One platform for a process that used to live in seven places

The client is a bakery & pastry producer with recurring orders, multiple customer types and a delivery operation where speed and accuracy aren't preferences — they're the product. The company processes an estimated 600–1,500 orders per month, and each one ripples through sales, production, quality, logistics, transport, accounting and shift-based office operators.

Every customer added its own twist: different platforms, different document formats, different ways of working. Some had integrations with third-party platforms; others sent corrections on paper — or as a photo of a hand-modified delivery note.

The same order was retyped again and again on its way through the company: read from a PDF, entered into an Excel, consolidated into a bigger central Excel, copied into a separate "Routes" file for drivers, then entered once more into the accounting system. Printed sheets carried information between office and production. The daily friction:

• Excel as operational infrastructure — not a helper file, but the place where orders, production prep and delivery data actually lived, with no audit, traceability or sync.
• The same data entered 3–4 times, with every re-entry adding 15–25 minutes of cumulative admin work per order — and another chance for error.
• Order changes were the worst case. A late change had to be propagated by hand through both Excels, batches, pallets, routes, documents and accounting — 25–45 minutes each, an estimated 150–220 times per month.
• Standard delivery notes that didn't fit every customer, forcing manual completion and adjustment of documents — sometimes corrected with a pen at reception.
• No shared picture of reality. No single place showed active orders, current batches, pallet groupings, planned routes or what could be invoiced — reality had to be reconstructed from files, sheets and conversations between shifts.

At an estimated 2–4% error rate, that meant 30–60 operational corrections a month — wrong quantities, misassociated batches, outdated routes — and roughly €4,000–€7,000/month in manual administrative cost.

A generic order-management tool would have solved a slice of the problem and hard-coded rigidity everywhere else. This operation had real particularities — orders arriving in different formats, customers on different platforms, frequent changes, a direct line from order to production batches, pallets and routes. The platform had to be shaped around that.

• One source of truth. Orders, batches, pallets, routes and documents connected in a single system, instead of synchronized by people across files.
• Automate the re-entry away. Integrations with the customers' third-party platforms bring orders in automatically — no PDF reading, no retyping.
• Make change cheap. The most expensive event in the old process — a late order change — became the platform's defining feature: batches recalibrate automatically instead of being recalculated by hand.

The right description of this project isn't "an order management tool". It's a custom platform for digitalizing the operational flow between orders, production, logistics and documents in a working bakery plant — one where information used to be reconstructed from spreadsheets, paper and memory, and now lives in a system every team reads from.

The pattern travels. Any producer whose orders drive production and delivery — food manufacturing, FMCG, fresh distribution, any plant with batches, pallets and routes — has a version of the same problem: the order changes, and people pay the price. The same operational principles solve it.