Data warehouse migration and modernisation services.

It depends far more on dependencies than on data volume. Discovery and dependency analysis on a mid-sized estate typically takes weeks; the migration is then scoped domain by domain, each reconciled before the next begins. We would rather give an estimate after discovery than before it, and discovery can be bought as a standalone piece of work.

By comparing it with the old one, automatically and repeatedly. Record counts and control totals per load, date coverage and gap checks, null and duplicate profiles, key uniqueness, and metric-level comparison of reported figures. Every difference is investigated and then either fixed or documented as a deliberate change, including the cases where the legacy number turns out to be the wrong one.

That is what the dependency analysis is for. Before anything moves we trace which reports, extracts and semantic models read which objects, and flag the business logic hiding inside them. Dashboards are then repointed and reconciled visual by visual, usually while both platforms run in parallel, so the old estate is switched off only after the new numbers have been signed off.

Usually a mix, decided per domain. Lifting and shifting is faster and keeps reconciliation simple, but it carries forward the modelling problems and often the cost profile. Redesigning is right where the grain is wrong, where legacy logic is unmaintainable, or where query patterns will change on the new platform. Discovery is what tells you which objects belong in which category.

Yes, and it is a common one. The shape of the work is the same: find the sources behind the spreadsheets, document the rules that live in formulas and in people’s heads, model them in the warehouse and reconcile the output against the last few months of manual figures. The difference is that more of the discovery happens in conversation than in the information schema.

We quote after a scoping call, because the effort depends on what discovery finds in your estate. Discovery can be bought as a standalone piece of work, and migrating one reporting domain first gives an estimate for the rest of the estate based on your data rather than on a template. Commercial terms are agreed per engagement.

Yes. MV.tech is a remote-first data engineering consultancy based in Ahmedabad, India, with working hours planned around your timezone. The migration runs on written work both sides can review: a source-to-target mapping that both sides sign off, automated reconciliation results at every stage, and written sign-off per domain before legacy objects are retired. Handover includes the mappings, ER diagrams, validation results, schedules, ownership and operating runbooks.