How do you move spreadsheet data into a new business app?

Agree what each record means, rehearse the import, and verify the result before switching daily work.

Start with the records, not the upload button

Suppose a service business replaces its job-tracking workbook with an internal app. The file imports without errors, but some customer codes lose their leading zeros and completed jobs appear in the active queue. The transfer worked technically; the business meaning did not survive.

This is an illustrative scenario, not a client result. Before requesting a migration estimate, identify the authoritative workbook, its owner, and the sheets the team actually uses. Include linked files, lookup tables, formulas, and any manual conventions such as colours that indicate a blocked job.

Decide which history belongs in the new app and which can remain in an accessible archive. Moving every old column can recreate confusion rather than remove it. Keep an unchanged source copy and record the export date so the trial can be reproduced.

Write down the rules hidden in the sheet

A short mapping document should explain the destination of each field and any transformation. Do not let an importer silently decide whether an empty amount means zero, whether two similar customer names are the same company, or which date format is intended.

Give records stable identifiers where possible. Keep a mapping between old and new identifiers when the app generates its own. A name or email address may change; related jobs still need to point to the right customer.

Source detailDecision to makeExample check
Customer or job codeStore as an identifier, not a quantity.Code 00124 remains distinct and keeps its zeros.
Date and timeAgree format, time zone where relevant, and blank handling.An ambiguous date is flagged rather than guessed.
Status and colourTranslate every used label or visual convention.A blocked job stays blocked after import.
Formula or calculated totalChoose whether to migrate the result or rebuild the rule.A known example gives the expected result in the app.
Repeated customer nameAgree matching and review rules.Similar names do not automatically merge separate customers.

Find exceptions before they become app records

Review missing required fields, repeated identifiers, unexpected statuses, and values that do not fit the agreed rules. Separate records needing a decision from records ready to import. Give each exception an owner; deleting awkward rows is not a migration strategy.

Microsoft documents Power Query profiling tools for column quality, value distribution, and statistics. It also notes that profiling defaults to the first 1,000 rows, with an option to inspect the entire dataset. If you use that tool, check the scope before calling the workbook reviewed.

Ask the person who knows the work to resolve uncertain meanings. Keep a record of approved corrections and transformations so the production import uses the same rules as the rehearsal.

Rehearse, then check more than totals

Use an isolated test destination and representative records: ordinary jobs, older entries, blanks, cancellations, duplicate-looking customers, and linked records. Use synthetic or appropriately protected data, and keep test imports from sending real messages or starting live workflows.

AWS DMS illustrates the distinction between loading and validating data: its validation compares source and target rows and reports mismatches. That is a database-specific capability, not a recommendation to use DMS for a workbook. For a spreadsheet project, ask your supplier for equivalent evidence suited to the app and agreed transformations.

Compare counts by meaningful groups, key field values, and relationships. If two source rows intentionally become one destination record, explain that difference rather than demanding identical totals. Check that each imported job opens under the right customer and that an authorised staff member can complete a normal task.

Ask what happens if the same file is submitted twice or the import stops partway through. The trial should show how repeats are detected, which records were rejected, and how unfinished work is recovered without duplicating accepted records.

Choose when the new app becomes authoritative

Agree the last time staff may edit the old sheet, who captures changes made after the rehearsal, and where new work goes during the switch. For a small workflow, a planned editing pause may be simpler than maintaining two editable systems. If both must remain active, define which system owns each field and how differences are reconciled.

AWS's cutover guidance calls for validation and explicit rollback criteria, data handling, and a decision owner. It also warns that the old system can become stale after new transactions begin. Applied to a workbook migration, restoring yesterday's spreadsheet is not enough if today's jobs exist only in the app.

Write a fallback that accounts for records created or changed after the switch. Agree who can stop the release and how staff will be told which system to use. Do not delete the old source immediately; agree archive access and retention appropriate to the business.

Use a six-point acceptance checklist

Ask for these answers before approving the move. The right level of testing depends on the records and consequences, but each item should have an owner and visible evidence.

  1. Scope: the authoritative files, included history, and excluded or archived data are listed.
  2. Mapping: field formats, formulas, statuses, identifiers, and duplicate rules are approved.
  3. Rehearsal: representative records load correctly, and rejected records have explanations.
  4. Reconciliation: agreed counts, values, and relationships match; approved differences are documented.
  5. Switchover: the editing cutoff, final changes, staff instructions, and go/no-go owner are clear.
  6. Recovery: repeat imports and partial failures are tested, with a fallback covering new data.

Make migration part of the project scope

Separate building the app from preparing and moving its data. Ask who cleans the source, resolves exceptions, runs the trial, approves results, and supports the first working day. This avoids an attractive new interface arriving with an unfinished operational handover.

Bring KazeZeus a description of the workflow, your field list, and a safe sample of the difficult cases. Udbhav can scope the internal tool and migration together, including the checks needed before your team relies on it.

Sources & further reading

Apply this to your business.