Back to resources

Find the file problem before it becomes a pay question

Check the structure, dates, signed amounts and staff attribution of an itemised salon sales CSV before reviewing commission.

A sales export can open perfectly in a spreadsheet and still be the wrong input for a commission run. It may contain report totals, name the cashier instead of the earning person, or mix date conventions. Trimsum separates file validation, column matching and preview from the later decisions about rates and exceptions.

The useful part

First establish what each row and column means. Then check whether the file can be read. Passing those checks makes a file usable; it does not prove the underlying sales are correct.

Start with itemised evidence

Choose a file with one sale or refund item per row. Preserve its original reference, date, earning staff name, item description and signed sales amount after discounts. Use AUD amounts that include the tax supplied by the report. Subtotals and daily aggregates cannot provide the same trace to an individual service. Retain the untouched export before making a working copy so a later reviewer can distinguish source values from preparation changes.

Match meanings, not familiar headings

The Match columns step suggests recognised headers, but each selection is yours to check. Map the reference, invoice date, earning staff member, item and signed tax-inclusive sales. Category, tax and reported commission are optional. Save a named mapping for reuse; one uniquely matching set of exact ordered headers can apply automatically. A notice identifies its name and version. Other compatible mappings can be chosen manually. These conveniences preserve column choices, not the meaning of a changed report, so inspect the earning-person field and a known line each time.

Use signed amounts consistently

The fictional refund below reverses part of a service. Its sales and tax are both negative, giving a negative basis when supplied tax is excluded. The import rejects a non-zero tax amount with the opposite sign or a tax amount larger in magnitude than the sale. Blank tax means no explicit amount was supplied; zero is an explicit zero. That distinction can change the calculation under the workspace’s selected basis.

Fictional file check · AUD
LineSalesTaxBasis excluding supplied tax
Original service220.0020.00200.00
Partial refund−110.00−10.00−100.00
Incorrect refund tax sign−110.0010.00Import rejected

Read a failure as a repair instruction

The parser requires a header and at least one sale, unique non-empty headers and a consistent number of columns. It reads comma, semicolon or tab-separated files and reports broken quoting. Files must be smaller than 8 MB and contain at most 15,000 sales. A bad date, blank reference, blank item or invalid money value rejects the import as a whole. Correct the working copy or export again; no valid subset is silently added.

Use the preview as a reconciliation checkpoint

A file can pass structural validation and still contain a wrong amount. Compare its coverage and totals with the original report before treating the preview as complete evidence.

  1. Confirm the chosen date convention for dates with slashes.
  2. Compare the preview row count, staff-name count, refund count and signed sales total with the intended report.
  3. Inspect the first five displayed lines for swapped fields or unexpected names.
  4. Choose whether this replaces or adds to an existing draft, then confirm the import statement.

Expect business questions after a valid import

Missing staff rates or links, uncovered sale dates, unknown categories, refunds, possible duplicates and commission differences appear in the review workflow as applicable. These are distinct from a malformed CSV. A successful import therefore begins the calculation review rather than bypassing it. If you only need to inspect a file’s structure first, use the public CSV health check without creating an account. Importing a real recurring run is the next step when you also need saved settings, decisions and a completed record.

Put it to work.

Inspect a working file before adding it to a commission workspace.

Check a CSV first

A record you can check.
A number you can explain.

Explore a sample run