5 ERP Import Mistakes SMEs Make (and How to Avoid Them)

ERP import failures cost small and mid-size manufacturers far more than the time spent correcting files. A rejected catalog upload delays purchase orders, stalls production planning and erodes trust between operations and IT. The frustrating part: most failures are caused by the same five mistakes, repeated across organisations and ERP platforms.

Mistake 1: Uploading files without pre-validation

The most common mistake is also the most straightforward to fix. Most teams submit their catalog file directly to the ERP and discover errors only after the system rejects rows — sometimes hours or days later, depending on batch processing cycles. Every rejected row then requires a correction cycle: identify the error, fix the source file, re-submit, repeat.

Mistake 2: Ignoring character encoding

Catalog files exchanged with suppliers frequently arrive in Windows-1252 or ISO-8859-1 encoding. An internal file edited on a Mac and re-saved may convert to UTF-8 with a BOM. The ERP expects one encoding; the file delivers another. The symptom is garbled special characters — accented letters, currency symbols, degree signs — that appear valid in Excel but cause field-level failures in the ERP.

• Always specify encoding when opening CSV files in Excel (Data → From Text/CSV)
• Save as UTF-8 without BOM for maximum ERP compatibility
• Validate encoding as part of every pre-upload check

Mistake 3: Multiple contributors, no shared format

When four sales representatives, two purchasers and one external supplier each maintain a section of the catalog, formatting inconsistencies are inevitable. One contributor writes prices as "1 249,00 €". Another writes "1249.00". A third leaves the currency symbol out entirely. Each format may look correct to a human reader. None of them may be correct to the ERP.

Mistake 4: Missing required fields on a subset of rows

Required fields are well understood in theory. In practice, a file may have supplier codes populated for 98% of rows and absent for 2% — the rows added last week by a new contributor who did not know the field was mandatory. The ERP rejects those rows. The team discovers the omission three days after submission.

The key risk is not the obviously incomplete row (where the supplier code column is entirely empty). It is the row where the value looks present but is formatted incorrectly — a space, a trailing character, a code from a previous ERP that is no longer valid in the current system.

Mistake 5: No audit trail between validation and submission

When an ERP import fails a month after a clean validation, the team cannot answer a basic question: was the file that was validated the same file that was submitted? Without timestamped validation reports, it is impossible to determine whether the error was introduced after the last clean check. Audit trails are not a nice-to-have; they are a diagnostic requirement for any import process that involves multiple contributors.

The common thread

Each of these mistakes shares a root cause: validation happens too late in the process, or not at all. Moving validation to the moment before submission — not the moment after rejection — converts a reactive, IT-dependent correction cycle into a proactive, self-service quality step owned by the operations team.