Most CRM migrations fail quietly. The data moves, the new system goes live, and then a few weeks later someone notices that a handful of six-figure deals lost their close dates, or that lead routing stopped firing, or that the forecast no longer ties to reality. A CRM migration is not a database export — it is open-heart surgery on the system your revenue team depends on every single day. The job is to move platforms without losing a single open opportunity, breaking a single automation, or eroding the trust your reps place in their pipeline.
This guide lays out the migration plan we use in our engagements to protect pipeline and data integrity from kickoff to cutover. It is opinionated, sequential, and built for B2B teams that cannot afford a blackout window.
Why CRM Migrations Lose Pipeline in the First Place
Pipeline does not disappear because the export tool malfunctioned. It disappears because of decisions made — or skipped — long before anyone touched a CSV. The most common failure patterns are predictable:
- Silent field mapping gaps. A custom stage, a multi-select picklist, or a currency field has no clean equivalent in the destination, so it gets dropped or flattened.
- Reference breakage. Opportunities point to contacts that point to accounts that point to owners. If the IDs re-key during import and the relationships aren’t rebuilt, records orphan.
- Automation cutover gaps. Routing rules, lead scoring, and workflow triggers exist in the old system but go live late — or never — in the new one.
- No reconciliation. Teams trust the import summary screen instead of comparing record counts and dollar totals before and after.
- Bad timing. Cutover lands mid-quarter or during a campaign push, so any disruption hits when stakes are highest.
Pipeline is rarely lost during the migration itself. It is lost in the assumptions you didn’t validate before you pressed go.
If your current instance is already messy, that mess will migrate with you and get harder to untangle later. This is the moment to take honest stock of what you’re carrying over.
Start With an Audit, Not an Export
Before you map a single field, you need a clear picture of what you actually have. The temptation is to treat migration as a technical lift-and-shift, but the highest-leverage work happens upstream. We typically begin every migration with a structured marketing operations audit so we know exactly which objects, fields, automations, and integrations are load-bearing.
Your pre-migration audit should produce four artifacts:
- An object and field inventory. Every object (accounts, contacts, leads, opportunities, custom objects), every field, its type, and whether it’s actively used. Fields touched by nothing in the last year are candidates to retire, not migrate.
- An automation map. Every workflow, routing rule, scoring model, and trigger, plus what fires it and what it writes to.
- An integration list. Every system reading from or writing to the CRM — marketing automation, billing, support, BI, enrichment, ad platforms.
- A data-quality baseline. Duplicate rates, completeness on key fields, and obvious anomalies.
That last artifact deserves real attention. Migrating dirty data just relocates the problem and inflates your record counts, which makes reconciliation harder. Run a focused cleanup pass — our CRM data hygiene system covers the deduplication and standardization steps — before you freeze anything. Clean first, then move.
Decide What Not to Migrate
Not everything earns a seat in the new system. Apply a simple test to each object and field: is it actively used in reporting, automation, or a rep’s daily workflow? If the answer is no, archive it to a flat export and leave it behind. Closed-lost opportunities older than a few years, dead leads that never engaged, and orphaned custom fields rarely justify the migration risk they introduce. A leaner dataset migrates faster and reconciles more cleanly.
Map Fields and Relationships Before You Touch Data
Field mapping is where pipeline integrity is won or lost. Build a mapping document — a spreadsheet is fine — with one row per source field and columns for the destination field, the data type on each side, transformation logic, and an owner sign-off.
Pay special attention to these high-risk areas:
- Opportunity stages. Stage names, probabilities, and the underlying sales process must map exactly. A misaligned stage corrupts your forecast on day one.
- Picklists and multi-selects. Destination values must match source values or you’ll silently lose data. Pre-create every picklist option before import.
- Owners and teams. Record ownership drives routing, reporting, and rep compensation. Map every user, including inactive ones whose records still need an owner.
- Currency and dates. Watch for format mismatches and timezone shifts that quietly move close dates.
