Moving from Excel to Airtable: prepare the data migration

Inventory the workbook in use

Work from the file that is actually authoritative, not an old copy. Identify active sheets, formulas, lists, macros and links to other files. Ask who uses each result and for which decision.

Keep a dated copy before cleaning the workbook. A data migration does not automatically transfer every behaviour of the file.

Give rows a consistent meaning

Define what a row represents in each sheet. Separate data rows from headings, comments and subtotals. Merged cells or several objects in one row need preparation before import.

Keep an identifier to find each item in the source and destination. Clean duplicates using an explicit rule; two rows with the same name are not necessarily the same object.

Check types and values

Map each Excel column to its Airtable field. Test dates, numbers, percentages, codes with leading zeros and missing values. Automatic type conversion needs checking against the actual data.

For relationships between tables, validate matches using reliable identifiers. Do not rely only on the spelling of a name to link the right records.

Rebuild the calculations needed

Do not assume that an import rebuilds your calculations. Airtable's import documentation states that formula conversion uses the computed result and does not convert the formula itself. Identify which calculations need to remain active after transfer.

Recreate the required rules in the target system or retain a separate process, then compare results against known cases. Also check workbook-specific processes, particularly macros and external links.

Validate a pilot import before switching

First import a sample that includes difficult cases. Compare identifiers, row counts, fields and calculations with the source copy. Have relevant users carry out a normal task.

Then decide where new data will be entered and how to capture changes made during the test. Keep the source file and a rollback procedure until the checks are complete.

• An authoritative source and a dated copy.
• Consistent rows and identifiers.
• Checked field types and calculations.
• A pilot compared with the source before switching.