Reports

All date-ranged views share one date-range control: Today, Yesterday, 7 days, 30 days, This month, Last month, This year or a Custom from/to pair. The selection is remembered for the browser session, so switching between dashboards and documents keeps the range. Headline metrics are compared with the previous period of the same length (▲ / ▼ with the percentage change; bounce and exit rates read “good” when they go down).

Dates are interpreted in the Matomo website’s time zone, exactly as in Matomo itself.

Analytics section

Referrer tables show the brand’s logo beside its name when it is a known search engine, social network, AI assistant or site — grey at rest, in colour when you hover the row.

Overview

The tab the section opens on, and the only one that answers a question no single source can: how much of this site’s traffic is people, and how much is AI agents.

Headline cards — human visits, AI agent fetches, the AI share, and how long ago an AI agent last fetched anything.
Human and AI traffic over time — one daily chart with both, on two y-axes. They are deliberately not on a shared axis: a visit is a session and an agent fetch is a single request, so the two lines count different things.
What the agents came for — the fetches split by purpose: retrieval (an assistant fetching because somebody asked it something), training, and AI search indexing. The same colours as the AI Crawlers tab uses for the same purposes.
Visits from AI assistants — people who arrived from ChatGPT, Claude, Perplexity and the rest.
Google Search — clicks and impressions for the range, and index coverage: how many of the site’s pages Google has indexed.

The page renders whatever this site has. With only a traffic provider installed it shows the human half and names the package that would add the AI half; with only AEO, the reverse; with neither, it explains how to start. One source being slow or unreachable costs its own cards and nothing else — a muted line at the bottom says which, and why.

The AI share is a share of page requests — agent fetches against agent fetches plus human pageviews — not of visits, for the reason above. The index coverage card is a snapshot rather than a range figure (it is the latest known status of each page), so it carries its own “as of” date.

Agent figures come from Flowcourier AEO, which records them itself. They will not match AEO’s own dashboard exactly for a nominally equal period: this page counts whole UTC days, while the AI Crawlers tab’s “30 days” counts back from the current moment.

Realtime

Who is on the site right now:

World map of where the current visitors are: one bubble per city (or country when the city is unknown), sized by visits, with the details on hover. Needs city-level GeoIP in Matomo — the default GeoIP 2 database includes it; without coordinates the map stays empty and the overview says so.
Active visitors in the last 30 minutes (RealtimeMinutes) and visits in the last 5, with visits and actions.
Visits per minute across that window.
Latest visitors — last seen, location, device, browser and OS, referrer, actions, duration and the last page viewed.
Devices, countries and pages being viewed in the window.

Refreshes every 30 seconds while open. Source: Live.getCounters and Live.getLastVisitsDetails.

Visitors

Headline cards: Visits, Pageviews, Avg. visit duration, Actions per visit, Bounce rate — and Unique visitors when your Matomo computes them for custom ranges.
Visits over time: visits, unique visitors and pageviews per day.
Countries, regions and cities.

Source: VisitsSummary.get, Actions.get, UserCountry.*.

Behaviour

The same headline cards and a pageviews over time chart.
Pages — pageviews, visits, average time on page, bounce and exit rates for the 25 most viewed pages.
Entry pages and exit pages.

Source: Actions.getPageUrls, Actions.getEntryPageUrls, Actions.getExitPageUrls.

Acquisition

Cards per channel: Direct, Search engines, Social networks, Websites, Campaigns and AI assistants (ChatGPT, Perplexity, Copilot … — reported by Matomo 5.x) with their share of visits.
Top referrers chart and channels donut.
Tables: all referrers, search engines, social networks, AI assistants, websites, campaigns.

Source: Referrers.getReferrerType, Referrers.getAll, Referrers.getSearchEngines, Referrers.getSocials, Referrers.getWebsites, Referrers.getCampaigns.

Technology

Cards: Desktop, Mobile (smartphones and phablets) and Tablet shares.
Browsers chart and device types donut.
Tables: browsers, operating systems, device types, device models, screen resolutions.

Source: DevicesDetection.*, Resolution.getResolution.

Visits

Shown when the installed provider can answer per visit — Matomo can, Google Analytics cannot, so on a GA install this dashboard does not appear at all.

The visits of the selected range, newest first, one page at a time, with a filter for a single visitor id or user id. Opening one shows its journey: the entry page, each pageview with the time spent on it, and the last marked as the exit. Events, goals, site searches, downloads and outlinks are shown on the pageview they happened on rather than as steps of their own, so the path stays a path. Earlier visits by this visitor opens any of them in the same panel.