- Relationships. Decide how you’ll preserve the links between accounts, contacts, and opportunities. Most modern CRMs let you carry an external ID; use it as the join key so relationships rebuild deterministically.
The single most important technical decision is the external ID strategy. Stamp every record with its original system ID in a dedicated external-identifier field. This gives you an immutable join key for rebuilding relationships, lets you re-run imports idempotently without creating duplicates, and provides an audit trail back to the source of record. Skip this, and every re-import becomes a duplication risk.
Protect the Pipeline During Cutover
This is the phase that determines whether your reps trust the new system. The goal is a controlled cutover with a known freeze window and a tested rollback path.
Run a Full Dry Run in a Sandbox
Never let production be your first real import. Load a complete copy into a sandbox or trial instance, rebuild the automations, and have a few reps work live deals through it. A dry run surfaces the mapping gaps, broken triggers, and permission problems that no spreadsheet review will catch. Time the import while you’re at it, so you can size the freeze window accurately.
Sequence the Import Correctly
Order matters because of dependencies. Import in this sequence so every record has something valid to point to:
- Users and teams
- Accounts
- Contacts and leads
- Opportunities
- Activities, notes, and attachments
- Custom objects and junction records
Loading opportunities before accounts exist guarantees orphans. Respect the hierarchy.
Choose a Freeze Window That Respects the Quarter
Pick a low-activity window — typically a weekend, and ideally not in the final weeks of a quarter when reps are racing to close. Communicate the freeze clearly: a defined period where no one edits records in the old system. During the freeze, export, import, rebuild automations, and reconcile. Reps should know exactly when the old system goes read-only and when the new one opens.
Re-enable Automation Deliberately
Bring automations back on in a controlled order, and validate each before moving to the next. Lead routing is the one to watch most closely — if routing misfires after cutover, fresh inbound leads stall in limbo and you lose pipeline at the top of the funnel just as you’ve protected it everywhere else. Test routing with a few seeded records before you open the gates to live traffic.
Reconcile Before You Declare Victory
A migration is not done when the import finishes. It’s done when you’ve proven that what came out matches what went in. Reconciliation is non-negotiable, and it should be quantitative.
Build a reconciliation checklist and work it methodically:
- Record counts by object, source versus destination, with documented explanations for any deltas (the records you intentionally left behind).
- Open pipeline dollar value in total and by stage — this is the number leadership cares about most, so it must tie out exactly.
- Ownership distribution — confirm records landed with the right owners and none defaulted to an admin account.
- Relationship integrity — spot-check that opportunities link to the correct accounts and contacts.
- Required-field completeness on the fields that drive routing and reporting.
- Automation firing — confirm workflows, scoring, and routing actually execute on new and edited records.
Keep the old system read-only for a defined grace period after cutover. If a discrepancy surfaces, you’ll have the source of truth to investigate against, and the external IDs you stamped earlier make tracing any record back trivial.
A Condensed Migration Checklist
Use this as the spine of your project plan:
- Audit objects, fields, automations, and integrations.
- Clean and deduplicate; decide what not to migrate.
- Build a field-and-relationship mapping document with owner sign-off.
- Define your external ID strategy.
- Run a full sandbox dry run and time it.
- Schedule a freeze window outside quarter-end.
- Import in dependency order.
- Re-enable automations deliberately, routing last and validated first.
- Reconcile counts, pipeline dollars, ownership, and relationships.
- Keep the source system read-only through a grace period.
Migrations reward patience on the front end. The teams that spend their effort on auditing and mapping have boring, uneventful cutovers. The teams that rush to export are the ones explaining to leadership where the forecast went.
Move Platforms Without Moving Backward
A CRM migration done well is invisible to your reps and your forecast — the system changes, the pipeline doesn’t. That outcome comes from disciplined sequencing, honest data work, and reconciliation you can defend with numbers, not from the migration tool you happen to pick.
If you’re planning a move and want a partner who has run these without losing pipeline, that’s the work we do. Explore our services or get in touch and we’ll help you build a migration plan that protects every open deal.
