A successful CRM migration is not defined by the go-live date, but by the integrity of the data transferred. The post-migration audit is therefore the single most critical phase, yet it is frequently underspecified in project plans and service level agreements. Many organizations rely on superficial metrics like total record counts matching, which is a necessary but grossly insufficient condition for validation.
A rigorous audit must be a multi-layered process designed to verify completeness, accuracy, and relational integrity. I approach this as a series of structured sampling exercises, moving from aggregate to granular. The core principle is to treat the source and target systems as two independent datasets and perform reconciliation, not merely a spot-check.
**Phase 1: Pre-Migration Baselining**
Before any data is moved, you must establish a verifiable snapshot of the source system. This goes beyond a simple record count.
* Extract key aggregate metrics: total records per object (Lead, Contact, Opportunity, etc.), total open opportunity value, count of active accounts, count of records with non-null values for critical fields (e.g., email, phone).
* Document a sample. For each major object, programmatically select a random but traceable sample of records (e.g., 2% or 500 records, whichever is larger). Export the full data for these records, including all custom fields and their unique identifiers. This sample set becomes your primary audit corpus.
**Phase 2: Post-Migration Aggregate Reconciliation**
Immediately post-cutover, run the same aggregate queries against the target system.
* Compare total record counts per object. Any discrepancy triggers a deep dive.
* Compare calculated totals like open pipeline value. A mismatch here often points to failed opportunity stage mappings or incorrect currency conversions.
* Verify that record counts for filtered lists (e.g., "Contacts created in last 12 months") are statistically aligned.
**Phase 3: Granular Sample Record Validation**
This is where most migrations fail their hidden defects. Using your pre-migration sample set:
* For each sampled record, locate it in the target system using the migrated unique ID or a reliable composite key (e.g., email domain + last name + company).
* Conduct a field-by-field comparison. Focus first on mission-critical data: email addresses, monetary values, dates, and picklist values. Pay particular attention to:
* Truncated text fields.
* Mismapped picklists (e.g., a source value of "In Progress" dumped into a target "Other" field).
* Improperly handled nulls and default values.
* Flattened or lost hierarchical relationships (e.g., a child case record now pointing to the wrong parent account).
* Validate a subset of "edge case" records manually: your largest opportunity, the oldest contact, records with the most custom fields populated, records with special characters in data.
**Phase 4: Process and Relational Integrity Testing**
Data exists to support business processes. Static validation is not enough.
* Test that key automated processes work with the migrated data: does the lead assignment rule correctly route a sampled lead? Does a workflow email fire based on a migrated date field?
* Verify that reports and dashboards rebuilt in the new CRM produce logically equivalent outputs. A 20% variance in a "Sales by Region" report is a clear indicator of underlying data corruption or misclassification.
* Spot-check relational integrity. Select ten accounts and verify that all associated contacts, opportunities, and activities are present and correctly linked.
The best way to spot missing records is not to look for what's missing, but to prove that what should be present is indeed present and correct. A missing record is simply the most extreme form of data corruption. By investing in a structured, sampling-based audit framework, you move from hoping the migration was successful to possessing documented evidence that it was. This evidence is also your primary leverage should the migrating vendor or internal team need to remediate defects discovered during the audit window.
- Due diligence first.