> ## Documentation Index > Fetch the complete documentation index at: https://promptwatch.com/docs/llms.txt > Use this file to discover all available pages before exploring further. # Metrics overview > The six core Promptwatch metrics in one place: what each one measures, how they differ, and where to read the full explanation. Promptwatch condenses every AI answer into a small set of metrics. Each one answers a different question, and most confusion comes from reading one metric when you needed another. This page gives you the one-line version of each and links to the full explanation. ### The two building blocks: mentions and citations Everything starts with the distinction between your brand being **talked about** and your content being **used**. A mention is your brand named in the text of an answer. A citation is a URL the AI attached as a source. They are recorded independently and improve through different work: mentions through reputation, citations through content on your own domain. If you read one page from this group, make it [Citations vs. mentions](/docs/academy/citations-vs-mentions). ### Presence and prominence * **Visibility Score**: how prominently your brand appears when AI answers your tracked prompts, 0-100. Each response is scored on placement, share of attention, and depth, then averaged, with responses that don't mention you counting as zero. It is the single number that reflects both coverage and quality of presence. See [Visibility Score](/docs/academy/visibility-score). * **Share of Voice**: your percentage of all brand appearances, measured against your competitors. Presence only: a footnote mention counts the same as the headline recommendation. Read it next to the Visibility Score to separate "we don't appear" problems from "we appear badly" problems. See [Share of Voice](/docs/academy/share-of-voice). * **Position in answer**: where your brand lands in the order of brands within a single answer, counted from 1. Being the first name in a recommendation list is worth far more than being the ninth, and this metric tracks exactly that. See [Position in answer](/docs/academy/position-in-answer). ### Tone * **Sentiment score**: how positively AI answers describe your brand, on a 100-point scale where around 50 is neutral. Scored independently of visibility, so a highly visible brand can still carry negative sentiment, and vice versa. See [Sentiment score](/docs/academy/sentiment-score). ### Sources * **Citation rank**: when an answer cites your domain, where in the source list that citation lands, counted from 1. Sources listed first tend to shape the answer most, so a lower rank (closer to 1) means your content is becoming a primary source rather than a supporting one. See [Citation rank](/docs/academy/citation-rank). * **Self-citation rate**: the share of all citations that point at your own domain instead of third-party pages. It tells you whether AI models build their answers from your content or from what others write about you. See [Self-citation rate](/docs/academy/self-citation-rate). ### How to read them together The metrics form a funnel, and the combination locates your problem: * **Low Share of Voice, high Visibility Score**: the answers that include you treat you well. Your problem is coverage: get into more answers. * **High Share of Voice, low Visibility Score or late position**: you appear often but as an afterthought. Your problem is prominence. * **Mentions without self-citations**: models know your brand from third-party sources, but your own content isn't feeding the answers. Your problem is content and crawlability. * **Good scores, weak sentiment**: you are visible but framed negatively. Your problem is reputation, and the [Sentiment page](/docs/academy/sentiment-tracking) shows which responses are doing the damage. ### Where to see them The Brand Visibility chart on your dashboard, monitor pages, and prompt pages carries visibility, Share of Voice, and position as tabs. Sentiment has its own page under Monitoring. Citation rank and self-citation live in the Citations section. They share date range, AI model, prompt type, and topic filters; tags apply on monitor pages and in Citations, not on Sentiment or the project dashboard. Slicing by a single model is usually the most revealing cut: blended numbers hide the platform where you are losing. The Promptwatch Brand Visibility chart with line, bar, Share of Voice, and position view tabs. For definitions of the underlying objects (monitors, prompts, responses), see the [Glossary](/docs/academy/glossary).