Running a stock count
A stock count is a session: you open a sheet, walk the shelves entering what is really there, and apply the whole thing at once. Nothing changes until you apply, so a half-finished count is safe to leave.
Starting one. Count stock opens a sheet listing the variants in your current view — so filtering first is how you count one shelf or one supplier rather than the whole catalog. The sheet keeps its own name, who started it, when, and how many of its lines have been counted so far.
Three states. A count is in progress while you work, applied once you commit it, or discarded if you abandon it. Applied and discarded counts stay on record, so a disputed number later has a history to check.
What applying does. Every line where your counted figure differs from the system's becomes a correction in the stock ledger, with the count as its reason. It does not silently overwrite: the movement is recorded, so the history afterwards explains exactly what changed and why.
Counting by scanner. Scanning a barcode jumps straight to that variant, which is what makes a real stocktake practical — one person with a scanner rather than two with a printout.
Counting from a spreadsheet. You can import a count instead: a CSV with one column identifying each item, by SKU or barcode, and one holding its new count. Nothing is written until you review what the file will do.
Why bother. Committed stock, returns that were written off, and units that walked out of the back of the shop all drift the numbers apart over months. A count is the only thing that re-anchors available stock to what is actually on the shelf, and available stock is what your storefront is allowed to sell.