Inventory snapshots with CSV Reconcile
Monday and Friday stock snapshots are matched on a combined key of warehouse and sku, with an allowed difference of 2. A change of one unit in a count is ignored, a drop from 40 to 32 is not, and the Columns tab totals the movement of qty and value.
- Export the stock report from each date as CSV.
- Select two key columns, here warehouse and sku, because the same SKU exists in several warehouses.
- Under Comparison rules set an allowed difference if counts or values are rounded.
- Press Compare and open Columns for the total movement of each numeric column.
- Open Changed for the lines, and Added or Removed for stock lines that appeared or disappeared.
What to know
A SKU alone is not a key in a multi-warehouse report: A-100 exists in the AUS and DAL warehouses, and a comparison on sku alone would treat them as duplicates. Selecting warehouse and sku together makes each stock line unique. The combined key is matched as a whole, so AUS plus A-100 is never mixed up with DAL plus A-100.
The allowed difference of 2 is a deliberate choice, not a default. In the sample, AUS A-100 went from 120 to 119 units: within the tolerance, so it is not reported as a qty change. Its value, though, moved from 2,880.00 to 2,856.00, a difference of 24, which is above the tolerance and is reported. AUS A-101 lost 8 units and 472.00 in value. Totals on the Columns tab: qty changed in one row by -8, value in two rows by -496 in total.
Pick the tolerance from how the numbers are produced. A count from a handheld scanner has no rounding at all, so use 0. A valuation computed from a cost average often differs by a cent or two, so a small tolerance removes noise. A percent tolerance is relative to the value in the first file, which suits large values. Whatever you choose is written to the Rules sheet of the Pro Excel report, so a reviewer can see what was ignored.
Updated · Cosmovex
Questions
Why use two key columns?
Because one SKU can have a stock line in several warehouses. The pair warehouse and sku is unique, a single column is not. You can select up to three columns.
How does the allowed difference work?
Two numbers are equal if they differ by no more than the absolute allowance or by no more than the percentage allowance of the first file's value, whichever is larger. Both default to 0.
Are numbers with thousands separators compared correctly?
Yes. Values such as 2,880.00 are read as 2880 before they are compared. A column that is not numeric in both files is compared as text.
What does the Columns tab total?
For every compared column it shows how many rows changed and, where both values were numbers, the sum of the differences (second file minus first).
What if a warehouse code is missing in some rows?
Rows whose key cells are all empty are left out and a note says how many. A row with only part of the key filled is matched on what it has.
The free plan covers everything on this page. CSV Reconcile Pro ($29, paid once) is described on the CSV Reconcile page.
