A familiar monthly ritual: someone exports the store’s stock report, someone else exports the accounting figures, and the two are reconciled by hand in a spreadsheet. The gap is never zero. Everyone has stopped expecting it to be.
Where the drift comes from
- A sale on the website decrements stock; the same sale is entered in accounting separately, at a different moment
- Returns are processed in one system and remembered in the other
- A walk-in sale never touches the online catalogue at all
- Stock adjustments after a physical count are applied to one side only
- Cancelled orders release inventory but leave the invoice behind
None of these is an error by anyone. They are the predictable result of two systems each holding their own version of the same fact.
What integration actually means
The fix is deciding which system owns each fact, and making every other system ask rather than remember.
- One system owns stock levels. Everything else reads from it
- One system owns pricing. The storefront displays it; it does not store its own copy
- Orders flow one way, on a defined trigger, with a record of what was sent and what came back
- Failures are visible — a queue you can see, not a silent skip
The unglamorous part
Most integration work is not the connection. It is deciding what happens when the two sides disagree: a payment that succeeded while the order failed, a product that exists in one catalogue and not the other, a sync that ran twice.
Any integration can move data on a good day. The ones worth paying for are the ones that behave correctly on a bad one.
- E-Commerce
- ERP