Why it exists

This is the dashboard for the monthly meeting. It replaced a manual Excel build that took days and never reconciled between the two entities, because everyone converted currency with whatever rate they had on hand.

• Results by entity. Revenue, cost, and margin for each company, with its mix by branch and by trade direction.
• Billing against target. The month's progress against goal, using the same exchange rate on both sides so the comparison means something.
• Receivables aging and days to collect and to pay. How much is owed, since when, and to whom payment is due.
• Vendor spend. Concentration on the payables side.

Where the data comes from

Shipments, receivables, and payables come from the same warehouse as the rest of the suite. On top of that sits a fourth webhook with billing from the accounting ERP: some clients are billed directly there and never have a single invoice in CargoWise. If the target is built only from the logistics ERP it lands nearly 10% short and no one can tell why.

The exchange rate goes through a cascade of three providers: first the central bank's official rate, and if that doesn't answer, two public backups. The date of the rate that was actually used stays visible on screen — a margin converted at a rate no one can date isn't a margin, it's an opinion.

Each invoice's amount comes in the entity's functional currency, not the document's currency. Confusing the two fields distorts the figures by nearly eighteen times; it's written into the code right where that field is read.

Build decisions

• Never two currencies on the same axis. The separation comes from the data model, not a dashboard filter: the entity is classified by the invoice's company field, never by the shipment's branch.
• The cutoff is by posting date, not open date. That way it reconciles with accounting, and last month's number doesn't shift backward when someone bills late.
• Payables balances come back negative in the original feed: absolute value is taken before summing. It's the kind of detail you only discover once a total comes out backwards.
• Orphan invoices aren't dropped. Hundreds of them have no matching shipment in the feed; an inner JOIN would make them disappear from the total and the dashboard would understate the number.

What's different in this demo

Same production file with the data layer swapped out. The calculations, the cutoffs, and the currency rules are the real ones; what changes is where the numbers come from.

• Brand, clients, vendors, and amounts are made up.
• All three exchange-rate sources are intercepted and answer with a fixed value, so the cascade is visible and it's clear where the number comes from.
• The query panel answers with a local engine, not a model.
• Exporting to Excel is disabled: the viewer doesn't grant download permission.