Dashboard
The daily read on one brand: six headline metrics, the visibility chart, rankings, and how to tell a measured zero from missing data.
On this page
The Dashboard at /dashboard is the daily read on one brand. Six headline numbers sit across the top, the visibility chart and rankings sit under them, and the rest of the page breaks the same report into models, prompts, citations, and traffic.
Everything on the page describes the brand in the switcher, filtered by whatever the header filter is set to. Numbers come from the last completed report, so the page changes once a day rather than continuously.
Read the six numbers, then follow whichever one moved. Each links to the page behind it.
The six headline metrics
The hero row is one strip of six cells. Each cell is a link.
| Metric | What it counts | Opens |
|---|---|---|
| Visibility | How prominently models named you, weighted so higher-ranked mentions earn more | The latest report |
| Mentions | Raw count of times your brand appeared across every prompt and model in the last 7 days | Prompts |
| Rank | Your position among the distinct brands tracked alongside you, ordered by visibility | Competitors |
| Citations | Unique URLs models referenced when they mentioned you | Citations |
| AI Traffic | Sessions from AI-referred visitors in the last 7 days, read from Google Analytics | Visitors |
| Conversations | Pages fetched by AI bots during live conversations in the last 7 days | Crawlers |
The change figure beside each number uses its own window, named in the small grey line underneath: Visibility compares with yesterday, Rank and Citations with last week, and the other three cover the last 7 days.
AI Traffic and Conversations need a connection before they show anything. Until then the cell reads -- with a Connect GA or Connect AI Crawlers link. A metric your plan does not include shows -- with Upgrade or Not included instead of a number.
Metric stories
Click the change figure, not the number, and the metric opens its story: what moved, the window it covers, and what the reading is based on (how many models and prompts were counted, or how many brands were ranked). The story also hands the question to the Agent if you want it explained in context.
Visibility's story compares the move with that brand's own usual day-to-day swing. A move inside the normal range is titled "Visibility held steady" rather than "Visibility is down", so a routine wobble does not read as an incident.
Visibility over time
The chart on the left is headed Visibility over time. It plots your daily visibility score for the selected range.
The range control offers 7D, 14D and 30D. A new brand only sees 7D. 14D appears once the series covers fourteen calendar days and 30D once it covers thirty. Thirty days is the longest window on this page. For 90 days, all time, custom windows, or weekly, monthly and quarterly grouping, use Reports and exports.
The chart has two compare modes, and they are mutually exclusive.
| Mode | What it draws | How to turn it on |
|---|---|---|
| Competitors | Your line plus a line per selected competitor, up to six | Click competitors in the Rankings panel |
| All markets | One line per tracked market | Filters, then All markets |
Two states are worth knowing. "Recalculating..." appears beside the heading while a tag filter is applied. "No tracked mentions in the latest scan" means the report completed and found nothing, which is a real zero rather than missing data.
Rankings
The panel to the right of the chart is headed Rankings. It lists every competitor plus your own brand, ordered by the rank in the latest report, and scrolls to put your row in view. Manage takes you to the competitor list.
Clicking a competitor adds its line to the chart. Six is the limit; after that, clicks are ignored until you remove one. Each row also carries a menu item that opens that head-to-head on Competitors.
Brands tracking more than one market see a Markets panel here instead whenever the chart is in All markets mode.
Filters
The header carries one filter control covering markets and tags. Its label adapts: a flag and market name when the brand only has markets, Tags when it only has tags, and Filters with a count when it has both.
| Option | What it does |
|---|---|
| A named market | Scopes the whole page to that market |
| All markets | Draws every market on the chart as its own line |
| Earlier reports | Reads results from before this brand tracked markets separately |
| Tags | Scopes the page to prompts carrying the selected tags |
| Clear all | Returns to the default view |
A brand with one market and no tags sees no filter control at all.
The rows below the chart
| Row | What it shows |
|---|---|
| Actions | The featured piece of work and the state of this week's plan, from Actions |
| Top Prompts | Your tracked prompts with scores, sparklines and change, taking two thirds of the row |
| Site Health | The audit score and issue counts, with an Optimize link |
| Perception | How models describe the brand, with a Details link |
| Models | All eight models ordered by visibility, with a Models / Markets switch for multi-market brands |
| Citations, AI Traffic, Conversations | Three cards carrying the recent detail behind the matching hero metrics |
The client portal is a trimmed version of the same page. It hides sharing, the citations and traffic cards, and site health, and shows a Results band in place of the Actions strip.
Share, download, and deeper views
Two buttons sit at the right of the header.
