Our Methods

Usability testing

The core method. You give a tester a task ("go find the profile page") over a prototype hosted inside the tool, and urTRACE captures every click, tap, and pointer move as they work through it. Those recordings become per-frame click and movement heatmaps - individually or aggregated across everyone - plus task success rates, time spent per screen, and a session-by-session view of the exact path each tester took. It's how you see where people hesitate, backtrack, or get stuck.

Reaching the goal is not one finding but three, so urTRACE separates them: a path diagram of the moves between screens, split into direct success, indirect success and unfinished. Draw the routes you expect people to take and each is measured on its own terms - useful when a study has more than one right answer, or when an A/B test's arms start in different places. Draw none and the benchmark is the shortest route your prototype actually permits.

First-click testing

The classic signal for whether a layout reads the way you expect: where do people click first, not just eventually. urTRACE takes each tester's earliest click on a screen and pools them into one map, so a high "background" or wrong-element score tells you the design isn't guiding attention where you intended - before the tester has had any chance to explore.

A/B testing

Show different testers different versions of any screen in the flow and compare the results side by side. Each tester is randomly assigned a variant; urTRACE reports success rate, average time, and clicks per variant so you can see which one actually performs better, not just which one looks better.

The comparison sits in the study's own results, as a section beside the heatmaps and the sessions rather than on a page of its own - and picking an arm scopes everything else to it, so "how did B do on this screen" is the same page you were already reading.

Surveys

Attach questions to a test - asked before the task (to capture who the tester is) or after it (to capture how it felt), as multiple choice, rating scales, yes/no, numbers, or free text, plus an optional closing comment. The survey view aggregates the answers into per-question charts with outcome cross-tabs, so you can tell whether a particular group struggled - e.g. did the people unfamiliar with the product take longer or fail more?

A survey can also stand alone - no prototype and no task, just the questions. Write them, share the link, and the responses chart themselves on a results page like any other study. A standalone survey drops what needs screens to mean anything: there are no heatmaps, no route analysis, and no pass rate, because nobody was asked to do anything.

AI insights

An AI-written summary of the friction points and patterns across a test's sessions, including how the survey answers relate to who succeeded. It's bring-your-own-key - point it at your own Claude, GPT, Gemini, or a local model - so no test data leaves for a vendor you didn't choose, and there's no AI cost baked into urTRACE.