Data Migration in Cranbury, NJ

By making loss structurally visible rather than trusting care: the source is frozen and backed up before the move; the mapping accounts for every field deliberately (nothing travels by default and nothing is dropped by accident — exclusions are written decisions the principals approve); the rehearsal proves the path on real data; and reconciliation audits the result in three currencies — record counts by category, value totals to the unit, and human-read samples checked by the people who know the data — with discrepancies hunted to zero or explained in writing before the old system is allowed to retire. ‘Guarantee’ in this discipline means an audit trail: at every stage, someone can answer ‘where is every record?’ with evidence. Any provider whose assurance is a personality (‘we’re very careful’) rather than a procedure is telling you which one they’ll rely on.

No — the calendar is a design input, not an inconvenience: the rehearsal runs alongside live work (on a copy), the live cutover sits in the quiet window the practice chooses (the closed day, the small hours), the old system stands throughout as the fallback, and the move is sequenced so the live-critical records — today’s matters, the open files — are the last to move and the first verified. The duration is known from the rehearsal’s measurement, so the window is honest: if your data genuinely needs longer than a night provides, the plan says so in advance and stages the move in parts, rather than discovering it at 4 a.m. Practices remember these migrations as a quiet weekend; the ones they remember as wounds were the un-rehearsed kind.

That is the normal starting condition, and the discovery stage exists precisely to surface it before it can ambush the move: duplicates catalogued with their scale, formats identified (including the predecessor vendor’s), drift and malformed records counted and sampled. The plan then treats the mess as decisions rather than as dirt to sweep: duplicates resolved by merge rules the principals approve (which record survives, how histories combine), the irreparable excluded in writing, the archival separated from the live. Moving messy data without this stage doesn’t clean it — it embalms it in the new system at migration prices. The rehearsal’s reconciliation is where the cleaned result is proven, in the three currencies, before anyone is asked to trust it.

They get a decision, dataset by dataset, instead of a default: what the retention obligations and the practice’s own needs require travels live; what must be kept but is rarely touched is placed in a searchable archive — readable, retrievable, out of the live system’s weight and licence costs; what is genuinely dead is let go, recorded as an approved decision. The distinction matters more than it appears: dragging thirty years into the new platform inflates its cost and slows its daily use, while discarding by accident is the one unrecoverable error in this whole discipline. The archive is designed to be opened — formats the future reads, an index the practice controls — so ‘archived’ never becomes a synonym for ‘lost politely’.

As the records’ own obligations require: access to the data during the programme is named, minimal and logged; copies exist only where the procedure needs them (the frozen source, the rehearsal copy, the backups), each accounted for in the migration’s audit trail and retired at the end; the rehearsal environment is secured to the same standard as the live one, since it holds the same records; and the whole transfer is documented so the practice can evidence to a client or a reviewer exactly how their material was handled, by whom, and when. A practice that holds other people’s affairs cannot treat its own migration as a casual file copy — this programme is built so it never has to explain that it did.

The old system is the standing answer: it retires only after validation passes and the new home has carried real work, and until then it remains the court of appeal — any discrepancy is checked against the frozen source, corrected by procedure, and recorded. After retirement, the audit trail (mapping, rehearsal results, reconciliation reports) is the evidence base for any late-surfacing question, and the new home’s own backups cover operational accidents. The discipline’s deeper protection is upstream: the errors that haunt practices for years are mistranslations, not losses, and those are the ones the mapping and the human-read samples exist to catch while the source is still standing. Post-move surprises in this method are rare, small, and — by design — provable either way.

Priced from the inventory: the discovery and mapping as a fixed engagement (its documents are yours regardless), the move itself against the rehearsed procedure and measured duration, validation included rather than sold as an extra. The variables are honest ones — volume, mess, the archive’s share — and the rehearsal converts most of them into knowns before the live price is final. The comparison is the un-rehearsed move’s true ledger: the emergency rates when it fails mid-window, the months of quietly wrong reports from a mistranslated field, the client conversation nobody wants to have. Custody-grade migration costs what careful work costs; it is priced, scheduled and evidenced like the professional obligation it is.