Your Order Still Walks Between Systems

Forget "integration strategy" for a moment and just follow one order instead.

A customer emails a PO. Someone on the order desk copies it into a spreadsheet first, because that is how they check margin, then keys it into the ERP. Warehouse picks from a printout that may or may not match what sales actually promised, and despatch confirms on a shared drive list nobody officially owns. Finance invoices from a third version of the truth, then spends Friday chasing why the credit note does not match the return that operations swear was already booked.

That order did not flow - it walked, person to person and system to system, clipboard to screen, and every step along the way was a place to drop it, delay it, or quietly invent a new number for it. This is not abstract architecture. It is your order-to-cash and purchase-to-pay reality, and until you name the journey honestly, you will keep buying systems that do nothing to shorten the walk.

Name the journey - out loud

Take order-to-cash and write the stations on a whiteboard: enquiry or PO in, commercial check, order entry, credit check, pick, pack, despatch, proof of delivery, invoice, cash application, exceptions and credits. Now circle every station that requires a human to re-enter data, export a file, or chase another team for confirmation - those circles are your real operating model, not the slide that says "ERP at the centre."

Do the same exercise for procure-to-pay: requisition, approval, PO, goods receipt, invoice, matching, payment. If invoices arrive by email and get typed while goods-in lives in a completely different tool, the walk is the process itself, not a symptom of it. People who live this every day do not need a maturity model to explain it to them. They need fewer circles on the whiteboard.

What the walk actually costs

Re-keying looks cheap because each individual instance is short, but it is not cheap once you add it up properly. There is the labour cost - minutes per order, multiplied by volume, multiplied by roles, with order desk, warehouse admin, and finance assistants effectively paid to act as human APIs between systems that should be talking directly. There is the error cost, where a wrong SKU, wrong price, wrong ship-to, or wrong tax code becomes its own chase through customer service, credit notes, stock adjustments, and supplier queries. There is the delay itself, as the order waits in someone's inbox while they finish another walk elsewhere, cash sits longer than it should, and customers get "I will find out" instead of an actual status. And underneath all of it sits a kind of management theatre, where ops directors spend meetings reconciling three versions of the same order book because no single system journey is trusted end to end.

A new ERP demo will always show a clean happy path with no email, no spreadsheet, and no "can you just..." That demo is not your Tuesday.

A new ERP will not stop the walk by itself

Replacement programmes sell end-to-end process, but by the time go-live arrives, ecommerce still sits outside, the WMS has become "phase two," and finance keeps a parallel spreadsheet because they do not trust the cutover. The order simply learns a new route. It still walks.

If your problem is handoffs, the first fix is handoffs - through data and systems integration and clear process ownership - not a platform swap that assumes the journey will magically become digital just because the logo on the login screen changed. We are not anti-ERP. We are anti buying a system to avoid confronting the walk. Sometimes ERP renewal is genuinely part of a wider redesign, but it is rarely a substitute for one.

Fix the handoffs before you fund the programme

The practical sequence starts with mapping one real order and one real supplier invoice end to end, including the informal steps - the WhatsApp confirmations, the "reply all" approvals, the shared mailbox nobody officially owns - because if it is not on the map, it will quietly survive the project anyway. From there, decide which systems are allowed to create truth at each stage: sales must not invent stock, warehouse must not invent price, and finance must not invent goods receipt. Boundaries beat heroics every time.

Automate the boring transfers next - order creation, acknowledgement, despatch confirmation, invoice generation, three-way match signals - because this is exactly where business automation earns its keep, not as a science project but as the removal of walking itself. Kill any spreadsheet that exists purely to move fields between two systems; if it is a staging area for an integration you never actually built, either build the integration or admit the process is unfinished. And measure chase volume properly: how many internal emails per hundred orders exist only to ask whether something has been entered, picked, or invoiced. That single metric tells you whether the walk is genuinely shrinking or just changing shape.

The practitioner test

If your customer asks where their order is, how many people does your team need to ask before they actually know? If the answer involves more than one system and more than one person, the order is still walking. You do not have a connected operation - you have a relay race with clipboards.

Vendors will offer you a better baton. Fix the handoffs first, make the journey digital and properly owned, and only then decide whether the core system needs to change - with evidence from a shorter walk, not hope from a longer sales cycle. Your order still walks between systems. Stop calling that normal, and start closing the gaps it keeps falling into.