A tour of the Analytics page: the headline numbers, the four tabs, what every metric means, and the one mistake to avoid when reading per-product revenue.
The short version: Analytics leads with what matters — attributed revenue, buy rate after try-on, opens, and try-ons — then breaks the story into four tabs: Funnel (where shoppers drop off), Performance (which products and pages work), Engagement (how deeply shoppers use it), and Preview widget. The attribution window is adjustable (7, 14, or 30 days) and every stat is labeled with it. Read the store-level revenue number as the headline; never sum the per-product rows.
The top of the page answers "is this working?" in one line — try-ons drove $X in sales, up or down vs the prior period — and flags your weak spot: the funnel step with the biggest drop-off. Under it, four tiles:
| Tile | Definition |
|---|---|
| Attributed revenue | Sales of tried-on items, counted under the attribution rules — same shopper session, order within your attribution window, only the tried-on line at the price actually paid, refunds netted. The strict version of the story. |
| Buy rate after try-on | The percentage of try-on sessions that went on to purchase. |
| Shoppers who opened it | Widget opens in the period — every open counts; unique sessions are shown separately in the Funnel tab. |
| Total try-ons | Renders in the period, with successful ones counted separately. Failed renders release their credit and don't count. |
Next to the time range you can set how long after a try-on an order still counts: 7, 14, or 30 days. Every stat is labeled with the window in effect, so numbers are never ambiguous. Billing statements always stay on the standard 7-day basis regardless of what the dashboard shows — the dial changes your analytics view, never your invoice. Full mechanics in how attribution works.
The session funnel: widget opens, then try-ons, then cart adds, then purchases — each step a smaller share, and the biggest drop is where your attention should go.
Never sum the per-product rows: The per-product view is product-influenced: an order containing two tried-on items is credited under each of them, so the rows intentionally add up to more than the store total. The store-level number is order-deduplicated — that's the headline. The two won't match, by design.
Export CSV offers five raw datasets for your own analysis: try-on events, widget events, cart events, purchases, and aggregated sessions — each scoped to the selected time range.
On the Free plan: Free includes the core layer — the headline band and funnel basics. Engagement, Preview, and the advanced Performance views unlock on paid plans, and because everything is tracked from day one, upgrading reveals your full history immediately.
A shopper ran a successful try-on and then bought that same product within the attribution window, in the same shopper session. Only the tried-on line items count, at the discounted price actually paid — never shipping, taxes, or the rest of the cart — and refunds are subtracted when they happen.
Opens count every time the widget is opened; unique sessions count distinct shopper sessions that opened it at least once. Device splits use distinct sessions, not raw opens.
Per-product rows are product-influenced — a multi-item order is credited under each tried-on item in it — while the store total is order-deduplicated. Read the store total as the headline; use per-product rows for ranking, not summing.
Yes — the attribution window selector offers 7, 14, or 30 days, and every stat labels the window in effect. Invoices and billing statements always use the standard 7-day basis regardless of the dashboard setting.
A product tried on frequently but purchased rarely — at least five try-ons with a conversion rate at half your store average or less. It's a signal to check that product's photos, description, or pricing.