Data migration is where Salesforce projects most often surprise their teams. Records that looked fine in the old system load with errors, relationships break, and the first reports after go-live disagree with what people remember. A repeatable checklist turns migration from a scramble into a procedure. This guide lays out the steps from scoping to post-load validation.
Talk to an Expert →Decide what moves. Typically accounts, contacts, open opportunities, active cases, assets, and key reference data move, while stale history stays in an archive. Assign an owner to each data set who can answer questions about meaning and quality. Without owners, mapping questions float for weeks. Also agree the cutoff date and the freeze rules for the old system, so the data you load is consistent.
Fix problems before they cross: merge duplicate accounts and contacts, standardize country and state values, normalize phone and email formats, and remove records no one uses. Cleansing in the source is repeatable and reduces load errors. Where you must transform during loading, keep the rules in scripts or documented steps, so each trial load applies them the same way.
Create a mapping document for each object: source field, target field, transformation, and rules for blanks. Pay particular attention to picklist values, since values that are not in the target list will fail or create inconsistent data. Load in an order that respects relationships: accounts before contacts, contacts before opportunities, products and price books before quotes. Keep external identifiers from the source on each record so child records can find their parents and the migration can be reconciled later.
Our Salesforce practice →Decide who owns each loaded record. Assign owners deliberately or the loading user will own everything. If you want the original created dates and creators preserved, plan for the settings that allow it. Automation such as triggers, flows, and email alerts can fire on loaded data and send unwanted messages, so disable or control it during migration and re-enable it after validation.
Rehearse in a sandbox at least twice with full-volume data. After each load, compare record counts with the source, sample records for accuracy, check relationships, and run reports that business users recognize, such as pipeline by owner or open cases by queue. Log every error and its cause, and fix the rule rather than patching the data by hand. By the final rehearsal the process should be uneventful.
NetSuite integration services →Loading the data is not the end of the migration. Ask business owners to review a sample of records in the real interface, run the reports they trust, and confirm that totals and key accounts look right. Re-enable automation in stages and watch for unexpected behavior. Keep the source system read-only for a defined period, and record the final counts and sign-offs, so questions months later can be answered with evidence instead of memory.
Share where you are today and a Cold Sun consultant will recommend a practical next step.
Talk to a Salesforce Expert →
Active accounts, contacts, open opportunities, open cases, and reference data. Older history is usually archived.
At least two full-volume rehearsals, each followed by reconciliation and sign-off.
They preserve relationships between records during loading and let you reconcile Salesforce with the source later.
By disabling or controlling automation and notifications during migration, then re-enabling after validation.
Often yes, with the right settings enabled for the load. Plan for it in advance.
Yes. We profile, cleanse, map, load, and reconcile your data, with rehearsals before cutover.