The PPS score
How the Prompt Performance Score is calculated, what each component weight means, and how to move it.
The Prompt Performance Score, or PPS, compresses four measurements into a single number from 0 to 100. It exists so that AI visibility can be reported to people who will never open a dashboard.
How it is calculated
The score is a weighted blend of four components. These weights are fixed and identical in every workspace:
What each component measures
Citation rate, 50 percent
The share of checks in the period where your domain was cited. It carries half the score because presence is the precondition for everything else: position and sentiment are undefined if you were never in the answer.
This is also the component you can move fastest, because going from absent to mentioned on a single high-volume prompt shifts the whole score.
Answer position, 20 percent
When you are cited, how early you appear relative to the other brands named in that answer. Being first in a list of four is meaningfully different from being fourth, and buyers read from the top.
Position is measured per answer and averaged. A workspace can have a high citation rate and a poor position, which reads as "we are always in the conversation, never the recommendation".
Sentiment, 20 percent
Whether the mention is positive, neutral or negative. A citation that calls you overpriced or hard to learn is worth less than a neutral listing, and the score reflects that.
Sentiment is available from the Starter plan upward. On the Free plan it is not computed.
Consistency, 10 percent
How stable your results are across engines and across runs. A brand cited by seven engines every day scores better than one cited by two engines sporadically, even at the same average rate, because the second brand's visibility is fragile.
Consistency needs several runs before it settles, which is why the PPS moves in the first week even when nothing about your content changed.
Reading the number
The dashboard shows the score with a word label. Treat these as broad bands rather than precise thresholds:
| Band | What it usually means |
|---|---|
| Strong | Cited on most prompts, usually early, described favourably, across most engines on your plan. |
| Moderate | Reliably present on some prompts and absent on others. Usually a content-coverage gap rather than an authority problem. |
| Weak | Cited occasionally, late in answers, or on only one or two engines. Competitors are winning the answers you care about. |
The lite score on Free
The Free plan shows a lite PPS based on citation rate alone, because sentiment is not computed on Free and a single engine cannot produce a meaningful consistency measure. It is directionally useful and not comparable to a full PPS from a paid plan.
How to move it
In rough order of leverage:
- Close outright absences first. Find prompts with a zero citation rate and read the stored answers to see what the engine cited instead. Half the score sits here.
- Fix stale and negative framing. A negative sentiment mention drags 20 percent of the score and is often traceable to one outdated source.
- Improve position on prompts where you already appear. Being cited late usually means your page is being used as corroboration rather than as the recommendation.
- Widen engine coverage last. Consistency is only 10 percent, so chasing a single stubborn engine is rarely the best use of effort.
Fix suggestions automates the first step by naming the specific gap behind each miss.
What the PPS is not
- It is not comparable between workspaces. Different prompt sets and different engine counts make cross-company comparison meaningless. Use Benchmarks for that.
- It is not a traffic or revenue metric. It measures presence in answers, not clicks or sales.
- It changes when you change prompts. Adding five hard prompts will lower it even though nothing about your visibility got worse. Change the instrument deliberately and note when you did.
Next
Read Citations and sentiment for how the underlying detection works.