How to Validate a CDE Migration (and Prove the Record Came Across)

23 June

A migration is not finished when the files appear in the new platform. It is finished when you can prove the record came across, completely and intact. Those are very different things, and the gap between them is where most migrations quietly fail.

The usual way of checking a migration is for someone to log in, click through a few folders, see that the files are there, and call it done. That is not validation. That is hope. Real validation is the step that turns "we moved it" into "here is the proof it arrived whole," and it is the part that protects you when a client, an auditor or a dispute comes asking.

You cannot validate what you did not measure first

Validation starts before the move, not after. To prove the destination matches the source, you need to know what the source actually contained: how many files, how many versions of each, which attributes, what structure. That baseline is the thing you check everything against later.

Skip it and "the destination looks complete" is just a guess, because you have nothing to compare it to. This is why a pre-migration audit is not paperwork. It is what makes proof possible at the other end.

What a real validation checks

A proper validation goes well past "are the files there." It confirms:

Completeness. Every file that should have moved did, reconciled folder by folder, source count against destination count, so nothing was silently dropped.
Version fidelity. Each file carries the right number of versions, in the right order, with the right dates, not just the latest one.
Metadata accuracy. Custom attributes, status and suitability codes and ISO 19650 classification are not only present but carry the correct values and are mapped to the right fields.
File integrity. The content of each file is intact and uncorrupted, confirmed by comparing checksums between source and destination rather than assuming a successful upload means a faithful copy.
Relationships and links. Issues still point to the right documents, markups stay attached, and supersession chains remain intact.
Structure and permissions. The folder tree and the access controls match the intended target state, not a flattened approximation of it.
The audit trail. Historical activity is preserved where the platform allows it, and where it cannot be carried across, it is exported and archived as evidence rather than lost.

A spot-check is not a validation

There is a real difference between clicking around a few folders and validating a dataset. The gaps that matter are random and invisible, a missing version here, a dropped attribute there, and sampling by eye is almost guaranteed to miss them.

Real validation is an automated, full reconciliation across the entire dataset, backed by targeted manual checks on the high-value and high-risk items. The automation catches what a human never could at scale. The manual checks confirm that the things that matter most came across exactly as they should.

The output is evidence, not reassurance

The purpose of validation is not to make everyone feel better. It is to produce a record you can put in front of someone.

A validation report documents what was checked, against what baseline, and what the result was. That report is what lets a client formally accept a handover, what an auditor or an assessor asks to see, and what protects you years later if anyone ever questions the record. "We migrated it" is a claim. A validation report is proof of it.

Validation is what lets you say "done"

Without validation, "done" means the files are in the new system and everyone hopes the rest came too. With it, "done" means the record has been shown to have arrived whole, and signed off on that basis.

For a handover especially, this is the line between a clean close-out and a liability you have quietly passed to the asset owner. A migration nobody validated is a migration nobody can stand behind.

How CDE Migrate handles it

Validation is built into every CDE Migrate engagement, not bolted on at the end. Each migration starts with an audit that captures the source baseline, runs with version history and metadata preserved, and finishes with an automated reconciliation against that baseline, plus spot-checks on the items that matter most. You receive a completion report and an audit trail you can hand straight to a client or an assessor.

Want a migration you can actually prove? Talk to us at info@utopiadigital.io.