Sampling sessions without turning people into clips

Session quality analysis depends on evidence. Evidence depends on trust. We refuse engagements that treat recordings as marketing footage or that expand a sample past what consent allows.

Consent first

Before intake closes, we ask which consent notice covers session recording, how long files are retained, and who inside your organisation may watch them. If those answers are vague, we pause. Australian Privacy Principle expectations are not a footnote for us; they shape whether the work can proceed.

Sample size that stays honest

More sessions are not always better. For a three-path audit we usually aim for a stratified sample: successful completions, recoveries after error, and abandonments near the same step. Twenty carefully chosen sessions often teach more than two hundred skimmed ones.

Redaction habits

Names, account numbers, and free-text fields that contain personal stories are masked in our working notes. The findings report quotes paraphrased speech unless you explicitly approve a verbatim excerpt under your own policy.

What we never do

We do not install new trackers to “get better data.” We work with what you already lawfully collect. If that is insufficient, we say so and propose a clinic on live walkthroughs with staff accounts instead.

Questions about how a sample would look for your application belong in the contact form — send the consent text you currently show users and we will read it before proposing scope.