Two things worth knowing:

Time on the last page reads “—”, not “0s”. Matomo cannot know how long someone stayed on the page they left from unless the site’s own tracking code calls enableHeartBeatTimer(). These packages only read Matomo’s Reporting API; they do not emit tracking code. A dash means “not measured”, which is the truth.
Identity is the user id when your site sets one, and the per-device visitor id otherwise — the visitor id changes when the same person uses another device, so a history resolved by it is narrower.

The Realtime dashboard’s visitor rows open the same panel, marked Active now, using data that report already carries — no extra polling.

Access is Analytics section access, checked on the server; Flowcourier:Analytics:VisitorJourney:Enabled turns the whole feature off. Nothing is stored in Umbraco. See Configuration for both.

Search queries

Shown when Search Console is enabled: clicks, impressions, average CTR and position with previous-period comparison, a daily chart, and the top queries (with a filter), pages, countries and devices for the property.

AI Crawlers and AI Referrals

Shown when Flowcourier AEO is installed: which AI crawlers fetch the site and what they take, and who arrives from an AI assistant afterwards. These are AEO’s own views — see AI crawler analytics and AI referrals — shown here because this is where a site’s measurement belongs. On a site running AEO without Flowcourier Analytics they live under Settings → Flowcourier → AEO Dashboard instead.

Reading crawler data needs either the Analytics section or the Settings section.

The document Analytics tab

Open any document and switch to Analytics. The tab has one sub-tab per connected source — Traffic (the provider) and Google Search — and remembers the one you used last.

Google Search is both of Google’s views of the page in one place, because an editor reads them as one question: does Google have this page, and is it fast when it gets there? It appears when either feature is enabled and shows whichever half this site can serve, saying in one line what the other would add.

From Search Console: the page’s index status with what to do about it, its canonical and discovery detail, clicks and impressions over time, and the queries, countries and devices behind them.
From PageSpeed Insights: the Lighthouse scores and the real-user (Chrome UX Report) numbers, for mobile or desktop. It shows the last analysis the server holds and waits for Run analysis before starting a new one — a run takes 20–60 seconds and spends PageSpeed quota. Lighthouse’s own detail is a link away in Google’s tool, which runs a fresh test there, so its numbers will differ.

The tab appears only on documents that are web pages. Umbraco cannot tell a settings node from a page by itself, so the packages decide per document: your configuration first (see which documents get the tab), then a template, then traffic on the page’s URL, then whether the public site serves it. A page that is new or has no data yet keeps the tab with a notice; a node judged to be data gets none.

The tab is shared by every Flowcourier extension, so the other ones you have installed add their own sub-tabs beside these: SEO (the SEO Agent: this page’s SEO fields, and its AI-readiness score where AEO can produce one), AI Traffic (AEO: which AI crawlers fetched this page and who arrived from an assistant, with a dot on the sub-tab when one has been here in the last day) and AI Visibility (Analytics AI: whether the sampled assistants cite this page, for which prompts, and who was cited instead). Nothing to configure — installing the package adds its sub-tab, and a package installed on its own simply has the tab to itself.

The Traffic view is scoped to the page’s tracked URL for the language you are editing:

Pageviews, visits (visits that viewed the page), avg. time on page, bounce rate, exit rate and entrances, each compared with the previous period.
Pageviews over time per day, with the page flow directly beneath it.
How visitors reach this page, and where they go next — one flow instead of two lists. On the left, where they arrived from: other pages of your site and external referrers. On the right, what they did next: other pages, links off the site, downloads, the visits that came back to this page, and the ones that left the site altogether. Each side shows its top 8 plus one “other” row that carries the rest honestly, and every entry its share of that side. The List button swaps the graph for the numbers.

The bands run from each side into the page, never straight across it: without a segment applied the two sides are separate totals, so a band drawn from one arrival to one departure would claim a path through the page that the data does not actually show.

Narrow it to one channel by clicking an arrival group — search engines, social networks, direct, campaigns, AI assistants. Only the flow reloads, and the right-hand side then genuinely is conditional on the left: of the people who arrived from search, here is where they went. “All” puts it back.

The narrowed view is capped at the last 90 days and cached for 15 minutes. That is not arbitrary: unlike the rest of the tab, this report is built from raw logs rather than pre-archived data, and a wide range makes the Matomo host work hard for it.

