A score from 0 to 100 with the breakdown in front of it

• You have five different metrics on five screens and nobody in the meeting knows which to look at first.
• The scores you have seen in other tools do not say how they are computed, so you cannot defend them.
• You need one figure that fits on a slide without being a lie.

A score without its breakdown is an opinion with decimals.

1. 1

   It starts from real answers

   Nothing new is launched: the score is computed on the sweep you already have. With no sweep there is no score, and we say so rather than show a zero that looks like a result.

2. 2

   It weighs four signals

   Presence, citation, position within the answer and engine diversity. All four weights are shown beside the number, and the breakdown opens down to the specific answer behind it.

3. 3

   It compares against itself

   Today's value is worthless without last month's. The series stores each point's configuration, so if you changed engines or questions the chart flags it instead of pretending the curve is continuous.

• The big number, with the sentence explaining where it comes from and the count behind it.
• The table of dominant competitors in your category, including the ones you did not have on your list.
• The engine that helps you most and the one that ignores you most, so you know where to put the effort.

• Anyone reporting upwards who needs a figure that survives the next question.
• Agencies presenting to several clients who need the same scale for all of them.
• Teams that want a concrete quarterly target instead of saying they will improve visibility.

The score is on every paid plan and on the trial, because it costs no extra usage: it is computed on sweeps you already paid for. What changes by plan is how many engines enter the calculation and how much history you keep.

See the full pricing

What exactly are the weights?

They are published on the methodology page, with the full formula and a worked example. It is not fine print: it is the sales argument. If they ever change, the computation version attached to each point of the series changes too, so nobody compares two numbers that were not computed the same way.

Can I compare my score with another brand's?

Only if both were measured with the same questions, the same engines and the same country. Comparing brands inside one sweep is valid by construction. Comparing against the score another tool gave another brand is not, and anyone claiming otherwise is selling a ranking, not a measurement.

My score dropped and I changed nothing. Is it a bug?

Almost never. Models change version, web search returns other sources, and a competitor publishes something. The series marks model changes with their date, precisely so you can tell your own drop from someone else's update. If the drop lines up with an engine version change, you will see it flagged.

Is a score of 0 any use if I have just started?

More than an inflated number. A real zero with its denominator in front says exactly what you need to know: in those questions and those engines, today you do not appear. From there, any movement is measurable. What is useless is a scale that never drops below 40 so the report looks good.