On Dataconomy: What Happens Between Product Growth and a Server Purchase
Dataconomy published my walkthrough of how a single demand signal travels from a product decision to a hardware purchase order.
Dataconomy published a piece built on my work, following a single demand signal through a large technology company to show how infrastructure decisions actually get made – and how they move from spreadsheet guesswork to an explainable decision system.
The interesting part of capacity planning is not the forecast. It is the translation. A product team ships a feature that changes a usage curve. That curve becomes a demand estimate. The estimate becomes a hardware requirement, which becomes a purchase order with a lead time measured in months, which becomes a rack in a building. At every hop the signal is re-interpreted by someone with different incentives and different units, and the loss at each hop is where the money goes.
Most organisations only instrument the two ends. They know what the product shipped and they know what they bought. The chain in between is a set of spreadsheets and conversations that nobody can reconstruct six months later when the question is why we bought what we bought.
