Visibility

Visibility, under Analytics → Visibility, shows how often AI assistants bring your brand up. Check here whether your rate moved and why. Open Gaps to see the questions you are losing and who wins them.

The glossary defines the shared metrics: AI visibility, organic visibility, share of voice and average answer position. This page describes the shopping-card rates and Named, not known.

What is on the pageCopy Anchor Link

In Summary, the headline is the mention rate, pooled across the period. The change under it compares with the period before, on only the questions present at both ends, and says what Fond measured it over. If you added or removed questions, the line says so. If nothing is comparable, Fond withholds the change.

Visibility Over Time is the trend chart, with Total and Organic lines. A switch breaks it out by assistant.

If runs in the period searched for shopping, the page shows two rates. One is how often a shopping card appeared, and one is how often it showed you. Fond measures both on Google AI Mode only. A note says how many runs Fond scanned.

Share of Voice shows, for each question tier, the share of questions where at least one answer mentioned you. It also shows how many answers mentioned you and which third of the answer you appeared in. A bar shows the organic rate, with its number of blind questions. Under each tier are the other brands on its blind questions.

Beside it, Named, not known watches the quality of the mentions that Fond finds on questions that named you. It does not change the rates above it.

Share of Voice by Cluster groups the same measure by the topic clusters that you assign. It appears only after you assign clusters to some questions.

Gaps lists every question with at least one answer, where fewer than half of the answers in the period mention you. You can search the list and filter it by tier. Open a row to see who won that question.

Reading the trendCopy Anchor Link

The chart draws one point for each run, so a seven-day gap between runs is as wide as a two-day gap. The slope does not tell you how fast the rate changed. Read the points.

The dashed line is the headline rate, pooled over answers and not averaged over runs. So a run with more answers pulls it further, and it can sit against the direction that the line seems to go.

A marker under the line shows a change in what Fond measured. It marks a changed question set or tag, an assistant that joined or left, or a change in how Fond measures an assistant. A step at a marker can come from that change alone.

Two changes draw no marker: a new model behind an assistant, and a change to how many times Fond asks each question. If you know that one of these changed, read the step with that in mind.

In the per-assistant view, each line is the rate of that assistant over its own answers. The overall line is their average, weighted by answers, so an assistant can sit far from it.

Fond rounds each figure on its own, so a per-assistant number can differ from the overall one by a point. An assistant line bridges a run where it gave no answer. In the Total and Organic view, a gap stays a gap. A missing Organic point means that the run had no blind questions, not zero.

What moves the numberCopy Anchor Link

• The questions that you track. If you archive the questions that you do not win, the rate goes up with no change in any answer. The chart marks those runs.
• The tier of your questions. Far Reach is competition at the category level, and Safety is the tier you expect to win. A bank weighted to Far Reach reads lower. So Fond splits share of voice by tier.
• Your brand names. Fond counts mentions of your brand name, its aliases in your brand settings, and the variants that it matched to you. To count a misspelling, add it as an alias.
• Failed answers. Fond leaves a request that errored out of both halves of the rates that count answers, and does not count it as a miss. So a run that reads "Completed with errors" is measured over fewer answers. A run whose blind questions all failed shows Organic as n/a.
• Share of voice counts questions instead. A question whose every answer failed stays in its tier as one you did not win, and is absent from Gaps.
• The assistants and models that your runs use, up to what your plan allows. A brand measured on one assistant and on four gives two different numbers.

Why two assistants disagreeCopy Anchor Link

Assistants are different products, with different training, models and ways of searching the web. One assistant also answers the same question differently from one ask to the next, so Fond asks every question several times and reports a rate. A spread between assistants shows which one you are missing from.

Summary or one runCopy Anchor Link

Visibility, Results, Citations and Competitors share one control at the top right, and your choice carries over between them.

• Choose Summary to pool the runs in a date range that finished, are not hidden, and have at least one answer. Pick 7, 28 or 90 days, all time, or your own range. The default is the last 7 days.
• Choose Per run to see one run, which you pick from the run picker.

AI Traffic follows the same date range, with its own picker and no per-run view. Agent Readiness has neither, because it reports on your site.

What this page does not tell youCopy Anchor Link

• Why an answer changed. Fond records what each answer said, not its reasoning. The Query Results page has the answers.
• Where you rank against rivals. Share of voice is your rate within a tier, not your share of its brand mentions. Competitors ranks brands.
• Whether assistants can read your site. Agent Readiness checks that.
• Whether AI sent you traffic. GA4 counts sessions from AI surfaces, and Cloudflare counts HTTP requests, which are a different number. Edge measurement is not on every account yet.
• Anything about questions that are not in your query bank.