Methodology
A seller's business is spread across dozens of reports that arrive at different times. Rosetiq's job is to put the whole business on one screen — from the seller's own records — and turn it into decisions the seller can act on.
To give the seller the whole business at hand — what each product earns after every fee, what to reorder and when, where the advertising budget goes — with the same numbers on every page, because every page reads the same ledger. And to point at what needs attention rather than wait to be asked: cover falling below the seller's floor, a restock date approaching, a product tripping a rule the seller wrote, a note pinned to the day something changed. We work alongside the seller on this; the aim is a business that grows on its own numbers rather than on guesses.
Every figure is a calculation over records obtained from the seller's own accounts — nothing is estimated from industry averages or carried over from another seller, and no figure shown as a result is produced by a statistical or AI model. Profit is assembled from the account's financial events (revenue, marketplace fees, promotions, refunds, reimbursements), the seller's own cost of goods and the advertising reports. Fees reported per order are attributed to the order's SKUs directly; charges reported only at account or campaign level (storage, brand and display advertising) are allocated across SKUs in proportion to sales share. Days of cover come from the seller's recent sales rate and reported stock, including units in transit and reserved. On request, we show which records produced any figure and over which dates.
Different data arrives at different speeds, and Rosetiq refreshes each kind as often as the platform allows, so the seller sees the most current picture the platform can give.
| Data | Refreshed | How |
|---|---|---|
| Orders, offers, stock changes | Within minutes | The platform's event notifications, processed continuously through the day |
| Advertising performance | Hourly, then daily | An hourly performance stream during the day; the platform's daily reports once it has finalised them |
| Stock on hand, inbound | Several times a day | Fulfilment reports whenever the platform makes a fresh one available, at least nightly |
| Sales, traffic, search performance | Daily | As the platform publishes them; the platform itself revises recent days for about a week |
| Financial events, settlements | Daily | Charges and fees as the platform posts them; a day is considered settled once its events stop changing |
Each figure carries the state of its day: preliminary while the platform is still reporting, settled once its reporting has stabilised. Preliminary days are marked as such and revised as records arrive.
After each recomputation, automated checks confirm that the parts of a day sum to the whole and that no day was written from an incomplete source. A day that fails is held back and flagged rather than published with a wrong figure. Fees for orders the platform has not yet settled are estimated from the same product's settled fee rates over a recent window and replaced when the settled figure arrives.
Every record carries its seller's identifier, every query filters on it, and the database enforces the same separation through row-level security. No figure is ever aggregated across sellers, and no benchmark or comparison in the product is derived from other sellers' data.
Rosetiq