Share opens a menu with New share and Manage shares. On a paid plan it also carries Download PDF and Download CSV. PDF uses the browser print dialog with the chart converted to an image first, so choose "Save as PDF" there. CSV exports the detailed dataset behind the page.
Reports opens a menu with Reports for longer history and scheduled sends, and Data Explorer for slicing the same data by prompt, model, competitor, or tag.
States before the data arrives
A brand that has never completed a report shows one of these instead of the dashboard.
| What you see | What it means | What to do |
|---|---|---|
| First scan progress, with model chips | The first check is running now | Nothing. It usually finishes in about three minutes, and you can leave the page |
| "Your first report is on the way" | The brand is set up and the report is queued | Review your prompts while you wait |
| "Add prompts to build your dashboard" | The brand has no prompts, so there is nothing to check | Add prompts |
| "No dashboard data yet" | Tracking is paused. Your prompts and setup are saved | Resume tracking, or ask the workspace owner to |
Once a report exists, the dashboard stays visible and problems appear as a strip above it: "Showing the last available dashboard" when the refresh failed, "The latest report didn't finish" with a Retry report button, "Some ranking details are temporarily unavailable", or "Tracking is paused". A Refreshing chip in the header means a new report is being prepared and the numbers will update when it lands.
Zero and missing scores
A score of zero and a missing score need different next steps. Before reading either, check the brand, report date, model, market and filters on the page.
| What you see | What it means | Next step |
|---|---|---|
| 0 visibility with successful results | The measured answers earned no top-ten position points. Your brand may be absent or ranked below tenth. | Open the prompt and inspect the ranked brands and cited sources. |
| No score, an empty value or a dash | No usable measurement for this slice. It is not evidence of absence. | Check whether the report finished and whether this model or market has usable results. |
| First scan or a report in progress | The baseline is not ready. | Follow the progress panel and any blocker it names. |
| Missing AI Traffic or Conversations | A connection or measurement may be missing. | Check the analytics or crawler connection before assuming nobody visited. |
| Not measured on a page stage | The stage needs evidence or a connection before it can be judged. | Open the page record and follow its explanation. |
A provider or parser failure earns no position points and is left out of the visibility denominator. If every response in a slice failed, visibility is unavailable. If only some failed, a score can still show but rests on less evidence. Some older reports show zero for a slice with no successful responses, so check the model evidence before treating a historical zero as measured absence.
Presence counts usable answers that mention the brand. Failed requests are excluded, and collection coverage shows how many results were available. Fresh and reused answers are shown separately; reused answers keep their original rankings. Average position only describes answers where the brand appeared. Metrics gives each formula.
Check one low score
- 1.Open Prompts and choose a question that matters to a buyer.
- 2.Open its detail page and read the per-model results. Look for successful rankings, errors or unavailable data.
- 3.If successful rankings omit your brand, inspect the competing brands and sources. A mention below tenth still earns zero visibility points.
- 4.If your brand appears under another spelling, check Settings → Brands → your brand → Name variants. Add only aliases that truly refer to you.
- 5.Compare the same prompt, market and model across dates. Check when the prompt was added or edited and whether filters changed.
Compare like with like
A headline can describe the latest report while mentions, trends or change figures use a longer labelled window. Read the label next to each metric. A missing comparison can mean there is no earlier baseline; it does not mean nothing changed.
Changing the prompt set changes what you measure. Adding branded questions can raise the total without improving category discovery, so use tags to compare the same question group. Sentiment is separate again: a high position score does not mean a model describes you well. Perception covers that.
The public REST API, MCP responses and some integrations return 0 for an unavailable visibility score, so a zero there cannot tell absence from a missing measurement. Check the report in the app first. Your first API read covers the limit.
Common questions
Why can't I see 90 days on the chart?
The dashboard chart tops out at 30 days by design, and shorter ranges are hidden until the brand has that much history. Longer windows, custom date ranges, and weekly or quarterly grouping live on Reports.
Why is a competitor in the Rankings list but not on the chart?
The chart draws up to six competitor lines. Ranking rows are only added to the chart when you click them, and once six are selected further clicks do nothing until you deselect one.
Why does Visibility say it held steady when the number went down?
Visibility's story compares each move with that brand's own typical day-to-day swing. If the change is inside that range, the story says it held steady, because a move that size happens most days and is not a signal.
What does a visibility of zero mean?
If the chart says "No tracked mentions in the latest scan", the report ran and no model named your brand on any tracked prompt. That is a measured result, not a failure. A report that did not finish gives you a notice strip instead. Zero and missing scores explains how to tell the two apart.
Do the export buttons need a plan?
Yes. Sharing and downloads appear on paid plans only, and the client portal hides sharing entirely. Everything else on the page reads the same report.