CRM migration services that run on your hardware. Pre-flight gates, signed audit trail, no vendor in the chain of custody. Powered by Pathfinder.
Leaving Salesforce for a different CRM. Extract complete, transform deterministically, load to the new system. Customer keeps a signed chain of every record movement.
M&A integration, multi-org consolidation, NPSP → Nonprofit Cloud uplift, sandbox-to-prod migration. Pre-flight gates catch the disasters; replay manifests prove what happened.
Bloomerang, Raiser's Edge NXT, DonorPerfect, Dynamics 365, custom legacy databases → Salesforce Nonprofit Cloud or NPSP. The classic mid-market migration done right.
Your migration runs on the laptop of the person doing the work, under their account, with credentials gated by the macOS Keychain. The Keychain ACL boundary isn't new to anyone in IT — every Mac in the building already uses it for Wi-Fi PSKs and cert private keys. We reuse that trust boundary instead of asking you to bless a new SaaS vendor's KMS posture.
Pathfinder is on the laptop. CDM Migration Services run on the same laptop. The vendor — including us — isn't in the chain of custody. The math doesn't let us be.
One less SOC 2 to read. One less DPA. One less subprocessor on the donor-facing privacy notice.
Most migrations don't go wrong in staging. They go wrong on the third Tuesday after go-live, in a conference room. Pre-flight gates exist so that conversation never happens.
Every formula on every field mapping walks an HM-lite type checker. UPPER($Revenue) doesn't save. The classic TEXT(NumericField) zero-truncation bug? Refused at save time.
Runs every formula against real source rows via reservoir sampling. Traces success or failure of every expression. Never touches a target connector. The one row that takes down the load? You see it before you click run.
Pin known-good (input, expected) pairs. The "small tweak" to an eighteen-month-old mapping that quietly breaks production? It shouts before the diff reaches the target. CI for mappings.
Capture writes a row to schema_snapshot. Detector classifies diffs into info / warn / block. Block severity halts migration approval. Catches Friday-afternoon admin changes before Monday's run truncates twelve thousand records.
Fellegi-Sunter log-likelihood from per-field m/u weights. AutoMerge / ManualReview / Discard routed by configurable thresholds. Survivorship rules with the math on the page.
Every mutation chained via SHA-256 with frame-prefixed domain separation. Ed25519 seal every 256 leaves. Your auditor verifies the chain. Protocol published →
You describe the migration. Source, target, volume, deadlines, risks. I ask questions. We determine if there's a fit.
Fixed-scope, project-based engagement. Or retainer if the work is ongoing (multi-org sync, periodic audits). No commission structure. No referral kickbacks. No vendor partnerships.
Schema discovery, dry-run on a sample, conflict surface, gap report. You see what would happen before approving the full migration.
Every record movement signed and chained. Workspace archive at completion. The audit trail is yours forever — even if you stop paying us tomorrow.
Project-based engagements for fixed-scope migrations. Retainer arrangements for multi-org or ongoing migration programs. No published rates — scope determines fit, and fit determines pricing. Same anti-consultancy posture as the rest of the practice.