks furiously on a non-responsive element. That's a signal that something's broken, even if no error is logged. Luna: It's like having a usability lab for every user, all the time. Lucas: Exactly. And it catches things that traditional QA misses because QA testers don't behave like real users. They follow scripts. Real users do unpredictable things — like opening the same page in two tabs, or clicking a button before the page fully loads. Session replay captures those edge cases. Luna: I've also heard that session replay can be used for conversion funnel analysis — seeing where users drop off. Lucas: Absolutely. You can watch a user's flow through a registration process and see exactly where they get confused or frustrated. Maybe they try to enter a phone number in a format the field doesn't accept, and then leave. That's a UX bug, not just a code bug. And it's fixable. Luna: Let's talk about the elephant in the room: how do you decide which sessions to record? Recording every session is expensive and might feel invasive. Lucas: Most tools let you set a sampling rate. You might record 1% of anonymous sessions, or 100% of sessions on a specific page you're testing. Or you can record based on events — like only record sessions where an error occurs. That reduces the privacy impact and the data volume. Luna: And what about legal compliance? GDPR in Europe, CCPA in California — session replay could be considered tracking. Lucas: Yes, and that's why you need clear consent. Many companies now require opt-in consent before recording. And they make sure to mask any personally identifiable information. Some tools even allow you to block recording for users who opt out, at the network level. Luna: So it's not a set-it-and-forget-it tool. You have to configure it carefully. Lucas: Definitely. But when done right, it's one of the most effective ways to close the gap between what you think your app does and what it actually does in the wild. And it builds empathy — you see the user's frustration firsthand. Luna: That's a good point. Session replay humanizes the bug report. It's not just a stack trace; it's a person trying to use your product and failing. Lucas: Exactly. And that perspective is invaluable for prioritization. A bug that causes one user to rage-click for thirty seconds might get bumped up the backlog once you see the video. Luna: Alright, so if a team wants to start with session replay, what's the first step? Pick a vendor? Or set up privacy rules? Lucas: First, decide your goal. Are you debugging specific technical issues, or trying to improve overall UX? That will guide your sampling strategy and which events you focus on. Then pick a vendor — FullStory, Hotjar, LogRocket, SessionStack — they each have different strengths. FullStory is great for search and analytics; Hotjar is simpler for UX insights. Luna: And then configure the privacy settings before you even deploy the snippet. Mask all input fields by default, exclude any pages that show sensitive data, and set up consent. Lucas: Yes. Then run a pilot with a small percentage of traffic. Watch a few sessions yourself. You'll almost certainly find something surprising within the first day. Luna: That's a good test — if you don't find anything, maybe you're not looking at the right sessions. Lucas: Right. And remember, session replay is a complement to your existing testing, not a replacement. You still need unit tests, integration tests, and manual QA. But session replay fills the gap where users behave differently than you expected. Luna: And it gives you a direct line to the user experience, without relying on them to file a bug report. Lucas: Exactly. And on that note, if today's conversation gave you something useful, here's a quick thing — we keep this show ad-free, and that's only possible because a small group of listeners chips in each month through buy me a coffee dot com slash fexingo. It's not a big ask, just a way to support the work if you find value in it. Luna: Yeah, it's a simple way to keep the content independent and focused on what's useful. Lucas: Right. So back to session replay — one last thought: if you're a developer or QA engineer, I'd encourage you to try watching just five real user sessions of your own app. You'll probably see something that changes how you think about your product. Luna: I think that's a great closing challenge. And for our listeners, we'll have links to some of the tools we mentioned in the show notes. Lucas: Thanks for joining us. Until next time.