Skip to main content
Data Migration

Salesforce Data Migration Tool Checklist: What to Test Before You Choose

Every vendor demo loads tidy Accounts and Contacts. Run your shortlist against a slice of your own awkward data instead, and score what comes out the other side.

Salesforce data migration tool checklist: a sample of source records run through a candidate tool and scored against a list of tests

Data Migration12 November 20265 min read

Most migration tools look the same on a comparison grid: field mapping, scheduling, error logs, a connector to Salesforce. The differences show up later, often weeks into loading, when Opportunities point at the wrong Accounts or a rerun doubles the Contacts. This Salesforce data migration tool checklist replaces the grid with tests you run on a copy of your own data in a sandbox, so you choose on evidence rather than on a demo.

Start with what Salesforce already gives you

Salesforce’s own guidance on choosing a load method points to the Data Import Wizard for loads under 50,000 records on supported objects. For larger or scheduled loads it points to Data Loader, which accepts CSV files of up to 150 million records. Volume alone is rarely why teams buy a tool. They buy one for load order, relationships, repeatable runs and proof that nothing went missing. So the useful question when you evaluate Salesforce data migration tools is simple: what does it do that Data Loader and a spreadsheet of record IDs do not? Every test below is aimed at that gap, and Data Loader makes a fair baseline to score alongside the shortlist.

Build a test pack from your hardest data

For each tool trial, pick a small slice, a few thousand records rather than millions, that contains your worst cases. A good pack includes:

Load it into a sandbox, and mask personal data before it gets there; our guide to data masking for sandboxes covers how. Before the first run, write down the expected results: record counts per object and a few control totals, such as opportunity amount by close year. A tool is judged against those numbers, not against how the run felt.

The checklist: eight tests

1. Relationships resolve without hand-built ID maps

The target org issues new record IDs, so every lookup has to be re-pointed. Check that the tool loads parents before children, handles self-lookups such as parent accounts, and copes with records that refer to each other. Then open twenty migrated records and confirm each lookup lands on the right parent. We explain why this matters in preserving the relationship graph.

2. A rerun does not create duplicates

Migrations are run many times before go-live. Salesforce’s upsert uses an external ID field to update a matching record or create a new one. Ask how the tool matches records on a second run, then run the same job twice. The counts must not move.

3. History survives

If created dates and creators matter for reporting, test them. Salesforce’s system fields reference says you can set audit fields such as CreatedDate and CreatedById when you create records. That covers Accounts, Contacts, Opportunities, Cases, Tasks, custom objects and several others, once the “Set Audit Fields upon Record Creation” permission is enabled and granted. The values are set at creation, not afterwards, so confirm the tool maps them on the first insert.

4. Files arrive on the right record

Salesforce Files are stored as content versions and attached to records through a separate ContentDocumentLink record. Open ten migrated cases and check that each file is there, opens, and sits on the right record.

5. Automation is under control

Flows, triggers, validation and duplicate rules all fire during a load unless you plan otherwise. Ask how the tool expects you to handle them, then check what fired during the test run: emails sent, tasks created, fields overwritten.

6. The API budget holds

Salesforce counts API requests across the whole org over a 24-hour period, not per user, and Bulk API and Bulk API 2.0 share an allocation of 15,000 batches per rolling 24 hours. Note what the test run consumed on the System Overview page in Setup, multiply it up to full volume, and confirm your live integrations still have room.

7. Errors are visible and reconcilable

Your bad rows should come back with a clear reason, and you should be able to fix them and rerun only those. The tool should also report source and target counts per object, so the reconciliation does not depend on someone’s spreadsheet.

8. You know where the data goes

Ask for the data flow in writing. Does data pass through the vendor’s servers, an integration layer, or only Salesforce? Your security team will ask, so have the answer before you sign.

Score it, do not just watch it

The eight tests are your selection criteria. Give each one a pass, partial or fail for every tool. Weight tests 1, 2 and 7 most heavily: a broken lookup, a doubled record or an unexplained gap rarely announces itself, and finding it after go-live costs far more than a longer evaluation. Our account of 50 million records with zero data loss shows what that reconciliation discipline looks like at scale.

How we help

We built iSyncSF because the generic tools we used on client projects flattened relationship hierarchies and dropped lookups. It is 100% Salesforce-native, with built-in validation, logging and reporting. Hold it to the same checklist as everything else on your list. If you would rather hand the evaluation over, our data migration team can build the test pack with you, run it, and give you the scores, with a certified Salesforce consultant reviewing the results.

Key takeaways

Proud Salesforce ISV Partner iSyncSF Listed on AppExchange 4.8+★ on the AppExchange 50M+ Records Migrated Zero Data Loss on Every Migration Data Migration Powered by iSyncSF Agentforce & Einstein AI Specialists CPQ & Revenue Cloud 3× Faster Delivery vs. Traditional Certified Salesforce Consultants AppExchange App Experts Experts On Demand in 5 Business Days 24×7 Product Support Serving Clients Worldwide