What a CRM Migration Actually Costs You

Every CRM migration has two prices. The first is on the quote, and it gets negotiated, approved, and tracked. The second is never written down anywhere: what the move breaks on the way through. Teams budget the first price carefully and meet the second one by surprise, months later, one discovery at a time. This essay is about the second price — the four places it hides, and why none of them are visible on a project plan that defines the migration as “records moved.”

The data that doesn’t make the trip

Migrations lose data quietly, and almost never the data you’d notice on cutover day. Accounts and contacts survive — they’re the headline objects, everyone checks them. What goes missing is the connective tissue: activity history, notes, the custom fields someone added three years ago that turned out to encode how the business actually thinks about its accounts. Under schedule pressure, each of these gets triaged as nice-to-have. Individually, each call is defensible. Collectively, they amount to amnesia.

The loss surfaces on its own schedule. A renewal conversation where nobody can see what was promised. An account handoff where the new rep inherits a name and a phone number and none of the relationship. The record still exists, so nothing looks broken — it has simply forgotten everything that made it useful. A record without its history isn’t a smaller version of the record. It’s a stranger with the same name.

Reports don’t survive renaming

Every pipeline report, saved view, and forecast in the old system is built against its stage names, its field structure, its assumptions. Rename the stages, restructure the fields — as migrations almost always do, usually for good reasons — and that reporting layer dies in place. The first Monday after cutover, someone opens the pipeline meeting and the numbers everyone has navigated by are simply not there.

This is the most predictable failure in the entire move, and the most routinely unplanned-for, because it belongs to nobody: the migration plan says “move the data,” and the reports were never data — they were the lens. Translating them stage-for-stage is unglamorous work that has to be scoped as a first-class deliverable, before the move, or the weeks after go-live get spent rebuilding the instruments mid-flight.

Adoption doesn’t transfer

Whatever trust and muscle memory the team had in the old system was earned over years, and none of it moves with the data. Worse: a team that has been through a rough migration before types less into the new system, on the sensible theory that anything they enter might not survive the next move either. Every field that came across empty or wrong confirms the theory.

That’s why adoption can’t be a training session bolted to the end of the plan. It is mostly an architecture question — the fewer things the new system asks reps to type, the less adoption it needs to beg for. Capture that writes itself doesn’t depend on the team believing in the new CRM. It just works, and the belief follows the working.

The mid-quarter stall

A migration takes the system of record — the thing deals are worked out of — briefly offline as a source of truth. Schedule that against live deals and the cost is a week or two of reps half-working from exports and memory, in the exact weeks the quarter needed them selling. The move itself might go perfectly and still cost real pipeline, purely on timing.

The sales calendar, not the vendor’s calendar, should choose the cutover date. That one decision — boring, free, made early — is worth more than most of what gets negotiated in the contract.

What a survivable migration looks like

Every one of these costs is a scoping decision, which means every one of them is avoidable before signature and expensive after it. Survivable looks like this: the mapping agreed in writing before anything moves — stage for stage, field for field, including what happens to history and activity, with nothing silently triaged. Reports translated as a deliverable, not rebuilt as a surprise. Capture designed so the new system fills itself. A cutover date the quarter can afford. And an end date, so the migration is a project that finishes rather than a condition the team lives with.

None of that is heroic. It is what “move the CRM” means once the second price is written on the same page as the first.