How YRQR analyses voice — and why we don't use one scorecard for everyone

The analysis profiles

Every YRQR campaign runs against an analysis profile: a set of dimensions, labels and scoring guidance matched to the audience and context. A football club consultation is not a museum trail, and the analysis should not pretend otherwise.

Built-in profiles include Generic Voice, Football Club, Cultural Trail, Retail Activation, Treasure Hunt and Brand Event. Each profile defines what its dimensions mean, what strong and weak signal look like, and how scores should be read. Custom profiles are available on the Enterprise tier, built with our help.

The reporting structure stays consistent across profiles, so campaigns remain comparable over time even when the dimensions differ.

How scoring works

Every response is transcribed, then classified against the active profile's dimensions — eight per profile. Scoring is weighted toward concrete statements: specific praise, specific friction, named barriers. A response that says nothing scoreable is not forced into a score.

Alongside dimension scores, each response is enriched with sentiment, emotion, themes, urgency and trend context, and normalised with metadata for channel, event, topic, timing and language. Sentiment is the floor, not the ceiling: it frames the words, it never replaces them.

Quality flags

Reports retain caveats and quality flags so decision makers can see both the signal and its limits. Flags mark responses the analysis is less confident about — short or unclear audio, off-topic content, possible duplicates — so they can be reviewed or excluded rather than silently counted.

What we don't do

We don't force one universal scorecard onto every audience. We don't invent respondents, weight results to flatter, or strip caveats from reports. We don't headline sentiment percentages as if they were the finding. And participant responses are never sold as advertising inventory.