Reading your data

Revenue per visitor

Revenue per visitor (RPV) is a channel’s revenue divided by its visitors. It’s the number Statlark leads with, because it’s the one that survives a bad assumption: a channel can send little traffic and still be your best, and a channel can flood you with visitors who never pay. Conversion rate hides both; RPV surfaces them.

Concretely, rev / visitor = revenue ÷ visitorsover the range you’re looking at. Sort your sources by it and the low-traffic, high-value channels that GA buries rise to the top. Revenue rides on every breakdown row — sources, pages, countries, devices, people — so you never have to guess which slice earned.

Filtering the whole page

Most rows on the Overview are also filters. Click a source, page, country or device — the row is a link, so it underlines on hover — and the entire page (the revenue chart, the KPIs and every other card) re-scopes to that slice. A chip appears in the toolbar; clicking the chip, the row again, or the filter button clears it. Prefer to pick from a list? The filter button in the toolbar opens a builder with the top values for each of the four dimensions — handy on touch, where there is no hover.

A card never filters by its own dimension: filter by a country and that card keeps listing every country (the one you picked marked with a coral rail) so you can compare and switch, while every other card narrows. Stack filters to drill in — say paid visitors from the US — and read exactly what that slice earned.

The time range and trend chart

The range control in the toolbar scopes every number on the page. Pick a preset — Now (live), Today, the last 7 or 30 days, week / month / year to date, or all time — and step to the previous or next period with the arrows. For anything the presets don’t cover, choose Custom range and set an explicit start and end date. The granularity control in the toolbar (hourly / daily / weekly / monthly) re-buckets the trend and defaults to whatever fits the range.

The hero chart follows whichever KPI you select: click Revenue, Rev / visitor, Paying conv. or Visitors and the trend redraws for that metric over the same range and filters.

Paying conversion

Paying conversion is the share of visitors who became paying customers — distinct paying visitors ÷ distinct visitors, in the same window. Unlike a goal conversion (someone hit a target you defined), this is strictly about money changing hands, so it’s immune to vanity goals. A channel with a high RPV but a low paying conversion is earning big from a few whales; a high paying conversion with a low RPV is lots of small orders. Read them together.

How channels are classified

Every visitor is assigned a channel from their first touch— the source, medium and referrer frozen on their very first visit, so a later direct return never overwrites the marketing that actually earned them. Statlark sorts each visitor into one of six channels, first match wins in this order:

• Ads — an ad-click id is present (Googlegclid, Meta fbclid, Microsoft msclkid) or the UTM medium is paid (cpc, ppc, paid, display, …).
• Email — medium email/newsletter, or the source names an email tool (Mailchimp, Klaviyo, SendGrid, …).
• Social — medium social, or the source or referrer is a social network (Facebook, Instagram, X/Twitter, LinkedIn, Reddit, YouTube, TikTok, …).
• Organic — medium organic, or the referrer or source is a search engine (Google, Bing, DuckDuckGo, Yahoo, …).
• Referral— there’s a referrer or source, but none of the above matched: another site linked to you.
• Direct — no referrer and no source: a typed URL, a bookmark, or a link stripped of its referrer.

“Organic vs paid” falls straight out of this: Ads is your paid acquisition, everything else is earned. Break sources down by Channel for the six-way view, or by UTM source / Referrer / Ad-click to go finer.

The visitor journey

Each visitor carries an entry page(the path they first landed on, frozen at first touch) and a chronological timeline — pageviews, goals, and payments in order. That’s how a page earns credit for revenue even when the sale happens three sessions later on a different page: the entry page, not the checkout page, gets the attribution. Open any person to see their full cross-device journey.

People & lifetime value

A person is a durable profile, merged across every device they use — so one human who browses on mobile and buys on desktop is a single row, not two visitors. Their lifetime value is the sum of all their payments (net of refunds), stored on the profile rather than re-summed per range, so it can never drift out of step with your revenue. Sort People by LTV to see who your best customers actually are — then look at the channel that first brought them, and pour more in there.

To turn anonymous visitors into named people, call statlark.identify() — see Identifying users. To pull any of these numbers over HTTP, see the REST API.