Multi-site hospitality runs on thin margins and fast decisions, but most operators still assemble their numbers by hand, days after they mattered. Here is what it actually takes to get every restaurant, kitchen and channel into one reliable P&L view.
Multi-site hospitality runs on thin margins and fast decisions. A few points of food cost, a badly staffed shift, a slow-moving grocery line, across forty sites, small things compound quickly. And yet most operators are still making those decisions from numbers assembled by hand, in a spreadsheet, days after the week they describe. The problem is rarely a lack of data. It is that the data lives in too many places, arrives too late, and nobody quite trusts the version they are looking at. Here is what it actually takes to fix that.
Why hospitality reporting is harder than it looks
A restaurant group is not one business reporting once. It is dozens of sites, often on different systems, plus central kitchens, plus whatever grocery or retail channels have been bolted on, each with its own idea of what a sale or a cost is. The EPOS system knows takings. The purchasing system knows food cost. The rota system knows labour. The finance system knows the general ledger. None of them agree on the definitions, and stitching them together into a P&L that a general manager and a finance director both believe is where the real work sits.
Do it by hand and you get the familiar pattern: a heroic analyst rebuilding the same workbook every Monday, numbers that do not reconcile between sites, and decisions made on last-fortnight's picture. The reporting is not wrong so much as too slow and too fragile to act on with confidence.
What one view actually requires
Getting every site, kitchen and channel into a single trusted P&L is less about the dashboard and more about the foundation underneath it. Three things have to be true.
First, the data has to come from source automatically, EPOS, purchasing, payroll and finance flowing into one place without a person copying anything. For a smaller estate that can be direct connections; for a larger, multi-system group it usually means a proper data platform on Azure and Microsoft Fabric that lands, cleans and conforms the data before anyone reports on it.
Second, the definitions have to be governed. Food cost percentage has to mean the same thing at every site and in the board pack, or the whole exercise collapses into arguments about whose number is right. That is a modelling and governance job, not a reporting one, a single certified layer that every report reads from.
Third, it has to be timely enough to change behaviour. A P&L that a site manager sees the morning after, not the fortnight after, is what turns reporting into action, spotting the shift that was overstaffed while there is still a week left to fix the pattern.
What it took for one multi-channel food business
We rebuilt exactly this for Wasabi, a multi-channel food business running restaurants, central kitchens and grocery. The Microsoft estate underneath was underperforming, and the reporting sat on foundations that were not solid, the classic numbers people could not fully rely on problem. We rebuilt the data platform on Microsoft Fabric and Power BI, and the outcome was forty-plus restaurant P&Ls, central kitchen and grocery, in one view, with the programme now extending into machine learning for forecasting.
The headline is the single view. The thing that makes it durable is everything under it: the governed data layer that means every P&L reconciles, and the automated pipelines that mean nobody rebuilds a workbook on a Monday ever again.
Where to start
You do not have to boil the ocean. The right first step is usually one clean, governed P&L across your sites, proof that the foundation works, before extending into labour analytics, menu engineering or forecasting. Get the single source of truth right first, and everything else has something solid to stand on.
Related reading: cloud data platform and warehouse build and our analytics case studies.
Related reading
What the pub trade teaches about numbers
Basket analysis on data you already have
Forecasting and demand planning on the data you already have
From raw operational data to boardroom number
Simon Devine
Managing Director
Part of the Hopton Analytics team, delivering governed analytics programmes for UK mid-market organisations.
