Portfolio demonstration with synthetic data

Planning a CRM cutover

A fictional SaaS team is moving accounts, contacts, and open deals into a new CRM. Its first dry run loads the expected number of records. One deal loses its account link, and two contacts collide on the same normalized email.

My role

I checked for contacts left without a link during a pipeline cutover and contributed to moving a partner routing and approval workflow from sandbox to production. This example shows how I’d apply those checks to a broader CRM migration.

From export to cutover decision

Snapshot and map
Load into staging
Validate records and links
Decide whether to cut over

View field mapping

Record	Source field	Target field	Rule
Account	Account key	Legacy account key	Keep the source key and map it to the new account ID.
Contact	Email address	Email	Trim spaces and lowercase before checking for duplicates.
Deal	Account key	Linked account ID	Resolve through the account mapping. Queue unresolved deals for review.

Run the cutover checks

Apply reviewed fixes Merge duplicate contacts C-02 and C-03 into C-02 Restore deal D-02’s verified link to account A-02

Deal D-02 account link

Source deal D-02Source account A-02

Staged deal D-02

Source and staging record counts
Record type	Source	Expected after approved merges	Staging

Validation evidence
Check	Result	Evidence

Rollback plan

Before cutover, take a source snapshot, pause writes, and keep the new CRM isolated while running the final checks.
If a check fails before release, stop the switch, discard the staging import, and resume writes in the source CRM.
If a critical issue appears after release, pause target writes and capture changes made since cutover. Reconcile and replay those changes into the source CRM, verify counts and links, then route users and integrations back.