Reading your Numbers

Rankscale turns many AI answers into a few numbers. This guide explains how to read them without overreacting. For what each metric means, see 📄 Metrics in Rankscale.

Look at the Period, Not the Day

• AI answers vary from run to run, so single days jump around. The period average on a score card is a better guide than one day.

• Trend charts can show a Trailing average, which smooths those jumps. Switch Smoothing to Raw latest in the date picker to see each point as measured.

• For longer ranges, set Aggregation to Weekly or Monthly to see the direction more clearly.

Know What a Number is Based On

• Every score describes the Search Terms, AI Engines and Regions that ran in the period. Filters change that base.

• Read branded and non-branded Search Terms separately with Branded status in the Queries filter. Branded results are usually much higher.

• Scores only count runs that finished. A failed or paused run does not lower your score.

Zeros and Counts

• Charts that count things per day, such as citations, show 0 on days without runs. A 0 can mean nothing ran that day, not that you lost everything.

• Brands (N) on a cited source counts your Brand and your competitors mentioned with it, not every brand in the answer.

Tip: If days look shifted, check Timezone in the date picker.