Signal & ScienceAcademy

Course 3 · Lesson 2 of 7

Follow the live run

Read a diagnosis while it runs without treating progress as a result, and know why partial data becomes Unknown rather than a failure.

Lesson 2 of 7 in Run and Read a Diagnosis
  1. 1Start a diagnosis
  2. 2Follow the live run
  3. 3Read capability evidence
  4. 4Read citations and source-bound claims
  5. 5Correct owned profiles and business facts
  6. 6Read Unknown without a negative judgment
  7. 7Decide whether a reassessment is warranted

Why this matters

A run takes minutes, and the live view is built to show progress, not findings. Misreading it causes two costly errors: quoting a half-finished picture to a client or colleague, or treating a blocked source as a problem with the business. The Diagnose step only earns trust if what reaches Prioritize and Plan is the finished, dated evidence map. The live view tells you the run is working; the finished report tells you what was found.

When to use it

While a first run or a re-run is in progress, and whenever you return to the Diagnostic page and find a run still working.

Before you start

  • A run has been started from the Diagnostic page (see Start a diagnosis).
  • Expect several minutes. The research and analysis stages can hold the progress figure still while work continues.

Do this

  1. 1Watch the center of the run view: the business name, a progress percentage, a stage word, and an elapsed-time clock. The clock is the honest sign the run is still working when the percentage pauses.
  2. 2Read the headline for what is actually happening. When the app knows the stage, live status such as "Running grounded research across public sources" or "Generating strategic analysis" replaces generic waiting copy.
  3. 3Read the category nodes (for example "Google Business Profile", "Competitor Field", "Site Architecture") as scans finishing, not as findings. Which nodes appear depends on the business type.
  4. 4Read the line along the bottom with care: the percentage beside "CONFIDENCE" counts finished scan categories. It is not the confidence level of any finding.
  5. 5If a strip reports that sources returned partial data, note it. The areas those sources feed will usually read Unknown in the report.
  6. 6If you leave the page or lose the connection, come back to the Diagnostic page. It reconnects to a run still in progress rather than starting a new one.
  7. 7Wait for the report before quoting anything. Observed and Unknown states appear only in the finished report.

What you should see

When the run finishes, the view hands off to the report, or to the quick check if the business type was detected. If some sources failed, the finished page shows "Some data could not be collected automatically" or "Some data sources had issues", with the note "Unobserved data remains unknown." Missing evidence stays Unknown; it is never filled in with a guess.

How to judge it

You can say whether the run has finished, which sources (if any) returned partial data, and the date on the finished cover. If you find yourself describing findings from the run view itself, you are reading progress, not evidence.

What it does not mean

A blocked or timed-out source means the evidence was not reachable in this run. It says nothing about the business. The progress percentage is not a quality measure, a slow stage is not a failing one, and a partial picture is not an early version of a verdict, because the diagnosis does not issue one.

Try it in Signal & Science

Open the Diagnostic page during a run and note which sources, if any, returned partial data.

Open in Signal & Science

Put it to work

Know what to do next.Know whether it worked.

Signal & Science keeps the evidence, the plan, and the results in one place, so each lesson here becomes a working habit in your own workspace.

Sign upCompare your businessSign in