Move and improve: the migration where you don't have to choose between fast and good
Think about the last time you moved house. If the movers just tip every box, unlabelled, into the new place, you have technically moved. But when you go to cook dinner, the saucepan is in the garage, the pasta sauce is in a bedroom box and long past its date, and the spaghetti is anyone’s guess. You cannot put a simple meal together. Nicer house, same old mess.
Lift and shift means taking your existing warehouse and data pipelines and rehosting them on the new platform more or less as they are. It is fast, it is cheap up front, and when the clock is running it feels like the safe option. We will give it that. What it does not do is leave you better off. The tangle came with you. The data underneath is still hard to find, inconsistent, and undocumented, only now it sits on a newer bill.
That matters more than it used to, because of why most organisations are moving in the first place. The reason is rarely the platform itself. It is what the platform promises next: AI on your own data, and the copilots and agents that answer real questions accurately. But AI can only act on data it can understand, and you cannot build much on a pile of unlabelled boxes. So the whole point of the move, your data and AI maturity, quietly stalls six months after everyone celebrated the migration. What has changed is that the unpacking, the slow part that used to make the better path impractical, no longer has to be done by hand.
Move and improve is the other way to make the same trip. You still move to the new platform, but you unpack and sort as you go. The integration layer gets rebuilt properly on the way in, so what lands is organised rather than tipped in a heap: the pan, the sauce and the pasta all where you can find them, and nothing past its date. And because the change is confined to the layer underneath, the reports and dashboards people rely on keep working while it happens. Every customer is one customer, the history is complete, the relationships between things are explicit, the definitions are versioned, and you can trace any number back to where it came from. That is what AI-ready means, and it is the difference between a home you can live in on day one and a house you are still digging through at Christmas.
There is an honest reason people reach for lift and shift, and it is almost always a lack of time. Unpacking a whole house properly, by hand, takes months, and nobody wants to live out of boxes that long. That trade-off between fast and good is usually too much for most migrations, and it is the thing we set out to remove.
Launchpad is our framework for moving onto a modern data platform (whether that is Snowflake, Databricks, or Microsoft Fabric), and IRiS is the engine inside it. IRiS automates the part that usually eats the calendar, rebuilding the integration layer from your own metadata, so work that would take months by hand takes weeks instead. You get the organised house without the long wait in cardboard. That is the point: move and improve should not cost you the speed that made lift and shift tempting in the first place.
None of this is an argument about whether to modernise your data platform. Nearly everyone is, and nearly everyone should. It is an argument about how you arrive: ready to build, or ready to unpack.
If you are planning a move to a new data platform, we would be glad to talk it through and show you what a move and improve looks like on your own data rather than in the abstract. The Launchpad brochure alongside this sets out how the framework fits together, and our Intelligence Series goes deeper on the foundations underneath it.