Buying — bandwidth, storage appliances, servers, even virtual-server instances — is itself a stage of capacity planning, not an afterthought. It carries **lead time** (quote → approval → delivery → rack → OS install → provision) that must be baked into the plan, because capacity ordered too late arrives after the moment you needed it. Company size shapes the tradeoff: - **Small/startup** — nimble (early Flickr could get a quote and walk to the founder 20 feet away for a check; servers racked within a week) but pays more per unit. - **Large** — multi-level approvals and global data-center logistics slow things down, but bulk buying power lowers unit cost. Either way, the procurement segment of time takes effort like every other step — plan for it. --- *Source: [[The Art of Capacity Planning]] (John Allspaw, O'Reilly 2008) — Ch 1 — Goals, Issues, and Processes in Capacity Planning*