Reading a cell: where the value came from, and how sure we are

1. 1

   Open the cell rather than reading the table

   Click into a filled cell to open the detail view. The table shows you the value. The detail view shows you the evidence: the source, what the tool matched on, and any secondary fields it returned alongside the main value.

   

2. 2

   Read the source, not just the value

   A company website found from an official listing is a different thing from one inferred from a similar name. When two rows disagree, the source is what tells you which to believe.

3. 3

   Tell an empty cell apart from a miss

   An empty cell means the tool never ran on that row. A miss means it ran and found nothing. They look identical in the table and they mean opposite things. This distinction is the subject of an entire Practitioner lesson because it is the most expensive filtering mistake in the product.

4. 4

   Check what the tool matched on

   Tools that resolve an entity tell you what they matched. A people search returns the title it actually found and how it relates to the role you asked for. A verification tool returns pass, warn or fail with a confidence score and the sources it read.

What a confidence signal is for

1

2

3

1. 1Selected and Rejected are separate counts. A cell showing one profile may have considered a dozen.
2. 2The banner names the cause and the fix: quality scoring filtered these out, lower the strictness to include them.
3. 3A reason per rejection, so you can tell a bad match from a setting that is simply too tight.

Confidence is not decoration. It is a filter. A verification tool that returns pass, warn and fail lets you send the passes straight to outreach, route the warns to a human, and drop the fails. Treating all three as "the tool answered" throws away the most useful thing it gave you.