You see the output of every phase
A migration is mostly irreversible at the end and entirely correctable at the start. So the work is arranged to put the decisions where they are still cheap, with a gate between each phase that has to close before the next one begins.
The short version
We inventory the source: how many records there are, how much media they carry, who has access, and what the source system will actually hand over.
What you getA record census — counts by date, duration histogram, format breakdown, and the list of records the source will not hand over.
Phase by phase
- 01
Profile
We connect to your source system and take an inventory: how many records there are, how far back they go, how much media they carry, and which of them the source will actually release. Most of what people discover halfway through a migration is knowable here, for a fraction of the cost.
What you get: A record census, a duration profile, and the first draft of what will not make it.
GateNone — this is the cheap part, and it is where surprises are meant to surface.
- 02
Scope
We turn the inventory into a prediction: what will land, what cannot, how long it will take, and what it will cost. Telling you before the migration that four thousand of your sixty thousand calls are too short to appear is the difference between a trusted vendor and a support incident.
What you get: A written scope with the exclusion list, a timeline, and the decisions we need from you.
GateYou sign off the exclusion list. Nothing loads until you have.
- 03
Test
A representative sample goes into your real destination — a sandbox where one exists. You look at actual migrated records, in your own system, before anything touches production data at volume.
What you get: Twenty to thirty real records you can open and inspect, with their reconciliation output.
GateThe sample reconciles cleanly, or we do not proceed.
- 04
Resolve
Every user, contact and account in the source is matched to its equivalent at the destination. Ownership is permanent once a record is created, so this is a gate rather than a stage that runs alongside loading. Anything ambiguous — usually people who have left — comes to you as a decision.
What you get: An approved mapping, and a queue of decisions rather than guesses.
GateEvery person is mapped, or you have explicitly decided what happens to the ones who cannot be.
- 05
Load
Records are written to the destination on a schedule that leaves your live systems the capacity they need. Every record is checkpointed, so an interruption resumes from the exact point it stopped rather than starting again.
What you get: A live progress ledger you can watch, updated as records land.
GateNone — but progress is visible throughout, and a failure costs minutes.
- 06
Reconcile
We read every record back out of the destination and confirm it is really there — the right owner, the media attached, the relationships resolved. A write that reported success is not evidence; the read-back is. Anything that did not land is listed with the reason.
What you get: A line per source record: where it went, what state it reached, and why, if it did not.
GateThe report is the deliverable. This is what you are actually buying.
We run it. You decide the things only you can decide.
This is not a tool we hand you. We do the work, and you review the output at each gate — the exclusion list before scoping closes, the sample records before production loading, the mapping before anything is written permanently.
The decisions that stay with you are the ones with business consequences: which duplicate record survives a merge, where a departed rep’s history should live, whether a category of records is worth migrating at all. We will give you a recommendation every time. We will not make the call quietly on your behalf.
Start with the profile.
It is the cheapest phase and the one that answers most of the questions. Tell us what you are moving and we will come back with what we find.