The Executive Guide to Modern Data Platforms
Warehouse, lakehouse, or fabric? A plain-language guide for leaders weighing a data platform investment — what the terms mean and what actually drives cost.
Data platform proposals arrive wrapped in vendor vocabulary — lakehouse, fabric, mesh, medallion — and the terminology obscures a set of decisions that are actually quite simple. This guide translates the landscape for executives who need to approve an investment, not administer one.
The Decisions That Matter
Strip away the branding and three choices drive most of the cost and value: where your data lives, who is allowed to change its definitions, and how fresh it needs to be. Everything else — engine, format, vendor — follows from those three, and can be revisited later far more cheaply than the three themselves.
The most expensive mistake we see is buying architecture for imagined future scale. A governed, boring platform that your team can operate beats an ambitious one that becomes a second legacy system.
What to Ask Any Proposal
Ask who owns each critical data set once the platform is live. Ask what the monthly running cost is at your current volume, not the demo's. Ask how a business definition gets changed, and how long that takes. Weak answers to these questions predict programme trouble far more reliably than any technology choice.
The full paper covers reference architectures on Azure and Google Cloud, a cost model worksheet, and a phased adoption sequence that funds each stage from the savings of the last.