Stock counts from three days to one afternoon

The finding that changed the project

The brief we were given was "replace the accounting system". The discovery found something else.

Sitting through a real stock count made it obvious that the numbers were not wrong because of the software. They were wrong because roughly a fifth of stock movement was never recorded at all: inter-branch transfers done on a phone call, samples handed out at the counter, damaged goods binned without a movement.

No system, however good, produces accurate stock from incomplete input. We said so in the discovery report, and it changed the shape of the engagement: less custom development, more operational change, and a much better outcome.

What we built

A single Odoo instance replaced both the POS integration layer and the accounting package. Sales, stock and payments now land in one place with no export step, which removed an entire category of monthly reconciliation work.

The custom development was smaller than expected: inter-branch transfer handling and a variance dashboard by branch. Everything else was standard Odoo configured properly.

The unglamorous part that produced the result

Barcode discipline and cycle counting. Every movement scanned, counts on a rolling schedule, variances investigated where they happened, within the week.

That is a process change, not a technology one. It is also why the stock figures are now trusted, and it would have worked less well without the system change underneath it.

What we would do differently

We under-estimated branch training. Four branches were straightforward; the two with high staff turnover needed a second round and a laminated one-page guide at each till. On the next rollout of this shape, that guide exists before go-live.