A 42-store retail chain gave every manager the same numbers.

The chain sells kitchenware and home goods through 42 stores in France and Belgium and a growing e-shop. Store sales are recorded by a point-of-sale system on SQL Server, the e-shop runs on Shopify, stock lives in a Postgres warehouse management system, and store targets are set in a spreadsheet by the regional managers.

With a data team of two, reporting was a weekly bottleneck. The sales report reached stores on Thursday, too late to act on the previous weekend. Store managers built their own spreadsheets from POS exports, with their own definitions of like-for-like sales and their own treatment of returns, and regional meetings regularly ended in arguments about whose figures were right.

Meanwhile, stock-outs on best-selling products were costing sales nobody could quantify. Click-and-collect orders were placed online against stock that had already sold in store.

The chain connected the POS database, Shopify, the WMS, the targets spreadsheet and Zendesk to Kimo. The data team built a retail model with a small set of governed measures: net sales, like-for-like growth (stores open more than 13 months, returns netted), basket size, stock cover and stock-out rate on the top 200 SKUs.

Every store manager received a store dashboard filtered to their own store through row-level access, alongside a regional view and a ranking. The report refreshes overnight and is complete by 9:00 on Monday. Managers ask questions in plain language, such as “Which products are below one week of cover in my store?”, rather than emailing head office.

A stock-out alert watches the top 200 SKUs per store and warns when cover drops below five days, joining POS sell-through with WMS stock and inbound deliveries. Click-and-collect availability now reads from the same model, so the e-shop no longer promises stock that is gone.