Needs Matomo’s Transitions plugin. Without it this area shows its empty state and the rest of the Traffic view is unaffected.
Channels that brought visitors here, as a bar chart of the top 5. Clicking a channel drills into the individual referrers within it — the same shape as the countries card below. “Direct” is the exception: those visitors arrived with no referrer at all, so there is nothing to list, and the card says so rather than showing an empty table.
Top referrers to this page, again as a bar chart of the top 5 with brand marks.
Where the visitors are — countries as a bar chart, with a Table toggle for the exact numbers. Clicking a country drills into its cities; All countries goes back. Both levels come from the same request, so drilling in costs nothing and is instant.
Technology — device types and browsers in one card, switched with the Devices / Browsers buttons. The devices donut carries the whole site’s split in its legend for the same range (“Mobile · site 62%”), so the page’s mix can be read against the baseline.

Every one of these cards opens as a chart and has a Table toggle. The chart shows the top few, which is what a shape is for; the table shows the rest.

Source: Actions.getPageUrl for the page’s own metrics and daily series, Transitions.getTransitionsForPageUrl for the previous/next pages, and the referrer, location and device reports filtered with a pageUrl== segment.

Findings: what to look at first

The tab opens with a strip of findings — things worth changing on this page, each with the evidence and a button to the sub-tab holding the detail. They judge the last 28 days against the 28 before (not the date range picked below), so they stay put between visits, and a page with fewer than 50 pageviews in the window says so instead of guessing.

A page younger than 14 days gets no strip at all — not an empty one. Nothing useful can be said about a page nobody has had the chance to visit yet, and “nothing to fix” on a page published yesterday reads as a clean bill of health it has not earned. No source is even asked, so a new page costs no upstream calls. The threshold is Flowcourier:Findings:NewPageDays; set it to 0 to judge every page from the day it is created. Nodes that are not web pages are skipped the same way.

Finding	It fires when	What it suggests
Traffic down	Visits fell 30 % or more; the channel that fell most is named	Depends on the channel: search → title, description and queries; social → the posts have aged; direct → a link changed; campaigns → the campaign ended; AI → refresh the answer-first structure
No visits	An old page recorded no pageviews	Link it from pages with traffic; check the sitemap and index status
Visitors land here and leave	A landing page’s bounce rate is 1.5× the site median	Make the first screen answer what brought people here; add a next step
Visitors stop here	A mid-funnel page’s exit rate is 1.5× the site median	Add the next step where visitors currently stop
No page leads here	No internal previous page in the window	Add links from related pages
Visitors search right after	Site searches after the page reach 10 % of its pageviews; the terms are listed	Cover the searched terms, or link to the pages that do
Slow on the server	Server time is 2× the site’s and at least 300 ms	A developer or hosting problem, not a content edit
Not indexed	The last Search Console inspection is not a pass (see the advice table in the Search Console guide)	Per state
Queries just off the first page	Search Console queries at positions 8–20 with 100+ impressions whose words are missing from the title and headings (SEO Agent)	Work the term into the title or a heading
AI assistants read it but never cite it	Verified AI-bot fetches, zero visitors from assistants (AEO)	Direct answer up top, question-shaped headings, llms.txt

Thresholds are under Flowcourier:Analytics:Findings and the window under Flowcourier:Findings:WindowDays; both are shown on the settings pages.

When the tab shows a notice instead

Notice	Meaning
Nothing to report yet	The document has never been saved — publish it first.
This page has no public URL	The page is not published in this language, or no tracked base URL could be determined (set `SiteUrl`).
No visits recorded	Matomo has no pageviews for that URL in the range. Widen the range, or check that the tracked URL (shown in the notice) matches what Matomo records — see `SiteUrl`.
Connect Matomo to see analytics here	`Endpoint`, `TokenAuth` or `SiteId` is missing.
Matomo rejected the token	The token is invalid, expired or its user cannot view the site.

Where the numbers differ from Matomo’s UI

Unique visitors over a custom range are omitted unless your host enables enable_processing_unique_visitors_range; Matomo’s own UI hides them in the same case.
The document tab’s visits metric counts visits that included the page, so it can exceed the page’s entrances and never exceeds its pageviews.
Segment-based breakdowns are processed on the fly. On hosts with browser-triggered archiving disabled they can be empty until the segment has been archived; the page metrics and Transitions are unaffected.