A multi-format retail group — supermarkets, quick-service restaurants, and a central distribution warehouse — needed one trustworthy view of sales, margin, and cash across every location, every day. Here is how we built it.
Sales, stock, and margin data lived inside the group's ERP but reached decision-makers as static, manually-built spreadsheets — often days after the fact. There was no single, shared view of performance across more than twenty-five stores, restaurants, and a distribution network, and no way for a store manager or executive to simply ask a question about the numbers.
Rather than starting with charts, we started with the data layer: a governed warehouse fed by real-time and scheduled extracts from the core ERP, with clear staging, reporting, and analytics layers — so every number downstream traces back to a known source.
The same governed data now powers a web dashboard for daily operations, an automated financial reporting suite for finance, and an AI assistant that briefs executives every morning over WhatsApp — with no re-keying between them.
A scheduled and near-real-time pipeline extracts sales, stock, transfer, and general ledger data directly from the group's ERP and lands it in a modelled warehouse, organised into staging, reporting, and analytics schemas — the single source of truth for everything downstream.
Daily and intraday sales by store, department, category and product, stock-on-hand, and distribution service-level tracking — with role-based access so a store manager sees their own site and an executive sees the whole network.
A monthly flash profit & loss with budget, prior-month, and prior-year comparisons, plus a full statutory reporting pack — Statement of Profit or Loss, Statement of Financial Position, and Statement of Cash Flows — recomputed live from the general ledger on every page load.
Machine-learning models for demand forecasting, price and margin anomaly detection, and basket analysis, paired with a WhatsApp assistant that delivers a daily briefing and answers free-text questions from the same underlying data.
A reporting platform is only as trustworthy as its honesty about its own limitations. Where a figure doesn't fully reconcile to source — because a balance is tracked in a system we don't yet have access to, or because a subledger doesn't capture every posting — we say so directly in the interface, in plain language, rather than presenting a polished number that quietly hides the gap.
Every statutory statement we build ships with an explicit balance check and a cash-flow reconciliation, shown to the user as a pass/fail — not buried in a footnote.
Store, restaurant, and warehouse locations reporting into one platform
Automated executive briefings, replacing manual end-of-month decks
Statutory financials available on demand, not just at period close
Free-text Q&A over sales, margin and stock — no analyst required
Explore an interactive version of this kind of dashboard, built on public sample datasets.