Skip to content
English
  • There are no suggestions because the search field is empty.

Digital Twin Confidence Score

Every Digital Twin answer shows how strongly your data supports it. Use the Confidence Score to decide how much weight to give an answer before you act on it.

What the Confidence Score tells you

Each Digital Twin answer now carries a Confidence Score. It shows how much of the answer your data supports, so you can decide how much weight to give it before you share it or act on it.

The score is per answer, not per twin. The same twin can score High on a topic your surveys cover well and Low on one they barely touch.

What goes into the score

  • Share of the answer backed by data: each statement is marked Backed by data, Partly backed or Interpretation.

  • Smallest sample: a figure from 50 respondents scores a level below the same figure from 1,000.

  • Statistical strength: narrow confidence intervals and significant differences raise the score.

  • Closeness to your data: a question your surveys asked almost word for word scores higher.

What each level means

  • Very high (Decision Grade): the twin's data goes deep on this topic. Lead with it, and cite the sources behind the answer.

  • High: well supported by the twin's data. Share it with stakeholders as you would a survey result.

  • Medium: reliable on broad patterns. Results vary by segment, so check the segment you are targeting and the cited surveys before you decide.

  • Low: early data on this topic. Add a survey, turn on Panoplai Data, or test with real people. A Low score shows where your research has gaps. It is not a failure.

How to read a score
  1. Start with Overview. Read the level first, then the Why This Level note. It names the evidence used and counts statements by evidence type.

  2. Check Grounded in. Open a listed source document and check that the cited pages cover your question. Page ranges and links to the originals are shown.

  3. Read Confidence last. Every statement carries one of three marks. Read Backed by data first: the data shows it directly. Partly backed means there is some support in the data plus the twin's reasonable reading, so check the cited survey. Treat Interpretation as the twin's own reading, a hypothesis to test.

When to use it

  • When you want to know how far to trust an answer, open Confidence and start with Backed by data, then move to the sources.

  • When a level comes back lower, read Why This Level, then add data or turn on Panoplai Data and ask again.

  • When you introduce a new user to Digital Twins, run all three sections once on their own question so they see the pattern.

Where teams use it

  • Brief an executive: share an answer with the sources and page numbers behind it, so the room can check it.

  • Separate fact from hypothesis: see which statements the data supports and which are the twin's own reading.

  • Spot data gaps early: a lower level signals thin evidence, so you can add data before findings go further.

Need help?

Contact your Panoplai customer success manager or email support.