Evidence Units — Every Insight Traces Back to What Someone Said
Full traceability from dashboard score to the exact words that support it
When Savo produces a dimension score, you should never have to take it on faith. The scoring methodology produces findings you can trace — all the way from the number on the dashboard to the specific words a participant said.
That traceability is built around a unit called an Evidence Unit.
What an Evidence Unit is
An Evidence Unit (EU) is a segment of transcript — a passage from a participant's conversation — that contains evidence relevant to one or more dimensions. Each EU is:
- Tagged with the dimension(s) it speaks to
- Classified by evidence quality
- Linked to the dimension score it contributed to
- Identified by a unique ID (formatted as
EU-XXXXXXXX)
The scoring system produces evidence units as part of the analysis process. Every dimension score is the aggregate product of the evidence units the session generated for that dimension.
Evidence quality labels
Not all evidence is created equal. Savo classifies each Evidence Unit into one of three quality tiers:
Story — The strongest evidence type. The participant describes a complete situation, action, and outcome. They don't just say something was hard — they describe what happened, what they did, and what resulted. Stories are the most informative because they demonstrate behavior in context.
Example — A specific past behavior or event, without the full narrative arc. The participant references something that happened, but without the situational context that would make it a Story. Still strong, behavioral evidence.
Claim — A stated belief or assertion, without a specific supporting example. ""I feel like things are improving"" is a claim. Claims are the weakest evidence type — they're self-reported and harder to calibrate against the dimension anchors.
When you read a finding in your Insights report, the evidence quality distribution across supporting EUs tells you how behavioral the evidence is, not just how much of it exists.
The traceability chain
The chain from insight to evidence is fully navigable:
- Dashboard score — A dimension score appears in your report (e.g., Competitive Confidence: 4.2 / High)
- Supporting evidence — Click through to see the Evidence Units that produced that score, with their quality labels and the AI's explanation of why each was scored the way it was
- Source transcript — Each EU links to the specific passage in the session transcript
This means any finding can be audited. If a score looks surprising, you can trace it. If a stakeholder asks why Competitive Confidence came back low in the eastern territory, you can show them exactly what reps said.
Why this matters
Evidence traceability is the difference between a plausible-sounding summary and a defensible measurement. Savo holds every insight to the standard of the evidence — not because it has to, but because that's what makes the insights worth acting on.
To explore evidence units in practice, see the Evidence Explorer tab in your Insights report. For how scoring works, see Understanding Dimensions.