Every revenue number Ello shows a merchant is built from a ledger that can be audited line by line. These are the rules — all four must hold, every time:
Ello counts the tried-on items' line value only — after discounts, before shipping and taxes. Example: a shopper tries on a $38 shirt, then checks out with the shirt plus an untried $60 tote, $8 shipping and $5.12 tax. The order total is
Shopify refund webhooks feed the same ledger: if a tried-on item is refunded within 45 days of purchase, its value is subtracted. The number merchants see (and the number Ello's enterprise revenue share bills against) is the net one.
Ello's session identity lives at most about a week, so attribution can never stretch further than that — unlike the 30-day view-through windows common in ad attribution. Most attributed purchases land the same day as the try-on.
View-through conversions; untried items in the cart; shipping and taxes; modeled or projected revenue; industry benchmarks dressed up as store numbers; anything without a Shopify order ID.
Beyond attribution, Ello ships a built-in A/B holdout: a random slice of the store's own shoppers browses with try-on hidden, and the dashboard reports conversion with vs. without, plus a statistical-confidence verdict. Measured, not modeled.