Confidence Levels and Why They Matter
Every score comes with a confidence level so you know how much to trust it
Two scores can both be 4. One might be backed by a fifth of your cohort with direct quotes, the other by two participants. They are not equally trustworthy, and treating them the same would lead to poorer decisions.
Savo attaches an evidence label to every Dimension score so that distinction is always visible.
The three evidence labels
Strong evidence. Backed by many participants and supported by direct quotes.
Moderate evidence. Backed by several participants with consistent supporting quotes.
Limited evidence. Backed by a small number of participants, or by thinner supporting quotes.
What drives the label
The label comes from the share of your cohort behind a finding and how consistent that evidence is. As a guide, a finding needs support from about 20% of participants to reach Strong evidence and about 10% to reach Moderate evidence. On smaller events the participant counts behind those shares scale down, and never below two participants.
The label says nothing about whether the score is high or low. A Strength-tier Dimension can carry Limited evidence if only a few participants drove it, and an Opportunity-tier Dimension can carry Strong evidence if most of the cohort agreed.
Evidence labels and abstention
Evidence labels apply to scores that exist. When there is not enough evidence to produce a reliable score at all, Savo does not assign a Limited evidence score. It abstains and records "No Score."
The two are different kinds of uncertainty. A score of 4 marked Limited evidence means "we found some evidence pointing to a high score, but not enough participants said it to be certain." A No Score means "we did not find enough evidence to say anything reliable."
How to use the labels in practice
Treat Strong evidence findings as reliable enough to act on directly. Treat Moderate evidence findings as real signal worth investigating. Treat Limited evidence findings as hypotheses: they may hold, but look for corroboration before making a significant decision on them alone.
Limited evidence and No Score findings can also inform your event design. If a Dimension keeps producing thin evidence, the facilitation guide may need stronger signal anchors for it.
For how scores work, see Understanding Dimensions. For how to inspect the evidence behind any score, see Exploring the Evidence.