Owner Intelligence

Each country ran its POS as a separate, unlinked network: its own credentials, its own currency, its own set of locations. The loyalty program lived in a third system, delivery aggregators in a fourth, costs in spreadsheets.

Because of that, a simple question — «how did the week go» — turned into a project: collect exports, normalize currencies, remember where a promo ran, and still end up with a week-old picture.

At that latency, management decisions are made on instinct. A location's decline shows up in money a month later instead of as a signal the next morning.

Overnight collection and a morning digest. The loop pulls sales, checks, channels, loyalty and platform ratings, computes weekday medians, and by nine delivers yesterday to the owner in one message.

Anomalies against the location's own norm. Deviation is measured not «versus last year» but against the median of the same weekday over previous weeks — so a promo doesn't look like a problem and a quiet decline isn't lost inside overall growth.

Layers by priority. On top of the base layer came P&L by location, the menu plotted on demand × margin, channels with the real contracted commission rates, loyalty with RFM and cohorts, operations control and a fraud radar, and reputation.

An advisor on top of the numbers. A question is asked in words; the answer comes with figures broken down by location and channel, and recommendations are stated in money — what the deviation costs and what to do about it.

Three real deviations on the very first run. Not «anomalies in general», but specific locations deviating from their own weekday norm — with amounts and an hourly breakdown.

The main lever sits in the menu. Almost half of the chain's revenue comes from items priced below the margin norm. That is not a «remove the dish» decision but a pricing and cost decision, visible only once demand and margin are plotted on the same plane.

A gap between loyalty and reality. Cards with a high visit counter and zero orders in the archive are not champions — they are staff cards. While loyalty was measured by the counter, the chain was rewarding the wrong people.

Aggregator cost in money. Once real contracted rates replaced averaged ones, it became visible how much the chain gives away in each city — and where a channel brings turnover but not profit.