Functional & Business Analysis

Hire Data Migration Specialists

Specialists for the part of a go-live that decides it — extraction, cleansing, load and the reconciliation that proves it worked.

Tell us what the work is

Data migration is the part of a cutover most likely to be underestimated and most likely to stop a go-live. It is scheduled as a task and behaves like a project, because the real work is not moving the data — it is discovering what the data actually contains.

Profiling first is the whole discipline

Profiling first is the whole discipline. Not what the field is called, but what is in it: how many customers have no country, how many part numbers carry a suffix the new system will not accept, how many open orders reference a supplier that no longer exists. Every one of these becomes a decision the business has to take, and taking them in week two is a conversation while taking them in cutover week is a delay.

Profiling means counting the actual values in the source before anyone agrees a mapping. It is unglamorous, it takes days rather than weeks, and it converts most of the surprises on a migration from cutover-weekend problems into ordinary decisions with time attached.

The load will be run more than once

The load has to be repeatable. It will be run for a mock, run again after the mapping changes, run again in the dress rehearsal, and run once for real. A load that appends rather than matching on a business key produces duplicates on the second run, and duplicates in master data are still being cleaned up a year later.

Repeatable also means restartable. A load that fails at eighty per cent should be able to resume or be re-run in full without a manual clean-up step, because the manual clean-up step is where somebody deletes the wrong rows at three in the morning.

What a mock load is actually for

A mock load is not a rehearsal of the technology. It is a rehearsal of the decisions, and its output is a defect list rather than a loaded system.

Each mock should produce a measured record count, a measured elapsed time, and a list of rejected records with a reason and an owner. Three mocks with a falling reject count is evidence that the migration is converging. One mock with a green tick is evidence of nothing, and it is what most programmes have when they enter cutover.

Deciding what does not come across

Every migration leaves something behind, and the mistake is leaving it behind by accident. Closed transactions beyond an agreed date, customers with no activity in five years, attachments nobody can open — each is a legitimate exclusion and each has to be a decision with a name on it, because someone will look for that record in month three.

The corollary is that the old system’s retirement date is a business decision, not an IT one. If history is not migrating, somebody has to say how it will be retrieved and for how long, and that answer belongs in the cutover plan rather than in an email.

Reconciliation is what actually gets signed

Then reconciliation, which is what actually gets signed. Counts by object, values by control total, and a documented explanation for every deliberate difference — records intentionally not migrated, balances rounded, history truncated at an agreed date. Without that, nobody can say on the Monday whether the migration worked, and the argument is unwinnable in either direction.

Record counts per object, source against target, with the exclusions stated
Financial control totals agreed with finance before the load rather than after it
A reject report that has been worked through to zero or to a signed acceptance
A named person from the business against each object, who signs what they can defend

When you do not need a migration specialist

A single small master data file moving into a system your own team owns does not need this role. It needs a spreadsheet, a check and somebody paying attention.

Nor is this the right hire when the underlying problem is that the business has not decided what its data means. Two departments disagreeing about what counts as an active customer is a governance question, and running a migration over it moves the disagreement into the new system.

The role earns its place on a cutover with several objects, real volumes, a fixed date, and a reconciliation somebody senior will have to sign. That combination is where migrations fail, and it is the reason they are staffed separately from the implementation team.

Teams are built for companies in the United States and the Gulf — the UAE, Saudi Arabia, Qatar, Kuwait, Bahrain and Oman. The engineers are in Pune, which matters mostly for the clock. Dubai is ninety minutes behind us and Riyadh two and a half hours, so a Gulf team shares almost the whole working day. New York is nine and a half hours behind, so American engagements run on a written handover and one fixed overlap window rather than on a standing call — a real constraint, and better stated than discovered.

What these engineers do

Profiling the source before anyone agrees a mapping, so the surprises come early
Cleansing and de-duplication rules the business owns rather than IT inventing
Load design that is repeatable and idempotent, because it will be run more than once
Reconciliation by count and by value, signed off against the source system
Mock loads, dress rehearsals and a cutover plan with a defined point of no return

Delivered AI-first

Specialists use AI assistance to profile and classify large source datasets, draft mapping and transformation logic, and generate reconciliation queries. Every figure a business signs off is verified against the source by a person.

Other Functional & Business Analysis roles

Business Analysis

Analysts who write the exception path, not only the happy one, and produce something an engineer can build from.

What an unfilled engineering role costs while you hire — worked out on your own numbers.

Tell us what the Data Migration work is.

Roughly what it involves, the seniority you need, and when it has to start. We will say what it takes to staff it, or say honestly that we are not the right people for it.