Missing values in key columns
Customer IDs, emails, or SKUs blank on otherwise complete rows. The import may still succeed.
CSV files
Audit CSV data before you import or trust it
A CSV is easy to open and easy to underestimate. Downstream systems often accept the file. That does not mean the records are ready. The audit happens before the import — while the problems are still in one place.
This is about reviewing the file, not about replacing your CRM or warehouse.
The actual problem
The file can parse and still be the wrong source of truth
CSV looks like a table. Operationally it is a contract: columns mean something, identifiers should be stable, and categories should be controlled. When that contract is loose, every downstream mapping inherits the mess.
Teams export from one system, open the file in Excel, save it, then import it somewhere else. Along the way dates shift, leading zeros disappear, extra columns appear, and two versions of “United Kingdom” coexist.
Auditing CSV is not explaining what a comma-separated file is. It is asking whether this particular file should be allowed to become someone else’s system of record.
What can go wrong
Problems that survive a successful parse
Duplicate rows and reused keys
Exact copies, or the same ID describing two different entities.
Dates that do not mean one thing
08/11/2026, 2026-08-11, and 11-Aug-2026 in the same column.
Categories that fragmented
Active, ACTIVE, and Current treated as three statuses in reporting.
Numbers that are not numbers
Currency symbols, mixed decimals, or IDs stored in scientific notation.
Structure that shifted
Extra header rows, blank columns, broken quoting, or encoding that only breaks for some names.
Outliers the importer will not reject
A quantity or amount that is technically valid and operationally implausible.
Example scenario
An export leaving one system for another
A migration spreadsheet is built from a CRM export. The CSV opens cleanly. Mapping looks obvious. Then the receiving system starts creating partial records.
| Customer ID | Country | Status | Opened |
|---|---|---|---|
| C-1044 | UK | Active | 2026-08-02 |
| C-1044 | United Kingdom | active | 08/02/2026 |
| United States | Current | 2026-07-18 | |
| C-1180 | USA | Active | 18/07/2026 |
The importer can create two customers for C-1044, drop the row with a blank ID, and store two date conventions as different values. None of that requires the CSV to be “corrupt.”
What to investigate
Check structure, then meaning, then outliers
- Headers: duplicates, blanks, unexpected extra columns, values that no longer match the intended mapping.
- Identifiers: uniqueness, missing keys, IDs damaged by spreadsheet editing.
- Types: dates, numbers, booleans, and categories that do not match a controlled set.
- Row grain: does each row represent one entity, or a mix of people, companies, and events?
- Values that pass format checks and still fail context — amounts, quantities, impossible dates.
Where Auditere fits
Review the file while it is still a file
Auditere is an investigation layer for structured business data — including CSV and spreadsheet exports — before those records are treated as authoritative inputs.
Use it to surface missing information, duplicates, suspicious values, and conflicting fields, then review them with the surrounding row in view. Fixes stay explicit. The product does not claim to be a native connector for Salesforce, HubSpot, or any other CRM.
Practical checklists live on the blog: cleaning a CSV before CRM import, Salesforce-specific import checks, and HubSpot identifier and association issues.
Audit the CSV before it becomes someone else’s problem
Start a trial and investigate structured files with detect, review, resolve, and verify — under operator control.