Why plan data migration to Salesforce now
Adopting Salesforce is not just a change of tool, it is a transformation of the way you sell, serve your customers and manage your business. The quality, completeness and governance of the data you inject determine the speed of team adoption, operational efficiency and, ultimately, return on investment. A successful data migration to Salesforce is prepared like a product project: value-centered, iterative, measurable and secure.
Companies that excel in this transition share three characteristics: a clear and prioritized strategy, control of technical and regulatory risks, and pragmatic change management. This guide details the structuring decisions, essential steps and pitfalls to avoid so that your data becomes a competitive advantage in Salesforce.
Data governance and mapping: laying the foundations
Define critical data domains
Before any technical action, identify your key areas: accounts, contacts, opportunities, products, quotes, orders, support cases, activities, and any custom objects. For each, define its source of truth, its priority business uses, its business owners and its minimum quality rules.
Source of truth: ERP for invoiced accounts, marketing tool for leads, support base for cases etc.
Business use: sales priorities, renewals, cross-sell, marketing campaigns, SLA support.
Owner: appoint a business sponsor who decides in the event of arbitration on the model and rules.
Quality thresholds: completeness of addresses, formatting of telephones, valid emails, duplication tolerated at 0 or 1 per 1000 records, etc.
Map flows and dependencies
Produce a simple diagram: where the data comes from, where it passes, what are the consuming systems, and how the hierarchical and relational links materialize. This mapping makes it possible to anticipate the loading order, the need for external identifiers, and integration tests.
Hierarchical relationships: parent-child accounts, accounts and contacts, product families.
Transactional relationships: opportunities, quotes and orders, product-related opportunity lines.
Historization: activities, notes, attachments, emails, critical field histories.
Choosing the migration approach: big bang, iterative, hybrid
The choice of approach depends on your risk tolerance, the initial scope and your ability to operate alongside existing systems.
Big bang under control
A single switch with short application gel. Advantages: strong team alignment, quick end of double run, cutover clarity. Disadvantages: peak risk, need for flawless preparation and rehearsal.
Batch iterative
Incremental migrations by population, region or data domain. Advantages: progressive learning, risk control, smoothed adoption curve. Disadvantages: temporary cohabitation of systems, complexity of synchronization.
Value-oriented hybrid
Rapid migration of high-value critical data (accounts, contacts, active opportunities) in big bang, then iterative catch-up of history and peripheral objects
ques. Often the best risk/ROI combination.
Prepare and clean up: data above all
Deduplication and normalization
Merging duplicates upstream simplifies the mapping and avoids false closures. Define your correspondence keys by type: e-mail and name for contacts, SIREN number and company name for accounts, reference for products. Standardize addresses, telephones, countries, currencies and languages according to your standards.
Targeted enrichment
Take advantage of the migration to enrich what creates value: segments, standardized industries, company sizes, personas, MQL/SQL qualification fields, commercial codes. Light but relevant enrichment accelerates your first dashboards and sales playbooks.
Cleaning up technical legacies
Eliminate obsolete columns, dead values, never used statuses, heterogeneous selection lists. Document each deletion and keep a cold archive if traceability is required.
Migration tools and mechanisms
Data Import Wizard
Ideal for small volumes, one-off loads and simple standard objects. Limited in transformation and fine control, but perfect for POC and pilot migrations.
Data Loader
The on-premise Swiss army knife or CLI for medium to high volume migrations. Handles insert, update, upsert, delete, hard delete. Allows relationships via External ID, offers detailed logs and configurable batches.
APIs and ETL
For complex transformations, multiple dependencies or large volumes, use Bulk APIs and ETL/ELT platforms. You benefit from transactional control, staging steps, isolated rejections and replayability.
Bulk API 2.0 for large volumes and parallelization.
REST and SOAP API for fine-grained operations or personalized orchestrations.
Middleware to map, clean, enrich and log as you go.
Modeling and mapping: linking the business to the Salesforce diagram
Align standard objects and customizations
Maximize the use of standard objects: Account, Contact, Lead, Opportunity, Product, Pricebook, Case, Task, Event, Campaign. Create custom objects only if the business need is not naturally expressed through standards. Document each field: owner, definition, type, validation rules, allowed values.
External ID and upsert strategies
Create External IDs for each major object. They serve as a correspondence key, facilitate upsert, parent-child connections and restarts in the event of failure. Prefer stable, non-intelligent keys, without business meaning that could evolve.
Load scheduling
Step 1: Pricebooks, currencies, territories, users and sales teams.
Step 2: Parent-child accounts and hierarchies.
Step 3: Contacts linked to accounts.
Step 4: Products, families, pricebook entries.
Step 5: Opportunities then opportunity lines.
Step 6: Activities, Notes, Files and History.
Step 7: Support Cases, Installed Assets, Contracts.
Rules, validations and automations: control the impact during migration
Validation rules and flows intended for end users can block or corrupt your uploads. Adopt a secure temporary strategy:
Temporarily disabling or relaxing strict validations during import, with documentary proof and reactivation plan.
Using a specific Migration profile or permission set to
controlled bypass.
Marking migrated records via an Is Migrated field for easier auditing, reporting filters and traceability.
Compliance, security and GDPR
Legal basis and minimization
Migrate only the necessary data. Document the legal basis for processing, apply retention policies, anonymize or pseudonymize sensitive data if it is not essential to business processes.
Rights of people
Anticipate the exercise of rights: access, rectification, opposition, deletion, portability. Set up fields and processes to mark consents and processing restrictions, and sync them with your marketing and customer service tools.
Operational security
Encrypt migration files, limit access to staging repositories, log operations, validate error logs and delete temporary data after cutover.
Sandbox testing, rehearsals and success criteria
Types of environments
Developer/Developer Pro for mappings and unit tests.
Partial Copy for testing with a realistic sample set.
Full Copy for dress rehearsal and end-to-end testing.
Test plan
Technical tests: referential integrity, performance, rejection management.
Functional testing: key business journeys, reports and dashboards.
UAT: real scenarios by user role, objective acceptance criteria.
Set quality thresholds for go/no-go: import failure rate below a defined threshold, post-migration duplication rate, completeness of critical fields, and 100 percent validation of executive reports.
Cutover, rollback and contingency plan
Toggle Window
Choose a niche with low commercial activity, communicate a freeze on changes in source systems, and prepare a 30-minute schedule with identified managers.
Backtracking plan
Clearly define rollback conditions. Keep snapshots, undo scripts, and check the possibility of controlled purging in Salesforce if the quality threshold is not reached. Test the rollback at least once in a sandbox.
Post-go-live stabilization
Allocate a hypercare support cell, track data health KPIs, set up a virtual war room with incident tracking, and plan rapid fixes for priority irritants.
Flow integrations and orchestration
Many migrations are transformed into integration projects. Anticipate the systems that need to sync from day 1: ERP for billing, marketing tools for segments and consents, support for cases, telephony for activities, and data warehouse for analytics.
Event-driven architecture: Prioritize publish-subscribe integrations and queues to reduce coupling and improve resiliency.