Diarization
The process of splitting a recording into turns and deciding who is speaking at each moment.
Why does it matter for a brand?
In a focus group several people talk, interrupt and talk over each other. If a sentence is attributed to the wrong person, a finding can switch segments without anyone noticing: what a non-customer said shows up as a customer's opinion. That is why knowing who is speaking is not a technical detail but a condition for a report anyone can rely on.
How does WAV handle it?
Each participant wears their own microphone and a 360 camera records the room. Even so, in a real focus group we measured that a microphone's wearer was the loudest voice on it only 11 of 120 seconds. WR3, our software, separates who is speaking with 90.4% agreement with a human review, measured on 141 moments of a real session, and flags what it cannot decide so a person settles it by listening. The details are in how we measure.
What should you ask before a study?
- Does each participant have their own microphone, or is there only room audio?
- How accurately is the speaker identified, and on what sample was that measured?
- What happens to the moments when the speaker cannot be decided?
This term is part of the qualitative research glossary, which defines the fifteen terms that come up in any study; it is most closely related to verified quote, verbatim and one-way mirror room (Gesell chamber). To see it applied, browse the sample report, built from a fictional study: it is the same format a client receives.