- Keeping a legacy ERP alive on an aging machine for lookups fails the moment that hardware dies and takes your history with it.
- The modern pattern archives the legacy database into Microsoft Fabric and OneLake, or inexpensive Azure storage for smaller datasets, where it stays queryable through Excel or Power BI.
- Archiving the data and building reports on top are two separate projects, so scope the reachable archive first and save the dashboard wishlist for later.
- Reshaping the old normalized schema into a denormalized model lets your team query plain answers instead of wrestling with the original table structure.
Last month, the owner of an amusement-equipment company told us the real reason a new ERP scared him, and it wasn't the new system. It was the thirty years of history locked in the old one, on software that won't even install on a fresh Windows 11 machine. "If the old system isn't around anymore," he asked, "how do I look up old data?"
It's the right question, and it has a much better answer than most people expect.
"Keep the old box running" is a decaying plan
The instinct is to keep the legacy ERP (enterprise resource planning) system alive on an old computer in the corner, just for lookups. That works until it doesn't. When the application won't install on current hardware and the old machine finally dies, your history dies with it. Counting on aging equipment to outlive your need for the data is a bet you eventually lose.
Archive the data, not the application
You don't need the old front end to keep the old data. The modern pattern is to take the legacy database and archive it into a cloud data store, Microsoft Fabric and its OneLake storage, or for a smaller database, inexpensive Azure cloud storage, where it stays queryable through Excel or Power BI. The clunky old screens go away; the records remain searchable.
A useful way to split the work: archiving the data ("keep it reachable") and building the reports on top of it ("make it pretty") are two separate projects. Scope the archive first, because you can define it now. Save the dashboard wishlist for later, once you know what you actually look up.
One technical note worth translating: old systems store data in a normalized form that's efficient for the software but awful to read directly. Part of the archival work is reshaping it into a denormalized model that reports cleanly, so your team queries plain answers instead of wrestling with the original table structure.
You don't have to migrate everything
This also keeps your new ERP lean. Instead of dragging decades of stale transactions into the new system, you bring forward what you actively need and leave the rest in the cheap, searchable archive. That's less migration cost up front and a cleaner system to live in afterward.
If you're putting off an ERP change mainly because you're afraid of losing access to your history, that fear is solvable, and cheaply. We're glad to show you what a queryable archive would look like. Let's talk it through.