How to Migrate Inventory From Spreadsheets

An inventory spreadsheet migration is not a file conversion. It is a decision about which identities, quantities, locations, units, costs, and histories will become authoritative. Importing every column can preserve contradictions inside a cleaner interface. Start by defining the target operating model, then prepare evidence for the opening state and keep the original records available.

Buyer scenario: one SKU means several things

Imagine a product company with separate worksheets for purchasing, warehouse counts, sales, and returns. The same code identifies a case in one sheet and an individual unit in another. Old items share codes with replacements, locations use inconsistent names, and a formula treats reserved customer demand as physical stock.

The migration team should not resolve these conflicts silently during import. Each ambiguous identity needs an owner and documented decision. GS1 US describes how GS1 identifiers are constructed and assigned, but internal records may also use other governed identifiers. Software cannot decide whether two rows represent the same trade item, package level, asset, or location.

Create a decision log that links every transformation rule to sample source rows and an approver. Include trimming, capitalization, unit conversion, duplicate mapping, inactive status, and location normalization. Automated cleanup is acceptable when the rule is visible and reversible. Rows that do not meet a rule should enter a quarantine list with an owner; they should never inherit a guessed identity simply to make the import report complete.

Decision criteria for migration-ready data

Define required fields for item identity, description, status, base unit, purchase unit, sales unit, conversion, location, quantity state, tracking identity, and accountable owner. Separate physical on-hand from available, reserved, in-transit, damaged, and pending-return states. Retire duplicates through a mapping rather than deleting their history.

For opening values, document the source, cutoff time, count method, approval, and later correction process. IRS guidance permits electronic recordkeeping while requiring records that clearly support the business's activity. Retain source workbooks and reconciliation evidence according to the business's record policy; an import success message does not prove the opening data is correct.

Reproducible migration evaluation and rehearsal

Create a representative extract containing a normal item, a variant, a case-to-unit conversion, a bundle component, an inactive item, a duplicate identifier, negative quantity, and a tracked identity if relevant. Transform it using documented rules, import it into a nonproduction environment, and compare row-level exceptions with expected outcomes.

Run purchase, receipt, transfer, order, shipment, return, and adjustment scenarios on the imported records. Export the result and reconcile identity and quantity back to the approved source. Repeat after fixing the transformation rather than editing records invisibly. Inventory Software Guide has not executed this buyer migration; teams should rehearse with their own data and preserve each decision.

Edge case: transactions continue during cutover

Stock may move after the physical count or extraction but before the new system becomes authoritative. Define a freeze window or a controlled delta log. Assign one person to approve late receipts, shipments, returns, and adjustments, then apply them in sequence to the opening position.

If a connection publishes during the gap, capture its events and decide whether to replay or reconcile them. Prevent duplicate orders and quantity updates. The target system should not go live until users can explain the opening balance plus every controlled delta.

Conclusion: migrate governed truth, not workbook clutter

Successful migration means the new system begins with understandable identities and reconciled states, while the source evidence remains available. Clean only through documented mappings, rehearse representative exceptions, and control the cutover gap. If an ambiguous row cannot be resolved, quarantine it rather than inventing certainty. A smaller trusted opening catalog is safer than a complete import nobody can explain.