Blog / 2026

5 usability problems you can find in session recordings before you run a study

6 August 2026 · 5 min read

Not every usability problem needs a study. A lot of the most common ones leave a specific, recognisable trace in real session data, and if you already record sessions, that data is sitting there right now. Here are five you can go find today, roughly in order of how easy they are to spot.

1. Rage clicks — something looks broken

A burst of three or more clicks on the same spot in under a second is the least ambiguous signal in session data. It means a person clicked, expected something to happen, and it didn’t — so they clicked again.

The cause varies (a failed request, a missing loading state, a disabled button that doesn’t look disabled) but the presence of the pattern is never noise on its own. The full detection logic, including the false positives to watch for, is in what is a rage click, and how are they actually detected.

Where to look: filter sessions for rage-click events and check whether the same control shows up across multiple distinct sessions. One is an anecdote; a dozen sessions hitting the same button is a usability finding.

2. Dead clicks — controls that don’t register

A dead click is a single click where nothing visibly changed and no request fired. On its own it’s weaker evidence than a rage click, because plenty of dead clicks are just slow responses or purely visual changes the detector can’t see. But a dead click that repeats on the same element across sessions is a strong signal that something reads as clickable but isn’t — a heading styled like a link, a card that looks interactive but has no handler.

Where to look: the same filter you use for rage clicks, narrowed to single-click events with no following network activity or DOM mutation.

A user going from page A to page B, back to A, over to page C, back to A again is a navigation pattern that almost never means “browsing casually.” It usually means they’re comparing something across pages that should have been on one screen — pricing tiers next to a feature list, a form next to instructions, a cart next to shipping options.

This one doesn’t have a clean automated flag the way rage clicks do — you’re looking at the actual path a session took, ideally through a transcript rather than replaying video, since the pattern is obvious in a list of timestamped page views in a way it isn’t in real time.

Where to look: session transcripts sorted by page-view sequence. Anything with the same two URLs alternating three or more times in one session is worth a look.

4. Form abandonment at a specific field

Not “the form has a high abandonment rate” — that’s aggregate analytics, and you probably already have it. The usability finding is which field, and whether the same field shows a pause, a correction, or several failed submit attempts before the person leaves.

A field that gets typed into, cleared, and retyped multiple times before abandonment usually means the validation message is unclear or the expected format isn’t obvious (date formats and phone numbers are the most common offenders). A field where people just leave without touching it often means they don’t understand why it’s being asked for.

Where to look: sessions that reached the form but didn’t submit, filtered by how far into the field sequence they got before leaving.

5. Confused scroll behaviour

Fast, repeated up-and-down scrolling in a short span — as opposed to a single steady scroll down the page — usually means someone is hunting for something they expected to find and can’t. It’s a softer signal than a rage click (harder to threshold cleanly, more false positives from people skimming), but combined with a subsequent exit or a support ticket, it’s worth checking.

Where to look: this is one of the few on this list better watched than queried — pull a handful of sessions on the page in question and see if the scroll pattern is consistent across them.

Why this beats scheduling a study first

None of these require recruiting anyone. They’re not a replacement for usability testing when you need to know why — a moderated test is still the right tool for that, see session replay vs. usability testing — but they tell you where to point a study, which is normally the part that takes the longest to figure out on your own.

The practical blocker is usually time: watching session video to spot these patterns doesn’t scale past a handful of sessions. UserTapes turns every recording into a searchable text transcript, so a rage cluster or a repeated navigation loop shows up as a line you can scan or query, rather than a video you have to sit through:

- ⚠ 0:19 rage click ×4 · button "Pay now"
- ⚠ 0:21 POST https://example.com/api/pay · 502

You can also set a standing alert on any of these conditions — a rage-click alert on your checkout flow, say — so you find out the day it starts happening instead of the day someone thinks to go look.

UserTapes records 100 sessions free with no card and flags rage clicks, dead clicks, and failed requests automatically. If a pattern above turns out to be real and you need to know why, that’s the point to run a small moderated test on that one screen — not before.