Skip to main content
Heralded leaves a figure empty, shown as a dash, when it has nothing it can stand behind. An empty figure is never a zero. Search awaiting its first measurement shows Calculating instead of a dash. Headline figures describes its pending state and when it applies.

No report yet

Figures come from completed reports. A new brand’s first report starts as soon as you confirm its questions, and Home shows it as running until it lands. After that, figures refresh with each weekly report.

Nothing could be measured

A headline AI figure needs at least three measured answers behind it. With one or two, the API returns too_few_answers and a null value. If every answer it covers was not measured, the figure is empty. Understood and Praised show a dash when no answer named you. Perception names the available population beneath a hidden figure, such as “of 1 answer naming you”. The Snapshot report reads Not measured this run. The Heralded Score needs both Mentioned and Recommended. If either is missing, so is the score. A Europe figure is empty while one of its countries has no answers in the 14 days it reads. Europe topics lists what a Europe figure leaves out. On a Snapshot, an engine that measured no questions stays empty. Questions cut by the scan deadline are separate from questions an engine failed to answer. A deadline cut always carries a coverage note. Otherwise, partial coverage carries a note when at least one third of the engine’s questions were unmeasured. Repeated answers to the same question count as one cell for this coverage.

A week that failed its checks

Every weekly report, including your first, is only scored when enough of it worked. If three or more questions got no response on report day, fewer than two engines answered, or fewer than ten question and engine pairs succeeded, Heralded marks that week’s run as failed and publishes no report for it. Your figures keep their last values, and the week’s daily answers are pooled into the next weekly report.

A score version

When Heralded changes how the score is computed, it becomes a new score version. Weeks with frozen per-answer facts are recomputed on v8.0; older weeks without them keep their issued score and version. A Heralded Score is empty when it was not measured.

Intentionally not scored

Search has no Heralded Score or keyword score, even when its Mentioned and Recommended are measured. Its score has a null value and the API state not_scored. This means no score is assigned, rather than that a measurement is missing. The Heralded Score explains why.

A change is empty

A figure can be present while its change is empty. That happens when no weekly report comes before the selected period, whatever its length, or when Heralded changed how it reads inside it. Change lists the cases.

A report read two ways

A weekly report whose answers were read two ways inside one topic, such as under two answer-pack versions or two engine setups, can’t be scored as one week. Its figures are empty. Home shows a dash with the tip “New baseline: changes from the previous report are not called.” The next report read one way fills them. A market change between daily updates and the weekly report never does this: daily answers asked before it sit out of the report.

Not measured on AI visibility

On AI visibility, a cell marked Not measured has no readable answer from that engine in the latest report. It is a gap and doesn’t count against you.